Subscribe
Share
Share
Embed
After Bryan's binge reading of all things DEC (Digital Equipment Corporation), the Oxide and Friends talk through the its meteoric rise and slow descent into ... Compaq.
Show Notes
Oxide and Friends Twitter Space: June 13th, 2022
The Rise and Fall of DEC
We've been holding a Twitter Space weekly on Mondays at 5p for about an hour. Even though it's not (yet?) a feature of Twitter Spaces, we have been recording them all; here is the recording for our Twitter Space for June 13th, 2022.
In addition to Bryan Cantrill and Adam Leventhal, speakers on June 13th included Tom Lyon, Dan Cross, Tim Bray, Ian Grunert, and XXX. (Did we miss your name and/or get it wrong? Drop a PR!)
Some of the topics we hit on, in the order that we hit them:
- Pronunciation and mispronunciation
- Bryan's DEC reading list:
- The Ultimate Entrepreneur by Glenn Rifkin, George Harrar
- Learn, Earn & Return - My Life as a Computer Pioneer by Harlan Anderson
- High-tech Ventures: The Guide For Entrepreneurial Success by C. Gordon Bell, John McNamara
- Computer engineering: A DEC view of hardware systems design by C. Gordon Bell, J. Craig Mudge, John McNamara
- Creative Capital: Georges Doriot and the Birth of Venture Capital by Spencer E. Ante
- DEC Is Dead, Long Live DEC: The Lasting Legacy of Digital Equipment Corporation by Edgar H. Schein, Paul J. Kampas, Michael M. Sonduck, Peter S. Delisi
- @1:29:05 Ian mentions Computer History Museum's oral history program prompting strong recommendations:
- Ian: Bernie Lacroute
- Adam: Pierre Lamond
- Bryan: Dave Cutler
If we got something wrong or missed something, please file a PR! Our next Twitter space will likely be on Monday at 5p Pacific Time; stay tuned to our Twitter feeds for details. We'd love to have you join us, as we always love to hear from new speakers!
Creators & Guests
What is Oxide and Friends?
Oxide hosts a weekly Discord show where we discuss a wide range of topics: computer history, startups, Oxide hardware bringup, and other topics du jour. These are the recordings in podcast form.
Join us live (usually Mondays at 5pm PT) https://discord.gg/gcQxNHAKCB
Subscribe to our calendar: https://sesh.fyi/api/calendar/v2/iMdFbuFRupMwuTiwvXswNU.ics
Hey, Yeah.
Speaker 2:Hey. How you doing?
Speaker 1:Oh, man.
Speaker 3:That's an auspicious start.
Speaker 1:Well, I I just feel like I mean I mean, there's really only one word to describe today, honestly.
Speaker 3:So should a firestorm?
Speaker 1:Firestorm doesn't quite capture it. I feel like a firestorm, but, like, but but broader, hotter.
Speaker 3:Really? If only there were a good descriptor. Conflagration. There you go.
Speaker 1:That's it. Still sounds so wrong. It still sounds like I it it sounds like I feel like I'm saying, like, congregation. I feel like I I I individually like, I'm visualizing a pew engulfed in flames.
Speaker 4:Well
Speaker 1:So so just for do do you wanna, like, you created this mess or or at least
Speaker 3:after I I mean, sort of. I mean, like, arguably, you created this mess.
Speaker 1:I'm not sure how. Go ahead.
Speaker 3:Okay. So apparently, you've been saying the word conflagration as con conflagration?
Speaker 1:Don't don't make it sound like it sounds so Alright. Yes.
Speaker 5:For
Speaker 3:your life?
Speaker 1:My life.
Speaker 3:And I, you know, I think we we were talking this morning, a bunch of Oxide folks, as we do in the morning, about, I think we were talking on the subject of, like, words pronounced wrong?
Speaker 2:No. No.
Speaker 1:No. We weren't. Don't you dare leaventhal me right now. Okay.
Speaker 6:We were not
Speaker 1:we're not discussing
Speaker 7:at all.
Speaker 3:What what was the topic?
Speaker 1:The you like, you brought this up out of out of clear blue sky. I think. I think. I don't okay. We I think I think because I I think we've got some witnesses who were there.
Speaker 1:I believe that that's that we were talking I think we were talking about the the crypto bust or whatever.
Speaker 2:Yeah.
Speaker 1:And you have been listening to some recording. You had listened to a recording of me saying this recently.
Speaker 3:No. No. No. So you at the on the Twitter space, you know, 2 or 3 weeks ago or whatever, you had said it. And in the moment, I was like, that was a
Speaker 8:that was
Speaker 3:a that was an unusual slip or whatever. Yeah. Exactly. But, you know, I wasn't gonna have a whole Twitter space dedicated
Speaker 1:to that subject. Steve mispronounced a word, didn't he? Clavnick did. Then I called him on it, and then you called him. Yeah.
Speaker 1:Yeah. Yeah. Yeah.
Speaker 3:Yeah. Yeah. Yeah. Yeah. Right.
Speaker 3:Yeah. No. You're right. Don't don't
Speaker 1:you, Leventhal? I was actually Leventhal on you. I was actually Leventhal on you. I I was, the to by the way, is to recall history in a way that is favorable for oneself. Is that I guess that's that's the most abstract.
Speaker 3:I think that's I I think that it was it was coined because early in my career, I would say, well, Brian, like, you argued the opposite, like, a week ago. And Yes. Of course, like, you I mean, it it's a it's a bit like, there's not much you can say when someone claims that you forgot that you made an opposite claim. It sort of it puts one on the back foot.
Speaker 1:Right. This position that you've adopted so radically is exactly the position that I adopted when we had this discussion in certain period of time ago, and you shouted me down. That's that's right.
Speaker 3:That's right. And and and kinda spikes up strip someone. I mean, I don't I don't think I've ever used it, like, fraudulently, but then maybe not. Maybe I misremembered.
Speaker 1:That's right. But so we and what was I can't remember that that Steve is using. Anyway, so but and then this, of course, then I made the mistake of confiding in the Internet that I had done this. And I mean, it was interesting. Did you
Speaker 3:see all that? I don't know if I saw all of it, but I definitely saw a lot of folks, with a lot of interesting, I mean, I I think a lot of bearing of the soul, which you you helped people through.
Speaker 1:Well, so the the 2 of note were, flaccid.
Speaker 3:Yeah. I I didn't know, the the actually correct pronunciation. I did greatly appreciate that your, definitive source on that as it is on most things that you that you like was Silicon Valley, the HBO series.
Speaker 1:HBO Silicon Valley where Richard Hendricks, where Ehrlich insults a venture capitalist by by describing his genitalia as flaccid, and Richard Hendricks points out actually, it's pronounced flaccid. Most people don't know that. So and then the other one and this is one that to me is I so that one, obviously, I accept. The one you mentioned from your parents has not come up in that thread.
Speaker 3:I think because it's so and I say my parents. So that and it's actually my mom who's who's like an English major.
Speaker 1:You you refused with so your father's good name with us.
Speaker 3:Oh, my my dad would have he's just like, I care. It's fine. Like, whatever your mom says. But, I mean, my mom was, like, the editor of the college paper and stuff like that. So, but it I I I think, one time, my mom was saying that she frequented the Starbucks, and I was like, mom, come on.
Speaker 3:Stop. Like, don't don't just come on. Don't be so highfalutin. She's like, that's the word. Look it up.
Speaker 2:And And I did.
Speaker 1:And you did.
Speaker 3:And and she's, you know, she's basically right. And and with
Speaker 1:She's at the point of right. She she's a kind of right, but she said not but, you know, another broader or more deeper sense she's wrong.
Speaker 3:Well, I mean, I think the asterisk, depending on which dictionary you look at, will say, well, with with varying degrees of sneering, it's like, you know, the common folk now just say frequent. And so I guess that's fine.
Speaker 1:The the peasantry says frequent. You frequent the Starbucks. I mean, you if you frequent a Starbucks, you're gonna get what you deserve in that I feel that that no one is gonna let that pass. I don't think.
Speaker 3:Yeah. I mean, we were in, you know, in New England at the time. So we're a little more, you know, little more rectitude.
Speaker 1:That's a little more rectitude. Right. You just come out of your your town meeting. Oh,
Speaker 3:that's right.
Speaker 1:So the other one though that I definitely reject is did you see how do you pronounce b I o p I c?
Speaker 3:I I pronounce this as biopic. Like, the fact about bio. I I
Speaker 1:You do not.
Speaker 3:Honest to god. I don't have to tell you. I do.
Speaker 1:You so you've always pronounced that biopic? That doesn't make that
Speaker 3:Not like biopic. Like like, it's some kind of myopia?
Speaker 1:Oh, like a biopic? Like, it's a skin flick for, like, zoologists? I mean, it sounds like mammalian porn.
Speaker 3:Well, was I wrong?
Speaker 1:Well, sadly, you're not wrong. You're right.
Speaker 3:Sadly, it was mammalian porn. Thank you. Put those points on the board.
Speaker 1:Biopic sounds wrong to me. It's I so I that word, I'm just like, you know what? Actually, don't need that word, really. I don't I don't use it very frequently. Unlike conflagration, which I'd actually do use frequently and have now misused very I'm like, I am I'm gonna wake up in cold sweats remembering talks in which I've used this thing.
Speaker 3:Well, I if to to, to, you know, expose my own, screw up, there are words that I sort of, like, play with and, like, mispronounce on purpose, but then accidentally mispronounce. One of them that I think is an interesting word is Antipodes, which I which looks like antipodes. And as one is, like, looking at one scrabble board or whatever, you know, antipodes is the thing that screamed out to me. And the other day, like, not jokingly, I said ant antipodes. But in fact, like, I do know his antipodes, but I just, like, I got wrapped up
Speaker 2:in the act of that.
Speaker 1:When you said antipodes, I'm like, wait a minute. I've been What? Antipodes. Is that is that wrong? I mean, this is like a day I I knew it's, like, whole years where everything I've done is wrong today.
Speaker 1:This is real It's
Speaker 3:all crashing down around
Speaker 1:It is all crashing down around it. So there there were but anyway, a lot of interesting response in that. Right? We won't we won't belabor it here because we we we wanna talk about the conflagration that is Digital Equipment Corporation. So we will So
Speaker 3:so Yeah. Good. So what what what got you on this Deck bent? Like, clearly, you've been binge reading everything about Deck that has ever been written. What what kicked this off?
Speaker 1:So Deck, I feel, has been my white whale in that of I have I came up very much post Deck, and I I don't think I've ever worked on a deck machine. And I have known how influential it is, certainly. But I just have not spent a lot. I've always kind of been meaning to spend time understanding the company a little bit better, or a lot better and just haven't. And I recall the book that Deck is Dead Long Live Deck that I was that was in the missing box of books from the previous mental breakdown on Friends of Oxide, when we did our our book roundup.
Speaker 1:I think we mentioned Deck is Dead Deck is Dead Longwood Deck in that someone else had read that. Maybe, Dan, maybe it was you who'd read that. And so I I I finally like, I found that book, located the box of books, found that book, read it, and it's the deck is really interesting and I just ended up, you know, kind of one book led to another led to another led to another. So I think I've got 6 books in front of me.
Speaker 3:God. So okay. So what what is the reading list?
Speaker 1:So the reading list well, so what were you what? Do you want it in the order that I read it or the order that I'd recommend it? Because these books are not all good.
Speaker 3:Let let's get the, the the recommendations.
Speaker 1:So the the the my recommendation actually is that I think is a very it is much better than Deck is that long lived Deck is, but not a great title, but the the book is The Ultimate Entrepreneur, the Story of Ken Olsen and Digital Equipment Corporation. So this is an unauthorized book about DEC. The thing that's really interesting about I mean, there's a bunch of things about this book. It's, like, it's it's there are reporters, Glenn Rifkin and and George Herrard. They're they're reporters.
Speaker 1:It's well researched. It's, it's well written, and the, it's written at the height of deck, like 1989. But in reading this and so it's kind of like especially with, like, the the reason they titled it the ultimate entrepreneur is I think Fortune had named him the ultimate entrepreneur maybe the year prior
Speaker 6:to prior
Speaker 3:The the entrepreneur of the century.
Speaker 1:Entrepreneur of the century in Ken Olsen. And so you think, like, okay, this is gonna be kind of a hey, geography about, I'm literally checking every word I'm saying right now to make sure I'm not, like it's it's hagiography. Right? Can someone say that?
Speaker 3:I I I'm not brave enough to even attempt it.
Speaker 2:Actually actually, no. It's not. It's actually hagiography.
Speaker 1:Is that true?
Speaker 2:Yes.
Speaker 9:Get out. Get out.
Speaker 6:I'm I'm really sorry. If I
Speaker 1:find the button, there's a button here. You could you could, alright. Is this really hagiography? God. This is a I'm afraid to talk today.
Speaker 1:Alright. Well, the anyway, whatever. A a biography filled with nothing but praise, but it's not. It's it is very much a warts and all view of deck. And reading that book, and it's obviously it's hard because we can't go back in time and read it in 1989 or whatever.
Speaker 1:But certainly reading it now, that book is like, wow. This thing was a time bomb in so many different ways. Like, this was so this company, there is so much that is deeply wrong in this company that it is what of no surprise that this book that it is the height and it is kind of like the long distance from there.
Speaker 3:So so the the title was almost tongue in cheek?
Speaker 1:I don't think it was meant to be tongue in cheek. It's it is not a takedown of Deck. I think it is really trying that's part of what makes it good is it's trying to actually tell the story of deck, but it is, it's just very revealing about a whole lot that is deeply wrong with the company when I mean so in particular and there there are there are a bunch of things that we can kinda talk about about what's right with deck and what's wrong with deck. And I've got a bunch of folks who probably use DEC systems or maybe even work for DEC, so I wanna get a lot of other perspectives on this. But one of the and the here's why I think DEC is interesting.
Speaker 1:I mean, obviously, I I think computing history is is always I think history is interesting because, like, it has a lot to teach us. I think deck is interesting because they're you know, why and Deck is is very much kinda rhymes with Sun. Tom, obviously, it'll be interesting to get your perspective on this. But DEC and Sun both filled kinda similar cultural roles, but in different eras. And they were companies that people were rooting for that ultimately, even though they had a lot of success, like, we talk about the failure of Deck and we talk about the failure of Sun, even though both these companies have, like, totally outsized success.
Speaker 1:And, you know, one of the things that, like, keeps occurring is, like, why do we talk about the failure of these companies when they succeeded so much or they had so much success? And I think it's because because they were both at their core companies that were that were decent, that really had the the engineer at heart, the customer at heart, and they were the companies that people were rooting for. They wanted these companies to to have even more success and and kind of get kind of permanence and endurance. And even though both companies lasted a really long time, I mean, DEC is founded in 1957 and is is acquired in 1998. Like, that's a long run.
Speaker 3:I know. It's nuts. I I I definitely wanna talk a bit some point about, like, the Sun analogs because, man, they felt really strong. But but but before we do that, let's get to the rest of the list if if you think it's worth Yeah.
Speaker 1:Yeah. Sure. Yeah. So let me just I'll I'll I'll rattle this off because then we'll we'll kind of end up hitting on those. But so the ultimate entrepreneur, really good book.
Speaker 1:If there's one book to read on deck, it's that. That book sent me to a really interesting book written by the cofounder of DEC, who is damn near written out of the history books as kind of told by Ken Olsen. It's this guy, Harlan Anderson. And Harlan Anderson is really I think is is our is a really interesting person. So he has he has a biography, an autobiography, called Learn, Earn, and Return.
Speaker 1:Not can't have a great title, but My Life as a Computer Pioneer. And I I loved it. I thought it was it was very personal, and you can definitely see why he and Olsen why Olsen ultimately forced him out, and talk about that in detail. So that that is really good. And, actually so this is just a a kind of an aside on this.
Speaker 1:So the book that I have is autographed by Harwin Anderson, and it's made out to to somebody. It's made out to a name, to a Tracy Mackle. And, Tracy Mackle, thanks for the illegible. Best wishes, Harlan Anderson, January 2010. And this is from the it's a sticker from the Senior Men's Club of New Canaan.
Speaker 1:It is new is it New Canaan? Is that right?
Speaker 4:Yes. Yes.
Speaker 3:Yes. Yes.
Speaker 1:That's a Ken. I mean, it's like a pronunciation guide and everything. But it which it was, like, super personal. And as it turns out, like, this is a guy he clearly this this person, I found his obituary relatively quickly, lived in New Canaan and clearly knew Anderson, and, you know, I kind of had this, like, kind of got this, like, a cherished copy that I bought for, like, a buck 98 on Amazon. And, you know, it's it was almost, like, sad in a way because clearly this guy passed away somewhat recently.
Speaker 1:And, you know, his kids are probably going through all his books trying And that one is really, really good, I think. And then that led me to Gordon Bell's High-tech Ventures, the guide for entrepreneurial success, which is a time capsule, I would say. There there's some good advice in it, but it is basically a time capsule and it's mainly capturing it's written in 1991, and it's very revealing. It's got some good advice in it, and then it's got some atrocious advice in it. I tweeted something out.
Speaker 1:I just this thing I tweeted out, like, last week about talking about the Japanese method of software.
Speaker 10:Did you see this?
Speaker 3:Oh, yeah. I did see this. Yes.
Speaker 1:Very kinda cringey talking about how, like, software should be done in the big open room in the style of a factory. You should have thousands of people working on it and so on.
Speaker 3:And It contrasted with, you know, software engineers who wanna see themselves as artisans and Right. And craftspeople and and and sneering at that notion.
Speaker 1:Sneering at that notion. Exactly. And kind of that that's a reflection of kind of the the early nineties. So that's actually that's an interesting book in a revealing book. I don't know if I'd recommend it, but it's more of a time capsule.
Speaker 1:Then the the book that I also have by Gordon Bell is Computer Engineering by Gordon Bell, which is it it it's Gordon Bell, which I I know I saw John Masters here. I'm I I I'm sure that John's got a copy of this one. That is a I obviously see Gordon Bell, indisputable computing pioneer, did a lot of incredible work, and that's him describing all these machines technically in detail, which is was is also really fascinating. That one I did. I've not read cover to cover, but there's a lot of interesting stuff in there.
Speaker 1:And then the last book and in some ways, the in this one, I'm I'm only about halfway through, but I a book that you've gotta read, you're gonna love, Adam, is Creative Capital. And this is on George Doriot, the the general, who is
Speaker 3:And this is this was the the first investor Yeah. In digital. Yes. I gotcha.
Speaker 1:This guy is amazing. Amazing. I mean, this is an amazing life that and, just exceptional. Really fascinating person. And one of the true pioneers of venture venture capital.
Speaker 1:I mean, arguably, the pioneer of venture capital and really, really, really fascinating person, and I'm dying to know if our board member, Pierre Lamont, must have had had intersected with George D'Oreal at some point. So I'm kinda dying to ask him that trying to find the time the right time to ask him that question. But I it's a really good book. So there we go. Awesome.
Speaker 3:Yeah. And and what about the, Deck is Dead Long Live Deck?
Speaker 1:Oh, yeah. The Deck is Dead Long Live Deck. I didn't mention yeah. Then that was kind of the first one I read. Yes.
Speaker 1:And then we got Deck is Dead Long Live as well. So by Edgar Schein, which is interesting because he was something who someone who worked with the company. So did you read that one? I don't know if you
Speaker 3:Yeah. That's the one I read.
Speaker 9:Yes.
Speaker 3:It was interesting. I I feel like it would it's like the, if I got, you know, you and a couple other buddies together and we wrote the history of Sun, I think it would sound a little bit like that.
Speaker 1:I think you're right. So okay. Yeah. So what do you mean by that? I agree.
Speaker 1:Wait. What do you mean?
Speaker 3:What what I what I mean is it I mean, it felt I mean, maybe maybe I shouldn't include you and my other buddies who aren't very, like, who aren't great writers maybe, but, like, it was it was okay written, but but, it was, you know, each chapter had, like, a slightly different ax to grind or a slightly different, perspective. And and clearly, some of them were stuck on on the, like, deck was great and, like, we crushed it and it was everybody else who was wrong and all of our you know, it it felt like early chapters were sort of making the excuses for failure. In other words, maybe written by folks who are too close to both in terms of working there around it and too close in time to be really objective.
Speaker 1:Yeah. I I agree with that. I also like did you read Gordon Bell? So Gordon Bell has an appendix in that book
Speaker 3:where No I did not check out the appendix.
Speaker 1:So I thought that was actually better than the book, because Wow. The the well no because one of the things that he says that I really disagree with is that Deck failed because it didn't have the business gene. And I'm like, what does that mean? This, like, this company is
Speaker 3:bootstrapped in 1957. Yeah. I mean, that was thematic throughout the book. And you're and you're right. Sort of like, I don't know.
Speaker 3:Antithetical with everything else.
Speaker 1:Yeah. I mean, I just don't know how you can accuse a company that has bootstrapped itself of not having of, like, not knowing how to run a business. It's like you can say a lot of things about that, but, like, I I don't know. It feels like and and Gord Bell also calls that out. You're like, I don't know he's he said that didn't have the business.
Speaker 1:Like, that is that that was not the problem. Yeah. So I so where do we wanna, again, I'd be curious if I did other folks wanna jump in here with with with their their memories of deck? Like, why do we talk about it? Why is it interesting?
Speaker 1:I certainly have got my well, some observations that I would like your perspective on, Adam, in terms of, like, why I think and other people's perspectives on in terms of why I think it's interesting. Something about Olsen and then Olsen has got like, there's a lot to be said that's positive about Olsen. He's clearly like a, he especially in contrast to kind of today where there's so much distrust in executive leadership, He is clearly he is someone who, does not live lavishly, who is not committing you know, this is not committing transgressions. He's not, you know, closing Bitcoin exchanges or what have you. I mean, he he he he's someone who who you would you would call decent clearly, and and the way that the kind of the corporation's being run.
Speaker 1:That said though, he there and I don't know how much this came through in deck. It's dead long with deck. But there is this very curious kind of juxtaposition between autonomy and autocracy in deck.
Speaker 3:Do you
Speaker 1:know what I mean?
Speaker 3:Yes. No. Absolutely. And and this is part of what felt very familiar with Sun. But in terms of Olsen's failing, it felt like it it sort of worked until it didn't, which is sort of a fast out statement.
Speaker 3:But what I mean is, like, there there came a certain scale or perhaps a certain kind of professional manager where the do the right thing started started being inwardly focused rather than, like, holistic.
Speaker 1:Totally. And in particular, one of the things that Olson likes to do, it that I think is nuts is he likes to have internal competition. He likes to have different groups competing with one another inside the company. And Yeah. I think that is, I I think that's terrible.
Speaker 1:I I I really do. I think that that is really, really, really toxic. I don't think that that works. Again, curious if other people got perspectives on where that has worked. But I don't I mean I
Speaker 3:mean, that's that's like the AWS model. Right? I mean, the the very overtly.
Speaker 1:The so does Amazon I don't know how much does is that true? Does Amazon have a lot of internal competition?
Speaker 3:So I obviously, I don't know because of their bit inside, but it does seem like even within their service offerings, I mean, you see competing and overlapping technologies. And you know, that I I think I've mentioned that the Everything Store, the biography of Jeff Bezos makes it seem like that is, like, a a big chunk of what drove his management philosophy.
Speaker 1:Yeah. That's interesting. I I don't think that works. I think that that ends up creating a bunch of internal strife. You create you're kinda constantly creating losers in your company.
Speaker 1:You're creating these battles, and you you shouldn't assume that the best thing is gonna win that battle. And I don't know. It feels very, having worked briefly at a company that really believed strongly in that kind of internal competition. I just saw how corrosive it was, and it didn't it it it kinda created a bunch of diverse incentives.
Speaker 2:I don't think it actually worked.
Speaker 3:Yeah. The I mean, the description of deck was they kept on bringing up the term family and this notion of, lifetime employment. Yeah. And but it it sounds like, you know, clearly, he he encouraged, lively discussion and heated discussion. But maybe, you know, if it rests on that foundation of, like, family and trust, and it's like, when those are drawn, I'm gonna have a job in a a year or 2 or 6 or 10 or whatever.
Speaker 3:Yeah. And maybe as that starts to erode, then also, like, like, the internal competition means, like, you know, if I lose, like, the con there are more significant consequences for me or that that trust breaks down.
Speaker 1:Yeah. That's interesting. So and and kind of my view is that that the I think you're right. And I think that that's how he kinda pulls it off with this, like, you know, win, lose, or draw, you're gonna have a job. But I also feel that, like, man, to have a project and another a couple people here have had this happen in their careers, but or maybe many, maybe all, to have a project that you've really poured your emotional energy into get canceled and kind of lose at the hands of a rival technology internally.
Speaker 1:You're like, I don't care if I got a job. Like, that's, you know, you've taken so much of what's been meaningful for me for some period of time and taken that away. It's like, I don't I don't want the job that's left here. And and people are leaving deck, like, relatively early on, you know. I mean, so Harlan Anderson, it leaves in, I wanna say, 66.
Speaker 1:So he leaves he is more or less managed out. He is saddled with the the PDP 6. One thing I wanna say as long as also, as long as we're talking about mistakes that I have made in talks, I I guess we're not talking about that, but maybe I'm gonna mention that. This one is because this is also this has been haunting me as I've, like I just feel like learned a lot more about deck. Like, I don't do the PDP machines blend together for you at all, Kev?
Speaker 3:Absolutely. As a much less attentive scholar to computer history, absolutely.
Speaker 1:And I don't know, like, p to p 8, p to p eleven, p to p ten. I don't know. I mean, this this all seems like, you know, p to p one. I don't know. P p p 3
Speaker 4:Naming was not a strong point.
Speaker 1:Well, Ed, yeah, right now. The thing is sort of also, Ed, in particular, you know, I gave this talk actually, the last talk that I gave in person was Adam, the talk that I gave about oxide at Stanford.
Speaker 4:Alright.
Speaker 1:That was that was, like, right before the shot it was the end of February 2020. Yeah. And the, actually, this is so they there's a lot that happened in that talk. They invited me down to give this talk, and then I got dry mouth in the talk. Have you ever had this happen to you?
Speaker 3:No. But now looking back, I do remember you just, like, desperate for every bottle of water in that room.
Speaker 1:I was so I told them before the beginning of the talk, like, I'm I really need water before I I talk, and they're like, we don't have time. You have to start. I'm like, I need water. Like, gotta go. Like, I need water.
Speaker 1:I need water. I need water. I need water. So I'm like, alright. I guess I'm gonna do the best I can.
Speaker 1:So I started that talk, like, desperate for water. And at some point in the talk, I call out for water. I'm basically, like, if someone does give me water, I'm I had not had that happen before, but I'm, like, I'm not gonna be able to, like, open my mouth. So in that talk, I talked about the pdp 11 pdp 1170 and and and how they have these, like, Raspberry Pi emulators. It was pretty cool.
Speaker 1:And someone was asking, like, well, what operating system do you run on that? And I'm like, I don't know. Like, tops 10 or whatever. I don't know. I I volunteered tops 10 as an operating system that one might run on the PDP 11, which I now realize is a, like, major faux pas.
Speaker 3:It I know.
Speaker 2:This this is absolutely mind blowing. You didn't you didn't say unit?
Speaker 1:Like, you know, I was just you know, I mean, obviously, like, units would be I but I was you know, I was just trying to, like, be colorful, I guess. I don't know. Yeah. I mean, obviously, it's all like I had you know, the I mean, obviously, units ran on the PDPLF 70, but the I was just kinda throwing it. And I so I say, I don't know, like, top 10.
Speaker 1:And the audience, which is supposedly a Stanford class, but basically consists of retirees that are local to Palo Alto that just kind of, like the the audience absolutely jumped on me. It reminded me of when I was with a bunch of 1st graders and suggested that a dolphin was a fish accidentally, and then tried to take it back and, like, all the first graders are like,
Speaker 3:a dolphin's not a fish.
Speaker 4:It's a thing load.
Speaker 3:You lost them completely.
Speaker 1:I lost them completely. It's just like, who who is this jackass? And you know that a dolphin thinks that dolphin's a fish. And the I feel like it's the same way. So I'm like, I don't know.
Speaker 1:Top 10. And literally, like, the audience, like, tops 10 never ran on a 10th 11th. And at the time, I'm like, what is their problem? And now I'm like, no. No.
Speaker 1:They were right. I was wrong. I should not have been suggesting that the clearly, like, because the p d p 10 is a, like, a totally different machine in every capacity from the p
Speaker 2:p Oh, yeah. I mean,
Speaker 4:you know p p 10 p p 10 was a mainframe.
Speaker 1:Right. Which which I now appreciate. And so, Tom, did you work on a PDP 10?
Speaker 4:Only very briefly. Not I was mostly PDP 11.
Speaker 1:Because the PDP 10 plays a pretty interesting role in history that I guess I didn't fully appreciate. I mean, this is basically one of the very early time sharing
Speaker 2:No. Well, it it it's a little more complex than that. The the the p d p 10 was the follow on to
Speaker 1:the p d p 6, which
Speaker 2:makes no sense. Because, like, the PDP 8, for example, is completely unrelated. Right?
Speaker 1:And the PDP 6 would follow on to the PDP 3, I'd like to say. Well, you know, the
Speaker 2:the PDP 6 was a follow on probably to the PDP 1. PDP 3. I I believe you since you've just read the books. But, like, the critical thing about the p two p ten was that that was, like, kind of the machine of the ARPANET early on. And and and a lot of the early Internet work was doing that machine.
Speaker 3:It's like Yeah.
Speaker 6:That's it it it Yeah.
Speaker 4:And and it yeah. In the late seventies, you were a nobody in terms of computer science if you didn't have a PDP 10, which was the problem at Princeton. We didn't have a PDP 10.
Speaker 1:And so what did you have in lieu of PDP 10, Tom?
Speaker 4:Well, we had the little PDP 11, which was a small machine that few people cared about, and we had the big old IBM systems in the computer center.
Speaker 1:That yeah. So that's that is really interesting. And the Tom, if I recall correctly, the p
Speaker 4:d And no and no ARPANET.
Speaker 1:It didn't know ARPANET. Interesting. And then the the folks at Parq coveted at 10 didn't have it, and Chuck Thacker built his that was the maxi. Right? It was not like a PDP 10 I mean, back
Speaker 6:off Right.
Speaker 4:It was a p PDP 10 clone, and Stanford was a huge p d p 10 hotbed.
Speaker 2:A a a lot of a lot of machines that we think of as being sort of weird these days, like the early, you know, park machines and early list machines and so forth, a lot of those were almost directly inspired by the p
Speaker 4:to p 10.
Speaker 1:Yeah. That's interesting. And so the p p ten is is killed very controversially inside a deck in, I think, 1983. And that really rubbed a lot of people. Apparently, rubbed a lot of people the wrong way.
Speaker 1:I don't know, Tom, if you had any visibility to when that was happening. I guess at that point, I mean, you're already kind of starting at Sun.
Speaker 4:Yeah. Well, I mean, they they were pushing the backs hard by that point. So there were just too many too many things to keep alive.
Speaker 1:Right. And I think which kind of highlights, you know, the the all of this this internal competition that they had for for and the the the kind of the p d v six even was kinda thought to be a dead end And the part of the the way they forced Anderson out, Harlan Anderson out, is by saddling him with the p by having him run the PDP 6, which everyone else, like, well, this is a project that's not gonna go anywhere, but ultimately became the the the much more successful PDP 10. And, Dan, did I mean, you never obviously work on a p to p 10. This is just I did.
Speaker 6:Good or not.
Speaker 3:How? I
Speaker 2:I I honest to god, I did. I mean, I I didn't do a lot with it. I was mostly famous for kinda getting inside of the system cabinets when when folks were out of the room. And then when they came back in the machine room, I would jump out and scare them. But we we yeah.
Speaker 2:We still we still had a PDP 10 that that the system manager had gone to absolutely this we actually had a k a one or k a tens. Excuse me. K a 10. And, the system manager had gone to absolutely heroic lengths to keep this thing running. And university management was like, this thing's a power hog.
Speaker 2:We wanna get rid of it. And eventually, the paper tape drive that it booted from died. And so Brian, the assistant manager built and, you know, he was like, we need to fix the paper tape drive on the p d p 10. And they were just like, no. We're gonna finally get rid of this thing.
Speaker 2:And so this guy, who was actually an actual engineer, sat down and built a compatible interface that talked to a serial port on one end and emulated a paper tape on the other. And then he wrote a daemon that ran on under VMS on one of the VAXs that when the p
Speaker 9:d p 10 booted up, you know, the the sort of carrier sense
Speaker 2:line on the serial port would go high, and he would just sort of squirt the bootstrap over it. That's true. He continued to be able to boot the p d p 10, and the and the the administration was not immune.
Speaker 1:Yeah. How much virtualization of paper tape exists out there? That's gonna be that's gonna be somewhat unusual. That's amazing. Where was this?
Speaker 2:That was at Penn State University. I was in high school at the time, and I was kinda, you know, stealing some computer time from the local labs. And in return for doing some system administration work, they gave me accounts kind of everywhere. And the the the the PDP 10 was esoteric, especially running tops 20. If you could find a deck system 20, which is running top I'm I'm sorry.
Speaker 2:The PDP 10 running tops 10 was very esoteric. If you could find a deck system 20 running tops 20, that's a much more reasonable computer. And that that almost feels very familiar when you log on to it.
Speaker 1:And then so could you tell me more about this dream like being sent in for You know,
Speaker 2:I was, like, 16 at the time, and I was a lot skinnier. And these were, you know, these full height, like, 6 foot tall cabinets. And if I kinda just lounge a little bit, I could basically stand inside of the thing. And some of them had had all the electronics removed. The cabinet itself was still there because nobody wants to move these racks around because they're heavy and stuff.
Speaker 9:But, like, yeah, I mean, you could open up
Speaker 2:the it was an interesting machine because you could open up the system cabinets and, you know, it was
Speaker 9:it was all, like, discreet wiring, especially on a Ka.
Speaker 2:And you could, you know, log an oscilloscope into one of these things or apply a DMM and actually trace, like, what the RAM was doing or, you know, what the CPU board was doing or whatever. It was it was all very, like, accessible in that sense.
Speaker 1:Yeah. It's amazing. But the world was such that low speed world. And so the I mean, this is but this is an old machine, obviously, by the time you're dealing with
Speaker 2:you're Oh, yeah.
Speaker 4:10 years old. Yeah. Yeah.
Speaker 2:Oh, oh, oh, more than that.
Speaker 9:I mean, this machine was installed in, like, 1968 or something.
Speaker 1:Oh, wow.
Speaker 2:And, you know, at at this point, it's the you know, I was in high school, so it was sometime in
Speaker 9:the early nineties, probably, like, 93, 94.
Speaker 2:So, I mean, this machine's been in service
Speaker 9:for, like, 25 years. And, you
Speaker 2:know, very,
Speaker 9:very few people are still using it, which was another reason the administration really want to kinda get rid of it. But, you know, it was just like there was a small user base, and the system manager was really into it.
Speaker 2:I mean, he really liked the p d p 10. He thought it was a great architecture. And, you know, he would just go to these heroic lengths.
Speaker 9:You know, it was like a ball put burnout on the on the control panel or something, and you'd go in and replace it with an LED type thing. And so as sort of physical components on
Speaker 2:the machine were breaking down, they were being replaced with solid state electronics and and, you know, that kind of thing.
Speaker 1:And so and are they running tops 10 on there?
Speaker 2:That ran tops 10. Yes.
Speaker 1:Yeah. And tops 10 is so you used I I mean, I never used tops 10. But see, this is, like, one of the first, I believe, operating systems to really feature OS based virtualization and other things.
Speaker 2:I don't know that I would describe it as featuring OS virtualization. There so TOPS 20 included them. There was, like, some weird thing where you could basically boot a That's
Speaker 1:that that's what I'm thinking of. Right? That TOPS 20 could TOPS 20 run TOPS 10?
Speaker 2:Kind of. There was like an there was like a TOPS 10 emulator that would run under TOPS 20, if I'm remembering correctly. It's funny. I have tops 20 running here at home. I should log in and just see if I could
Speaker 3:As one does. Yeah.
Speaker 1:As one does. Yeah. And and then but, Tom, had you used so so you did you you coveted the p to p 10 but didn't have it. And had you used the, like, the PDP 11 and its kind of successors or or any of the VAX machines? Or
Speaker 4:Yeah. So, I I got into Unix on the PDP 11 45. So that was 1975. And then I was actually the sysadmin for that machine for a couple couple of years. And any So it it was a very sweet machine.
Speaker 1:Yeah. I was gonna say, like, yeah, fond feelings of fond memories of that machine.
Speaker 4:Yep. Yep. And I actually like the control the the panel colors even better than the 11 70.
Speaker 1:Oh, wow. Interesting. Yeah. The the thing that does seem of Ken Olsen as an engineer, the thing that seems to and I I don't know. I can't remember how much this this came through in the deck is in the little deck, but it's certainly in the entrepreneur and these other books.
Speaker 1:Like, Ken Olsen is a packaging engineer, kind of this is his background, and it really does show in the the kind of the the product that, and just the the packaging of the product is really seems really good, durable.
Speaker 3:Yeah. Dev Alright. Did did not alluded to his engineering interest in Acumen, but not like that aspect of it.
Speaker 4:One one of the interesting things is they I don't know when they started, but the PGP 11 used the standard 19 inch relay rack, right, which we're all familiar with today. But I'm not sure when computer is she using that.
Speaker 1:Is that That's a good question. That. Yeah. That
Speaker 4:was a telephone company thing?
Speaker 1:Yeah. Right. The kind of the first computer to be rack mounted. Interesting. Yeah.
Speaker 1:When and certainly, the just the the way, you know, the and, you know, we've got a VT 100 in the office. I I need to think of that. A VT 100 is just a beautiful piece of simple machinery. I I I I'm sure they are,
Speaker 4:I What operating system does it run?
Speaker 1:That is a good question. I that is you know, that's a question for mister Clu. Dan, I assume that you've got a couple of of BT terminals.
Speaker 2:Oh, yeah. Yeah.
Speaker 3:I'm I'm just I
Speaker 2:feel slightly embarrassed. It's like, I swear I'm not a hoarder, but I I like, I definitely have a handful of VACs just down in the basement and some BT terminals. They don't really run an operating system. I mean, there there was a
Speaker 4:Oh, no.
Speaker 2:Yeah. I mean, they they basically, it's microcontroller code. There's there's certainly a microcontroller inside of the thing.
Speaker 1:But that's not configurable or upgradeable. You have to you only run one thing. Right? So I Yeah.
Speaker 6:Yeah. I mean, like like like
Speaker 2:it boots off of a ROM and basically, you know, takes characters from a couple of different sources and and, you know, does something with them. It's it's a very, very simple machine.
Speaker 4:Yeah. That that was the later I mean, the the early ones, like the b t fifty, that was hardware. There was no none of that microcode stuff.
Speaker 1:Right. So and, Dan, do you have any I mean, maybe fast forwarding a bit to as things begin to crumple inside the deck. Do you have any of the deck PCs?
Speaker 2:I do not. No. That was I I never found that particularly interesting. I I think by the time I was by the time Deck was making really nice PCs, which they did kinda towards the end and and sort of the, you know, the the nineties. I was much more interested in machines that could run VMS or Unix.
Speaker 2:And, like, you know, to say that you couldn't run Unix on a deck PC is a little bit of misnomer because you couldn't sell for EBSD or Linux or whatever on one. But, like, you know, Alteryx on the back or or sorry. On the on the the the deck station 5,000,003,000 model machines was actually pretty reasonable. And o OSF 1 on the alpha was my favorite version of UNIX for quite a while.
Speaker 1:And the so in o s f one is as I understand it, it's kind of Olson and Dex response to Sun, an attempt to I mean, I think I didn't quite realize that the degree to which Deck was a belligerent in the UNIX wars.
Speaker 6:Oh, got you.
Speaker 3:Oh, got you.
Speaker 1:I mean, look, here's
Speaker 2:the thing. Sun and AT and T get together and do UNIX International, And everybody else who was shipping UNIX at the time was like, oh, shit. Like, what do we do? I mean, this is it. Like, they're just gonna take over the market.
Speaker 2:And so DEC and STI and HP and IBM, and I'm I'm assuming a few other companies that nobody remembers anymore, all got together and created the Open Software Foundation.
Speaker 3:And, you know
Speaker 4:And and OSF also also stood for oppose Sun Forever.
Speaker 1:Wow. God.
Speaker 2:The yeah. No. But, like, OSF 1 was pretty cool. It the the thing is, though, it it was gonna be this thing where, like, I think HP was gonna give the Windows system, and they did. And, DEC was gonna supply the kernel that was gonna be mock based or or CMU was gonna do that, I guess.
Speaker 2:And, like, so each company was gonna kinda give their own little thing. Mhmm. And then once they had a system that, you know, nominally worked, like, Deck was like, alright. We're shipping this on alpha, and everybody else was like, cool. We're we're we're not.
Speaker 2:You know, except for IBM. IBM actually shipped OSF 1 on the mainframe.
Speaker 1:So was was CDE on I mean, the interface by HP, I mean, that's gotta be the common desktop environment. Right? Is that not
Speaker 2:a Yeah. So HP had something that I think was called HP View. That was the predecessor of CDE. And that was kinda what they threw into OSF to to, you know, get to play in the party. But and, you know, it's it's kind of an interesting question.
Speaker 2:I mean, why didn't IBM, for example, dump AIX? AIX was was the weirdest and, frankly, it was kind of the least friendly version of UNIX out there. You know, and OSF 1 was was just a much nicer operating system. And hockey pucks was terrible. Right.
Speaker 3:Well, I
Speaker 1:guess that's why they put them on the HP's like, hey. What do we contribute? Like, hey. How are you how about the user interface? How about you how do you do that?
Speaker 1:And the the the UNIX wars, I feel, ultimately end or at least change tenor with the once the BSD case is settled and the kind of the rise of Linux and the open source, Unisease is kinda my readout. I mean, the Unisease, I feel kinda predate me, but the I I feel like the big thing that Deck seemed and I'd actually tell him your take on this, but you are in the Denyer as well. But the it feels like what Deck really missed was the rise of open systems. And, similarly, I think so they understood, like, great engineered well engineered systems, but did not see the value in an open standard.
Speaker 3:Well, they just very just the PC. I mean, among other things as, like, I think, Olson's line was that, he would have fired the engineer who had designed it.
Speaker 1:Which is true. I mean, in that but, like, IBM would have if IBM had known and there's like a weird, this this kind of weird circular dependency that IBM had known how successful the PC was going to be, they would have put different people in charge of it, and they would have spent much longer on it, and it would not have been successful.
Speaker 2:I I mean, like, tech certainly had some of that. Right? I mean but, you know, like, the the tech is interesting because not only did they hate the PC, they hated UNIX. Olson hated UNIX. He there there's this great quote in one of the old iterations of the fortune file where he's talking about sort of UNIX documentation.
Speaker 2:And he's like, Unix. You know, it's this toy system, and look at these 2 little skinny books. And, yeah, they tell you everything you need to know to program the system. But, like, when you're ready to do real engineering, you're gonna come over to VMS and, you know, it that that's when it requires, like, a, you know, 2 beefy dudes to carry around this, like, 6 foot bookcase full of our manual set. Just look at our documentation.
Speaker 2:And the UNIX people, I think, sort of wore that as badge of pride. They were like, my god. I mean, your system is so complex. It requires a bookshelf to document. You know, ours is very simple, and you have these 2 little sort of thin volumes.
Speaker 2:But, like Olson, he never saw the value in that, and he's I think you I think you hit it on on the head. He really did not see the value of open systems at all until it was just fabulously too late.
Speaker 1:Fabulously too late. And I and and the the analog to Sun is, Sun, born on open systems, saw the value of open systems, but Sun didn't really see the value of open source until it was too late,
Speaker 2:I think. Someone also you know, let's let's be frank here. I think Sun chained themselves to Spark for just way too long. Oh, for sure. For sure.
Speaker 2:Yeah.
Speaker 1:Yeah. Yeah. And I think I mean, this is where, like, the analogs were interesting. It's like because Sun was kind of like for Sun, it shackled itself to Spark, kind of the way Deck shackled itself to VMS. And, x86 at Sun was kinda playing the role of UNIX at deck.
Speaker 1:Like, we know we have to take this seriously, but we can't take it too seriously because we feel it undermines who we are. So I I this is why there are all these interesting analogs even though the companies are separated in time by,
Speaker 3:you know, 10,
Speaker 1:15 years. Yeah. I mean, Son Son lost control of, 15 years. Yeah. I just
Speaker 4:I mean, Sun Sun lost control of the architecture and was eaten from the low end, which is kind of the same thing that happened to Deck.
Speaker 2:Absolutely. I mean, Deck Deck really wanted to sell you a a large form vax, you know, what what people used to call a real vax back in the day. And I remember looking at a Deck catalog one time, sort of 91, 92, something somewhere around there. And, you know, like like, the cost of a vaxx 9,000, that that was a 1,000,000 plus dollar machine. And I remember looking at sort of, you know, the performance numbers of the thing, and it's like, well, if I bought, you know, 3 MicroVac 31 100, like, that's the same aggregate performance.
Speaker 2:And and, oh, by the way, VMS has, like, best and arguably still, in many ways, has best in class clustering software. Like, why the hell would I buy the big machine when I can just throw another microvacs into the cluster and and and and get better performance? Like, it it made no sense.
Speaker 3:Yes.
Speaker 4:You
Speaker 2:know? So, like, just their product line was just totally out of whack with what the market was looking at. And they just they didn't notice. They just they they could not rationalize that.
Speaker 1:Well and and then I think also you because my view of deck, and I have said this has been, my view going in and this view has, I would say, been strengthened, having read now way too much on deck. Deck had way too many people from the get go, way too many people. And especially for a company that prides itself on lifetime employment. I mean, Deck had a 120,000 employees in 1989, and that is for, you know, an $11,000,000,000 top line. That is way too many people.
Speaker 1:And when you have that many people, it creates I think you create so many problems because now you have implicitly created a high margin company. You can't actually do low margin things or lower margin things because you just have too many mouths to feed. And I feel like they've also they created these kind of civil wars, and then they also invented matrix management. And but that's, like, a good thing. So sorry, Tom.
Speaker 1:Go ahead.
Speaker 4:Oh, I was just laughing at that.
Speaker 1:Yeah. I mean, so I mean, Olsen feels that he has this big breakthrough with the invention of matrix management. And this is part of the reason that that Anderson leads in 66 is because that he did. This is like madness where you have kind of everybody reporting into everybody else, but then you've got Olson. And I think his view, which is I don't think entirely cynical, is that this was kind of Olson's way of retaining power over the company was by not letting any any individual really be a logical successor.
Speaker 1:And by kind of assuring that there's this kind of constant feuding, I mean, the kind of the dark read on it is, this was his way of consolidating his own power. He definitely had a habit of anyone who he he thought was getting too powerful inside a deck, he would would be forced out because he didn't wanna have his authority be challenged. And yet for the sorry. Go ahead.
Speaker 4:Yeah. When in the early days of San, when people were coming from from digital, I think they had already lost a lot of respect for Olsen. He was already in the autocrat mode.
Speaker 1:Yeah. That's yeah. Can you describe a little bit more about that? Because there are some early folks at Sun that were critical at at deck. So Bernie Lacroote, is that right, Tom?
Speaker 4:Uh-huh. Yeah.
Speaker 1:And did you I mean I mean, the company must've been I mean, obviously, you you, I assume, knew him at Sun.
Speaker 4:Oh, yeah. Yeah. I mean, he he was, like, our 2nd VP of engineering. First one didn't last very long at all. But, Bernie lasted a long time.
Speaker 4:And I I don't I don't remember what he worked on at deck, but, he was he was the right mix of, you know, someone who knew what he was doing and able to handle a startup.
Speaker 1:So could you speak to that a little bit? Because he the loss of Lucreta's son is definitely viewed as one of these kind of pivotal moments inside of deck by everyone writing from the kind of the deck side viewed that was like a gutting loss and viewed him as really instrumental in helping to build the the culture of Sun or helping build Sun. Is that I mean, could you speak a little bit to kind of the way he ran engineering?
Speaker 4:Oh, boy. Not really. I just know that he he he ran it, which was given the growth rate, it was quite an accomplishment.
Speaker 1:Yeah. So
Speaker 4:it I mean, there's just so much shit going on. Yeah. Interesting. Everyone everyone respected Bernie.
Speaker 1:And he was able to balance must have been it must have been extraordinary manic in those days with with I mean, revenue was absolutely exploding.
Speaker 4:Yeah. Yeah. And with, you know, Bill Joy and Andy, you know, providing input frequently.
Speaker 1:Oh my god. My diet and pepper almost came out my nose.
Speaker 2:I'm I'm curious, if
Speaker 3:this idea of,
Speaker 2:one of the things I'm wondering if that
Speaker 3:sort of gives rise to,
Speaker 2:what it what I would call these sort of insurgency movements within DEC. Yeah. I'm thinking specifically of, like, your Mondo Stettner and the UNIX people. Because for instance, I guess it was gatekeeper. Was was that the machine?
Speaker 2:Anyway, one of the early machines at digital was, like, it was a super important Usenet and UUCP hub. Like, everything Yeah. Deck DeckFax.
Speaker 6:Yes. Thank you. Way back.
Speaker 4:Yeah. Bill Bill Shannon, who was number 11 at Sun, came straight from the UNIX group at DEC. And so they were definitely, troublesome groups for DEC. You know? They're, the Bell Labs bought a huge amount of machines to run UNIX, but other than that, everyone hate that deck hated UNIX.
Speaker 2:But there were, like, these this group of guys who I remember it I think it's in Salus' 25 years of UNIX history. There's this sort of thing where at some point some accounting person comes to Stettner and is like, hey. You spent, like, $10,000 on on phone calls to the West Coast last month. What what is this? And he just sort of said, like, oh, the computers are talking to each other or something.
Speaker 2:And first one Yeah.
Speaker 4:They're not
Speaker 6:Okay. And, like, the
Speaker 4:number I saw pretty recently was quarter of a $1,000,000. I don't know if that was a year or a month or whatever, but it was a lot of money.
Speaker 2:That's a pretty big matzo ball. And I I and I guess the point is, like, even within this kind of autocratic environment, you have these people who are like, yeah, we're not gonna cowcouch at the party line. We're gonna do our own thing because we think it's the right thing to do and they weren't snuffed out, you know. And and like that and then team 82 or whatever, that's a pretty big line item for a telephone bill.
Speaker 1:Yeah. And I do think that that is that Dan, I think it's a very good this is where you get to the autonomy portion of this. This is what what makes Dex so, I think, interesting and but also also himself a bit of a sphinx because you did have this autonomy and people were encouraged to do the right thing, it seems. And that certainly, I mean, I I feel that certainly Sun had that same culture that came directly from that or not. But where you definitely had peep the boots on ground were very much, in charge.
Speaker 1:I mean, I I had said for years that that Sun's political model was like Somalia. It's like feuding warlords, and it couldn't possibly be federalized or controlled because the warlords were actually in charge. And if you wanted to get a food convoy through Sun Microsystems, you had to make sure that it was the the the warlords you coordinated with. And it feels like that, deck had that that that kind of similar, arguably autonomy to a fault. I think the thing that that it was different is that there were more it feels like there were more rival efforts inside of Deck.
Speaker 1:And then also just a lot more people, just an absolute ton of people there.
Speaker 4:Yeah. I wanted to circle back, and another parallel between Deck and Sun was that they they both made the computers that computer people wanted to have.
Speaker 1:Right. Right?
Speaker 4:So they're addressing the technical markets, and there's no one in their right mind wanted would prefer an IBM to a to a PDP in the 7 days. Right? But everyone was forced to use IBM except for the lucky lucky few.
Speaker 1:And was that I mean, obviously, I can envision many ways in which that was reflected. But was that I mean, Tom, what were some of the the kind of the dimensions in which this is a computer for computer people?
Speaker 4:Why? A lot of there was a lot of openness. So, you know, the unit bus and stuff like that provided a lot of interesting hardware variety. There was a range of operating systems, fair amount of customization possible. I don't know.
Speaker 4:Just a lot of
Speaker 1:It also a
Speaker 4:lot of choice.
Speaker 2:We're also really high quality computers.
Speaker 10:Right.
Speaker 2:Yeah. Every time I touch this every time
Speaker 9:I I I still have an alpha
Speaker 2:down on my basement that I recently had to replace a fan on because the fan died. And, you know, this machine was built in, like, 1996 or something. Right? And it's, like, still going strong and still running VMS. It's it's kind of an impressive accomplishment in many ways.
Speaker 1:Well, and I think
Speaker 4:It's funny because because in my point of view, I you know, the the bar was set by by deck. That was what I expected. And then when the PC came along, it's like, holy shit. How do they how how are they getting away with this?
Speaker 1:They being, like, the PC market. I would like they're like, this is not
Speaker 4:Yeah. Right.
Speaker 1:Well, and you can see why it was really hard for for Deck, to figure out how to enter that market, and obviously, they entered it poorly. The thing that I think is interesting is they had 3 rival piece the efforts inside of deck and they they released them all on the same day.
Speaker 3:So on purpose or by accident?
Speaker 1:Apparently on
Speaker 4:purpose Probably to mollify all of these competing groups
Speaker 1:I think it was a bit to mollify the the the competing group the competing groups. And, the but it it it ultimately, there's there's a great line where, Olson called in Miller, Folsom, and Loveland. These are the 3 execs leading these 3 or 3 folks leading these 3 of our teams to discuss the 3 headed monster deck's low end strategy I'm quoting now from from the old master. He he said, you know, I'd make it easy on everybody if I just said one of these products will come out. But I'm not gonna do it because they do different things, and the market will figure out who is right.
Speaker 1:And, Ofer Miller says, the market figured out who is right, and it was IBM.
Speaker 3:You know, I think this is where, certainly, the analog with with Sun breaks down. Because I don't think we could have ever coordinated 3 different teams to launch on the same day.
Speaker 1:Right. Exactly. Well, I no. I I
Speaker 4:absolutely We did we did we did have 3 different architectures going for a while. We had
Speaker 1:the Yeah.
Speaker 10:Yeah. For sure.
Speaker 4:Sun threespark, Sun 3 86i.
Speaker 3:Yeah. I just mean you couldn't get those 3 executives into the in the room and decide and, like, agree on a date, and there was no leadership that would that would dictate it from on time.
Speaker 2:I I I kinda feel like another critical difference between Sun and Deck. It just has to be acknowledged. There's there was a coastal difference. I mean
Speaker 1:Yeah. Yeah.
Speaker 4:Yeah. That's a great point. No. Definitely.
Speaker 2:Deck Deck always struck me as sort of like the the nerdy engineer company. And and I I mean that, like, in a in a in a in an affectionate way. This isn't, you know, but, like, I think of deck as, like, guys with black rimmed glasses that are taped together kinda like mine. You know, maybe, like, button down shirts, but with, like, short sleeves and a skinny tie. You know, it's gone to that 19 early 19 sixties image of an engineer.
Speaker 2:And and Sun was more like, you know, no offense, Tom, but I like like, I know people were smoking weed in the parking lot. You can't lie.
Speaker 3:You know?
Speaker 2:You know? Or or somebody was dropping acid. It's not or, you know, I like that. It was it was much more of, like, the hot tub, it'd be, you know, kinda make It
Speaker 4:was definitely definitely the hot tub thing. This was this was early eighties, pre aids.
Speaker 2:Right. Right.
Speaker 4:There was a lot going on. Yeah.
Speaker 1:So okay. So the hot tub thing is not merely, like because certainly, Dan, I agree with you. When I think of, like, Silicon Valley in the early eighties, I feel like I'm in, like, Santa Cruz or Scotts Valley in a hot tub. And because the the, Stephen Levy's hackers talks about this. Right?
Speaker 1:Where the, the the, Sierra online folks were constantly having these, like, wild parties in a hot tub. So so, Tom, this actually happened. This is not this is not merely apocryphal. These hot tub parties
Speaker 3:Oh, yeah.
Speaker 2:Sale had a hot tub at Stanford.
Speaker 1:I mean, it was like
Speaker 2:they actually had it in the AI lab. Yeah. That's not apocryphal.
Speaker 1:Alright. So, video is it sorry, Tom. Go ahead.
Speaker 4:I I I saw that hot tub by one visit to the AI lab.
Speaker 2:By the way, if you want some colorful history of of the PDP culture around that time, log in to a deck system 20 and run the b board command, but print out all of the entries, including the very first one. And you will get some flavor. This is the goodbye message from the sale AI system. When they shut down the PDP pen, somebody wrote this, you know, rather long, thing that I guess was posted to Usenet or or something, but it went into the v board file. V board was like a primitive bulletin board that kinda was built into the top twenty.
Speaker 2:I mean, it's all built with this type stuff. But,
Speaker 4:well, it turns out, you know, they they save the backup tapes for the sales system.
Speaker 2:Yes.
Speaker 4:And somebody has somebody has restored everything, and it's online. But, you you know, the private stuff, you still have to be the right user to get it. But Oh, wow. One of the really one of the really interesting things is that Andy Bechtelsheim did all the design for the early sun stuff on that system. So you can find the design files, you know, the wire wireless and all the stuff for the sun ones, sun twos, even some sun 3 stuff before they they got off of that system.
Speaker 1:And that is a PDP 10?
Speaker 4:PDP 10. Yeah.
Speaker 1:Wow. That's amazing. Matt, you had your your hand up.
Speaker 10:Yes. Yes. So, I'm I'm young enough that by the time I was aware of DEC, it was well and truly in decline in the mid nineties. But, I, you guys were talking about, DECK, you know, various DECK hardware that you've owned. And I, briefly owned a piece of deck hardware that I bought off of eBay about 4 years ago, and that was the DECtalk Express, speech synthesizer.
Speaker 10:So this was, this was basically a a portable box that could, you know, run off of a battery and and, hook up to a computer through a serial port and it was running the DECtalk software on
Speaker 3:a on
Speaker 10:a 386, from what I've read. But the thing that I thought was interesting about this and I bought this hardware for the specific, purpose of I I wanted to donate it to the Living Computers Museum in in Seattle because this was while I was at Microsoft. And I I thought this museum was cool, but it should at least 1 or 2 of the computers there should be accessible. So anyway, like accessible for blind people. But, anyway, the thing I found interesting about this hardware was the serial port connection was, that the port on the DECtalk Express box was not a standard RS 232 port.
Speaker 10:MMJ. Right. Yeah. MMJ, the modified modular jack. Is that is that what it stood for?
Speaker 4:I believe so. Yeah.
Speaker 10:It was like I think it was, like, similar to an r j 40
Speaker 2:5. Rjrj11 is probably the closest. But, yeah, it it has the offset clip, which is off So
Speaker 10:I had to buy another adapter off of eBay. Right. But I thought that this is, weirdly proprietary, I guess.
Speaker 1:This is weirdly proprietary. I okay. So I have no idea about this thing. I'm on the Wikipedia page now for the modified mod tour jack. And honestly, Adam, have you seen this thing?
Speaker 1:If I saw this jack on the back of the machine, I would assume that you are pranking me. I'd be like, what the hell is this?
Speaker 4:Are you talking about Sir serial ports have a sad history. I I used to complain at Sun that every hardware engineer had to design his own serial port connector. Yeah.
Speaker 1:Well, it it it does feel like it's the the apple of its day in terms of, like, but I need another dongle now? It's like, okay. Jesus. Perfect.
Speaker 2:Like, in fairness, Deck used that for everything. So, like, the later model VT terminals all have an MFJ connector coming out of the coming out of the back.
Speaker 3:Yes. I have definitely bumped into this in, in Brian, the in the Brown CS computer lab. The the terminals there had this weird thing.
Speaker 1:You're right. Right. Right. Right. Right.
Speaker 1:Little did I know that I was,
Speaker 4:yeah. It's kinda like the the Cisco standard for serial ports on a rj45,
Speaker 1:which is, you
Speaker 4:know, propagated around lots of places.
Speaker 10:And by the way, yes, the DECtalk express that I bought off of eBay was functional. I serial adapter. And some kind some serial adapter. And some kind somebody wrote an add on for the NVDA Windows screen reader to make it work with Tech Talk Express. But, yes, it was functional.
Speaker 1:That's very cool. And actually, Matt, I was thinking about that as I would because I I remember you had mentioned that in the the when we did the space on accessibility. And and I I I mean, I think you had said that, you know, this was way ahead of its time and and really remarkable engineering, and they've done a lot with a very small amount of of
Speaker 10:RAM and other resources. Well, yes. Original DECtalk came out in 84, and I I heard from one of my colleagues Microsoft that the original DECtalk was using a 68,000 processor, and then DECtalk Express was much later. I think it was late eighties or early nineties and was specifically designed for accessibility applications, you know, to be hooked up to, like, DOS PCs.
Speaker 4:Interesting.
Speaker 1:Yeah. That would but another degree in which the deck was I mean, again, like like so many times was was way way ahead of its time.
Speaker 2:Speaking of sort of text to speech, peripherals, Lauren De Cherry recently passed away. She had been a researcher at, Bell Labs working with the UNIX group in 1 127. And she had written a text to speech program that, plugged into some sort of speech synthesizer that they had, and I don't recall exactly what model it was. But, again, that was hooked up to a DEC machine. You know, all of that work was originally done on Digital Equipment Corporation Computers.
Speaker 2:And when are the one of the sort of funny stories about this, and this is secondhand, but I don't think I'm betraying any competencies here, is that somebody went into the lab one night very late, and they heard this sort of computer y voice coming from down the hall saying, fuck. Fuck. Fuck. And they went into this office and found Ken Thompson and Dennis Ritchie basically teaching the computer how to swear. It's like, what are you gonna do with it?
Speaker 2:And there was just had this string of obscenities coming out of the thing. So, you know This
Speaker 1:is like the Microsoft Tay circa 19 seventies era Microsoft Tay. Right. With with with the role of 4 chan being played strangely by by Dennis Ritchie and Ken Thompson.
Speaker 6:Yeah. I don't know. That might be a little bit of a stretch, but, you know, maybe not too much.
Speaker 10:Oh, so one more thing about deck talk, if I may. So when so you guys said Deck got sold to Compaq in in 98. Right? So when Deck got sold, Deck the ownership of Decktalk then changed hands a few times over the past few years. But I don't know if this is coincidence or not, but Decktalk, the the the quality of DeckTalk really suffered, starting right around the time that Deck got sold.
Speaker 1:It's amazing.
Speaker 10:The last good version of Deck Talk, in my opinion, was version 4.50, which is late nineties. And, in fact, when, when I was when I was working on Windows accessibility soft well, at one point I was working on an on a product that ran on both Windows and Linux around 2004. And, our primary speech synthesizer was DECtalk. And, going back and forth with the then current owners of the technology, we were able to dig up version 4.50 for Windows, but we could not get a good version for Linux because by the time it got ported to Linux, it had already gone to shit.
Speaker 1:Got yeah. Well, acquisitions will do that to you. Simeon, since you got your hand up, before you do, Tim, I saw you jumping in here. It's it's been a long time. How are you?
Speaker 7:Not bad. Can you hear me?
Speaker 1:I can hear you. Absolutely.
Speaker 7:Yeah. So I worked for DEC between 81 and 83, my first job out of university. And, you know, this was when the Vaxx tribe was winning the wars against the DEC System 10/20 tribe. It was it was it was, ugly. But I just wanted to tell you a funny war story, which I thought might amuse a couple of people.
Speaker 7:So I was working up in Canada, and so they took us on a tour of one of the factories where they built the PDP elevens. It was up in Ottawa. And so they had this big factory floor full of these wire wrap machines, and there were immigrant women. They were all immigrant women sitting there in front of these things, you know, putting a wire on 2 of the connectors on the back of the of the motherboard, by hand. 1111 1.
Speaker 7:And this is how the computers were being built in about 1982, fairly astonishing. But then, then they said, well, you should see how we do it in the future. So it went over to the place where it was being done automatically. And what they had there were these huge robot machines, and I think there are 6 of them. And each of them had 2 arms that held the motherboard, and then moved it around.
Speaker 7:And at the same time, 2 other arms came in and plunked a wire between 2 of the connections. And it was really pretty fast. Boing boing boing boing boing. So they were building those, you know, at great speed. And then after they said, would you like to meet the guy who wrote the code that does that?
Speaker 7:And we said, yeah. Sure. That'd be fun. And so they took us to this guy, and he was in the room off to the side, and and I and we said, well, how does that work? He said, it all runs on this PDPA here.
Speaker 7:And we were shocked. I mean, because the PDPA was not much of a computer. And we said, well, how can that possibly work? Oh, it's okay. We don't have any operating system to get on the way.
Speaker 1:Nice. Well, and the PDP 8, as I understand it, Tim, was used in a lot of these kind of industrial real time applications.
Speaker 7:It was a 12 bit computer. And, you know, and I can't imagine what the clock speeds were like, but you know, not fast. And, yeah, they sold tons of them. And they also had this really successful line of dedicated word processors at one point. And those were all PDPAs inside.
Speaker 7:And, you know, on the famous launch of the 3 DEC PC's in 82 or whatever that was, One of them was a PDP 8 word processor. Is
Speaker 3:this Deckmate, Tim?
Speaker 7:I believe that's correct. Right?
Speaker 3:Nice. Great. Yeah.
Speaker 1:Yeah. This is I wanted
Speaker 7:to tell the worst I wanted to tell the worst story.
Speaker 1:That's a great war story, Tim. And I believe it was the Deckmate 2 was the so that was the the kind of the 3, the the 3 rivals are the the Deckmate 2, the Pro 350, and then what is effectively a Deck PC, the Rainbow. Terribly, terribly named. And those are I I obviously, all 3 perished.
Speaker 4:But the guy the guy behind the rainbow was Barry James Folsom, who then, came over to Sun to do the 3 86 I.
Speaker 1:Oh, interesting.
Speaker 4:So he should've he should've learned his lesson, but, you know, we had to produce our own non PC compatible. So
Speaker 1:And and, Tim, does your story does actually because one of the things I've definitely wondered, especially, you know, is we've got I mean, obviously, we're making machines, and, I mean, it was really mesmerizing. When we were doing board bring up, they got the the the flying probe that test all these different connections and watching out things, seeing through all of this, you know, watching these pick and place machines. And you realize that, like, Moore's law has absolutely played a role in the way we manufacture computing. And we were able we were able to make computers better because computers themselves were getting better. I mean, there's there's a there's a very important role that computers were playing in the making of computers.
Speaker 1:And, Tim, like, it sounds like you're right on the co face of that,
Speaker 3:of the PDPA. Can't disagree.
Speaker 4:But going going back going back to the EPA thing here, there there's kind of a lost art of being able to write stand alone programs, which were fairly common back in the day. If you had some simple task to do, you know, who needed an operating system?
Speaker 1:Oh, well, you know, we we do this with bare metal ROS. We do this with you know, this the beauty of no standard in rust, right, is that you actually it's kind of bringing the lost stars back a little bit.
Speaker 4:Yeah. But, one of the one of the really beautiful books from DEC from the PDP 8 era I think it's called the Small Computer Handbook or or something like that. But but, basically, it starts with, here is the definition of a bit. It can be a 1 or a 0. And by the by the end of the book, you're writing a Fortran compiler that runs on the p d p a.
Speaker 1:That's right.
Speaker 3:And it
Speaker 4:just walk it just walks you all the way up the stack. It's amazing.
Speaker 1:That is right. Yeah.
Speaker 2:We would be remiss if we didn't mention the d d b 7 in this conversation since that was actually the first machine that ran what became UNIX And the first identifiable version of UNIX was written for the p d p seven and assembly language. And that is an interest. If you ever get a chance to use PDP 7 UNIX, it's actually kinda cool. Like, it it it's different, but you can recognize it.
Speaker 4:Yeah. I was amazed they got that up and running at at the Living Computer Museum.
Speaker 1:That is really cool. So and, Tom, the PDP 8 small computer handbook is online. Oh, cool. Yeah. It is it is really cool.
Speaker 1:This is really neat. I it's like you know, a question I definitely have is, like, if you were to teach a course on this today, I think it would it's still be very pedagogically valuable actually. Even though plenty of it is antiquarian, plenty of these abstractions are obviously remain the same.
Speaker 4:Yeah. Yeah. If if if you're the type of student who wants to learn some of the bottom up, it's it's amazing stuff. It is amazing stuff. But if you're a top down person, it's like, why are we doing
Speaker 1:this? That's right. The everyone in these photographs is dead. Simeon, you had your your hand up for a while with it. Sorry to sure.
Speaker 1:Sorry. Are you there? I'm not sure if Simeon is still there or not. May have given up.
Speaker 3:Can
Speaker 8:you hear me now?
Speaker 4:Okay. Yep.
Speaker 1:I can hear you.
Speaker 8:Yeah. I guess with the clock speeds now, you get that, you know, everything in analog problem is is, is is more relevant now than, you know, when when people were running things on on PDPs. 2 comments. So the one was you mentioned this internal competition in a company thing. I would say probably the modern case study might be looking at Cisco because they do a lot of that.
Speaker 8:Perhaps, one difference is they have business divisions that focus on specific markets, so they start out not necessarily competing. And then they try and attack each other's market, which is kinda weird. So, that was my comment. The thing that I wanted to ask is so my my, sort of, only proper experience with, I guess it would have been Compaq at that point, was, working on alpha machines. And, in fact, one of my first kind of student jobs at university was, together with a mate being a sysadmin for, at one of the very late Alpha machines at the point where, where, Compaq had struck a deal with Samsung to start manufacturing motherboards in, kind of, PC form factor, and you could buy, essentially a clone, like an Alpha clone.
Speaker 8:And there was a theoretical physicist at the Physics department that he wanted to run Mathematica, he wanted to teach a course with Mathematica, and he wanted a real machine, which meant a 64 bit machine. And so my mate and I, we were the Sys Admins. We ran an early Red Hat that was actually shipped, compiled for alpha on that machine. And I felt like they had a bunch of ingredients and they had good timing to really have a high performance processor, which ultimately failed, you know, comp you know, against X86. But it just seemed like a shame.
Speaker 8:Like, it ran open source software. It kinda had the PC model. It was high performance. It was 64 bit, and yet it failed. And I'm kind of curious if anyone has insights as to, you know, why that happened.
Speaker 8:What killed Alpha, you know, even though, DEC was pretty much dead at that point?
Speaker 3:Well, I
Speaker 1:think that that's probably that's probably a part of it. So I can tell you that a bunch of the alpha architects, the alpha folks came to Sun after alpha, so I actually worked somewhat closely with the guy who'd been the lead architect on alpha. And it could because he was the he was one of the the 3 architects on Millennium, inside of Sun. Very bad to name a when it was, like, 2003, and we were still working on Millennium. That's a bad sign.
Speaker 1:It's the danger of, like it's like maybe you should not code name your microprocessor after a year unless you are absolutely certain you're gonna stick your date. But the what he had told me, his perspective on why Alpha had failed was, Deck had very bad marketing. Thought
Speaker 3:Alpha
Speaker 1:had had a I
Speaker 4:mean, organically, not because of
Speaker 1:pretty bad it had I thought Alpha had had a I mean, organically, not because of pretty bad had a terrific brand, a brand that meant performance. I remember thinking, like, don't you think the problem was, like, that there were no apps? I mean, that feels like the absence of software really was more of a problem.
Speaker 2:I I don't think it was that. I mean, like, alpha fundamentally, alpha was too expensive. It it was a great architecture, but it cost it was it was, you know, double the cost of x86 or whatever, 4 times or however much more expensive it was. And, you know, the the corresponding increase in performance was not that high. You know, you were paying 4 times as much for a processor that was twice as fast.
Speaker 2:I mean, just buy 2x86 processors or hell by 4 of them, and you get an aggregate system that's 4 times faster. That I think had a lot to do with it. Alpha as an architecture did some weird stuff. I mean, it was effectively a word oriented machine machine with special instructions to access bytes, all that was byte addressed. And then there was, like, some weird stuff with the memory model.
Speaker 2:It was, like, Uber relaxed. If you read the RCU papers, it's the only one where in the read path you actually have to, like, do something. You know? So it was it was just a strange architecture, and and I think that had more to do with it. You know, like the Alpha was canceled for a reason.
Speaker 2:And and it wasn't that it was unsuccessful commercially. It was it was they had pushed the architecture about as far as they thought that they could reasonably push it.
Speaker 1:Yeah. The well, and and Alpha also served as the foundation for Piranha, which I thought was I I think is the 1st multi core design. It's really a very early one coming out of deck world. And I felt like that could have been a very I mean, obviously, it was the path that ultimately all of computing went, but DEC was definitely there first or very, very early with with Piranha. I think part of the reason that the I mean, I do think that the failure of DEC is really tied up in the failure of Alpha.
Speaker 1:I think by the time Alpha came out, Deck as a as a vessel was in deep trouble.
Speaker 2:I I would I would put the back 9,000 as as more of the root cause there.
Speaker 1:That's fair. Yeah. Okay. That's yeah. The backside 1,000 was also a, obviously, a complete disaster.
Speaker 1:Tom, what's your perspective?
Speaker 4:That was all after the time when I was paying attention to the deck, but but, the in I like to say there were 3 phases at Sun. There was beat beat Apollo, and we pulled that off. Then there was beat digital, and we pulled that off. And then there was complete chaos.
Speaker 3:And we
Speaker 9:pulled that off too.
Speaker 1:I I I guess I was just there. I think I came for complete chaos. I guess I showed up.
Speaker 4:Yep. Yep.
Speaker 1:The, I well, I actually think that there was a there was a beat IBM for sure inside of Sun.
Speaker 3:There there was also a beat Microsoft inside of Sun
Speaker 1:too. Yeah.
Speaker 4:Right. And and a beat beat Intel and a beat anyone and
Speaker 3:Hence, complete chaos because Right.
Speaker 4:It's
Speaker 3:the the shifting targets.
Speaker 4:I've I I posted that graphic a couple weeks ago. That was very hilarious. Which one was that? This, this it was, somebody did a cartoonish thing about all the directions Sun was trying to go and how screwed up things were. Yes.
Speaker 4:Yes. And, that's pretty funny.
Speaker 1:Yeah. And I think that, like, you know, we to to a certain degree, it's it's that that chaos is a luxury of having I mean, when you are a profitable company, you got the luxury of chaos and Deck had that luxury. When you when you're trying to beat Apollo, you just you don't you don't have I mean, you've got necessarily you've got a lot of focus. And
Speaker 4:And and and
Speaker 10:Sorry. And I can probably thank that for deck talk. The
Speaker 1:the right. Exactly. Absolutely. Then no question. And the the a lots of things inside of Sun too benefited from the chaos and the
Speaker 4:from from the autonomy.
Speaker 1:The the lots of great ideas came out because of that. So, like, there's there's something to be said for it. It's just that it it's, there's also obviously a a real danger. I think one of the things that is very clear is that Olsen needed to have left Deck earlier. I think that it was gonna be very hard for Deck to survive with Olsen in place.
Speaker 1:He'd just been wrong about too much that driven too many people out.
Speaker 3:Yeah. And And I think there was a bit of a a transition problem at Sun as well where, I'm not saying that Scott should have left early earlier necessarily, but clearly the transition. Didn't like take us where we needed to go. And it felt very familiar at at deck where he he pushed off. And as you say, he kind of consolidated power and made sure there was no clear number 2 until much, much, much too late.
Speaker 9:You know, I was out in the field then, and I think one thing that was really very very bad was out in the field sales organization. They were completely a 100% obsessed with with IBM, and quite convinced that if DEC just became more like IBM, then everything would be okay. And they kept bringing in these mid level doofuses from IBM and putting them in important jobs, and assuming things would fix themselves. Now at that point, you gotta realize that that IBM was, like, 10 times the size, and so the appraisal was, well, you know, IBM is growing at 1 deck per year. How can we win?
Speaker 9:Well, we gotta be more like them. And it just didn't end well.
Speaker 1:And that does not end well. And Tim, it goes to, like, the again, the the kind of the my, like, fundamental belief of more people, more problems. And Deck had way too many people. You need to have you should have a tenth the number of people that they had or or or a third, and force the company to focus a lot more. It it just felt like they were creating so many of these problems.
Speaker 1:And the idea that, like, headcount is a goal in and of itself, to me, is just anathema, which I believe I am pronouncing correctly, but now Patel
Speaker 4:Well, the one the other one of the, local symptoms of that internal competition stuff was the DEC Western Research Labs versus DEC SRC, 2 different research centers 2 miles apart, and little antipathy among the groups.
Speaker 1:It's okay. I didn't. So I don't think I realized that so deck world and deck circ were I I knew that they both existed. I don't think I realized that they both existed at the same time.
Speaker 4:Oh, yeah.
Speaker 1:Oh my god.
Speaker 4:And CERC CERC was the former Xerox spark people, but I think overall, I think WRL did a lot more interesting product stuff.
Speaker 1:Yeah. I wanna say that that I think Piranha actually came out of world. So, yeah, there's definitely some interesting stuff happening. And when they and I mean, like Sun, where you had different, you know, rival labs as well.
Speaker 9:Yeah. You know, they also had AltaVista. If they've been a little smarter, they could have been Google.
Speaker 1:Well, that's it. Yeah. I mean, it and they had so I mean, Tim, they had and, actually, I remember when I knew someone who had gone to Google very, very early.
Speaker 3:I I'd never heard of
Speaker 1:the company before, and was so it it Tim, it's kinda hard to remember. And that was, like, that was the late nineties. Right? I thought 98, 99, the search engine of choice was was AltaVista for everybody.
Speaker 11:I mean, not only not only did they have AltaVista, they had Jeff Dean and Sanjay at
Speaker 3:the same
Speaker 11:time. So, like, they had some of the core people at the heart of early Google, within the building, and they weren't able to, you know, effectively deploy them.
Speaker 1:Okay. So here's the question though for you. So for was deck ever going to make it? Because I kinda feel like the and I don't know, Tim, what you feel about Sun, but I I I kinda feel like, you know, you can talk about, like, how could you brighten the ship at deck? How could you brighten the ship at sun?
Speaker 1:But I also just kinda feel like, actually, the the ecology of Silicon Valley requires forest fire to have actual new new birth, and, actually, these companies had to burn to to to release their to release the Jeff Deans back into new enterprises where they could actually build something new.
Speaker 11:It's just so it's so hard to be able to navigate a technology change as fundamental as what Dick was going through. Right? Where they had to throw away that, you know, big metal time sharing business model that was the golden goose. They had to redirect funds away from that and pick what the next thing was. And there were a few people in the building who were trying to do that.
Speaker 11:They knew what they what the right thing was to to try and fund, but they just could not effectively, you know, shift the the ties of the company you were talking about, Bernie LeCruit earlier, and he has a oral history on the Computer History Museum's YouTube channel. And looking through the the transcript, there's some interesting stuff in there about his, you know, decision to leave deck and and join some.
Speaker 1:Oh, that's so just a quick pitch for those. I Adam, have you listened to any of these oral histories?
Speaker 3:Yeah. The one from Pia Laman and then I think 1 or 2 others as well.
Speaker 11:The Dave Cutler one the Dave Cutler 1 is a very good follow-up.
Speaker 1:The Dave Cutler 1 is a very good follow-up. It is a very good follow-up. It's a 2 parter. And my I I just had had that on whenever I was in the car, the Dave Cutter 1. And my kids just got used to, like, alright.
Speaker 1:If I dad's gonna give me a ride, I gotta, like, listen to this guy for 15 minutes talk about but it was it's really good. It it and those are, highly recommended. The the one with Pierre is terrific. I don't think Pierre realized it was gonna be publicly available, which makes it because I in my first meeting with Pierre, I referenced that. And he's like, how do you know that?
Speaker 1:I'm like, well, there was this oral history. He's like, did they make those available? And and I I think
Speaker 3:Oh, do you regret taking swings at all the people you did?
Speaker 1:Well and the thing that's funny is, like, this is literally the venture capitalist behind YouTube who is just surprised that it's online. And but I think that these oral histories are marvelous, and I I can't I missed the Bernie LeCoultre one. I, you know, can't wait to check that one out. I I wish they would turn them into, like, a podcast format because it'd be a lot easier to consume than YouTube. But these are really, must watch.
Speaker 1:I the Dave Cutler one is great. The other one so another character and, actually, there's a there's another book that I've actually got on the queue now as a result of all this. Adam, I made reference from earlier. This guy, Avraham Miller, is actually a really fascinating person who is at deck on one of these on one of the 3 rival efforts. He's on the the pro 350 side.
Speaker 1:He's he leaves DEC, and he ends up starting Intel Capital. But there's a really good oral history with with Asimov as well. So anyway. Sorry. Dan, I know I cut you off.
Speaker 1:I just wanted to make the plug for the computer history museum oral histories.
Speaker 2:No. That's I mean, that those those things are awesome. No no problem at all. I I wanted to make 2 sort of quick points about this question of, did Deck have to fail? So Gordon Moore wrote a retrospective, I guess, after he left digital.
Speaker 2:It's available on his website. And
Speaker 1:Gordon Moore or or or, Gordon
Speaker 2:Bell. Sorry. Gordon Bell. Excuse me. And I I I posted a put the link to this on Twitter the other day, and it's kinda fascinating.
Speaker 2:He basically made this argument that, like, look, if DEC was going to survive, DEC was going to have to transition from being a hardware company to a software company. Yeah. And nobody has ever done that successfully except for IBM. And and, like and and they just could not pull that off. And I think that really dovetails nicely with these things about, like, hey.
Speaker 2:They had Jeff. They had Sanjay. You know? They had AltaVista. Clearly, people were thinking about those things.
Speaker 2:Well, those are all fundamentally software projects. Right? And if you're a hardware company, you're like, well, that's cool that our research guys are doing that, but, like, whatever. Who cares? Like, we need to ship more alphas or, you know, people still wanna buy the vax or, hey.
Speaker 2:What's up with the titanium thing? I mean, you know, there are all these sort of questions that come down the pike about that kind of thing. And then sort of the other thing that I wanted to mention, if it'll come back to me because I, of course, forgotten what it was I was gonna say. Never mind. I forgot.
Speaker 1:And it it can it be interesting to know if this is the same, because, again, he's got this kind of this this, UAG 4 deck in deck instead of long of deck, and I think it's actually very interesting. And
Speaker 4:Yeah. The the those those are the same same thing.
Speaker 9:The same thing.
Speaker 1:Yeah. Hey, Johnny. Can I
Speaker 5:can I jump in? This is Shahin.
Speaker 1:Sure, Shane. How are
Speaker 3:you? Hey.
Speaker 5:How are you? Thanks for having me. Dan's been suffering me for the past couple of days on this Gordon Bell thing because I totally disagree with it.
Speaker 1:I think Oh, interesting. Yeah.
Speaker 5:Yeah. And, you know, I've been trying to make a case and, you know, Dan's making a great counter case, and I think we're having a good chat.
Speaker 3:Oh, yeah. Absolutely.
Speaker 5:I think the biggest mistake with these companies, Sun included, is that they abandon their installed base.
Speaker 1:Yeah.
Speaker 5:They lose heart. They they they sort of get tired of it. It, and they think that it's a losing battle, so they throw in the towel before they actually have lost the game. And the contrast for that actually is the IBM main Is that in the you know, 30 years ago, in, like, the early nineties, IBM mainframe was called dead, and their management was actively trying to kill it. And then Lou Gerstner gets appointed.
Speaker 5:And then he looks at the books and says, what's the problem? You know, we have an installed base. They want more mainframes. They're gonna give them more mainframes. And 30 years later, it's still, by itself, about a $2,000,000,000 product, plus carries a whole bunch of other storage and software and services and ISVs, etcetera, etcetera.
Speaker 5:And over 30 years, it's provided probably 1,000,000,000 of margin dollars that have enabled IBM to become a software company. So I think DEX's biggest problem was that they abandoned VAX. It was a cult following. You never abandon a cult following satisfied installed base. That's a sin.
Speaker 5:And frankly, Sun did the same thing for 4 years. I couldn't even find hardware on the website. It was all kind of, you know, the new and shiny object like the pie chart that that that that Tom shared. So I think that's the major sin, is that they chase markets that don't exist, and they abandon the market that does because it's not as cool as it was before.
Speaker 4:Yeah. I I I I think you're totally right. And and both these companies were addicted to super high growth new markets kind of stuff. Whereas, you know, IBM stuck to their knitting and was able to to grow that even though in the nineties in the nineties, the mainframe was the worst possible price performance. So it's really bad.
Speaker 1:Well, that's interesting because I I do feel I mean, I have to say, having read succeed? Because it is such a again, when they kind of invent matrix management and it is a total clusterfucking side. And I I think that part of the way they did that is that they and Sun 2, when you're sitting in this super high growth market, it's able to cover up a lot of sins and you don't have to be very operationally superlative and you get to have way too many people and you all these things that aren't issues when you have to go into a market and fight a lot more, honestly. I so, yeah, I think that's interesting. I think they kind of, I also feel that, like, you should never I feel, the the kind of the customers that brought you there, you wanna be sure that you are never denigrating those customers.
Speaker 1:And I do feel that Sun did that, a bit. I think Deck definitely did that. I noticed that, like, Amazon really does not do that to their credit. I think I think Amazon gets indisputably right. Amazon
Speaker 2:Amazon was also young compared to all of these companies. Right?
Speaker 1:Yes.
Speaker 4:It wasn't yeah. Every day, but yeah.
Speaker 1:Yeah. I mean, young ish. Right? I mean, Amazon put, like, 97, 98, something like that. And, you know, we're now I mean, it's you know?
Speaker 1:But they're a quarter of a century down the track, so it's not
Speaker 4:Okay. That's fair. But they they are way way past way away from the center of the curve. Most software companies can't keep an API the same for 2 years these days.
Speaker 1:And, you know, I kinda feel like that, you know, Amazon has got I I mean, obviously, you know, Tim was on the was was famously, infamously on the inside there and can and can speak to this. But I I I feel that the Amazon does a superlative job of if you're a customer building on Amazon APIs, you know that thing is gonna be there forever. You're not gonna have to wake up one morning to someone's OKR was to rip your thing out, which I feel it is is a, is a persistent fear, for those using anything from Google.
Speaker 9:Yeah. AWS in particular is, not only, you know, pretty good about keeping things around. They're smart. Like, there was the infamous, AWS SimpleDB, which, you cannot find on the website. It's not there.
Speaker 9:You cannot sign up to join it. But there is a big swath of happy customers who still use it every day and a team that supports it. Pulling that off is hard, but, you know, it it it it it really works well for
Speaker 1:them. Yeah. And I think it just makes I mean, certainly, from the outside in any way, there's I mean, Amazon has got, certainly an operational culture that would does not present at at DEC or SUN.
Speaker 3:They no one would accuse them of not having the money, Gene. Exactly. Unlike DEC, which is what, you know, the the the theme here.
Speaker 6:Yeah. Exactly.
Speaker 2:You know what I'm like like I I remember the thing I wanted to mention earlier, which was this video. It's a video that's a series of interviews with Olsen interspersed with other things about digital, and it was made probably sometime in the eighties. And it's quite fascinating. I actually think that in many ways, one of the things that made digital successful was that they they had such amazing engineering. And I think that one of the things that killed them was that in some ways, they were too far ahead of where they should have been.
Speaker 2:Like, they, you know, they had a vision of this completely networked world where you would have various sized systems, all of which were fax and or whatever. But, you know, I mean, still, you would have sort of a workstation that a manager or an engineer would sit in front of with a graphical display. There would be serial line terminals and concentrators and so forth. There would be on, like, the manufacturing line. There would be large fax and sitting in data centers acting as servers.
Speaker 2:All of this stuff is gonna be tied together with DeckNet. It was all very interactive networked and so on. Like, nobody was really thinking of I mean, Sun was to their credit, but, you know, that that was just IBM wasn't really thinking about that in those terms nor was anybody else in the industry. It was it was a very, like, dynamic environment that they were envisioning. And I I I think that they just weren't they were too early.
Speaker 2:They couldn't pull that off.
Speaker 1:Well and I think if you're gonna be that early, I think you can be that early, But if you're gonna be that early, you need to be willing to really grind and do some things that you may not want to do to get these things all the way to fruition. And I mean, one of the things that to me, at least in Sun's history, Sun had this, I think, very revolutionary product in Sunray. And infamously, the airlines were trying to figure out what their kiosk story was gonna be. This is when they were you know, they're they're first beginning to to, computerize all the the these these terminals for
Speaker 3:people to use.
Speaker 1:And Sunray was very, very tempting, but there was no USB printing support for for Sunray. And, Sun just like wasn't interested in USB printing. It was just really gross, not really sexy, not a fun problem, really difficult to do.
Speaker 3:Well, there there's some
Speaker 4:known impossible problems and among them are making printers work
Speaker 1:on units. That's right. Exactly. Right. And so they did and and I remember talking with the folks that because this is all happening.
Speaker 1:American Airlines is doing this. And I remember talking to the folks in Itasca, and and just in general in Chicagoland who were like and, Shaheen, I know you saw this too inside of Sun when you had these folks in the field that are, like, you know, red alert. Like, we need to get this we are gonna if we lose this business in the next 9 months, all the airlines are gonna do whatever American does. And they're all and we will lose this business completely. And there's this and it Sun just couldn't summon it until it was too late.
Speaker 1:And then when it was too late, they did try to summon it, but it was ultimately it was it was ultimately too late. And I, you know, I kind of feel like if you are gonna be early and SunRay was early, you know, Dan, all the things you're describing are obviously very early, you just have gotta be willing to really grit and grind to get it all the way through and to make it real. You can't just be attracted to the shiny object. You you you've gotta be you've gotta have, I think, culturally value getting those innovations all the way into customers' hands and at the coalface. And, you know, I thought, you know, Sun was kind of hit and miss on that.
Speaker 1:Like, sometimes it was good and sometimes it wasn't good. That's why we did Fishworks, honestly, because we were kinda sick of this stuff of some really interesting stuff not getting into customers' hands because it wasn't productized. So we wanted to really go after a very focused market with a product that was pulling together all of this stuff. But it was the exception, not the rule at Sun, and I I think probably I dare say a deck
Speaker 3:as well.
Speaker 5:I think it's just really challenging to cross those chasms and go from something that 5 customers want to something that everybody wants and figure out which ones those are. And Sun wasn't very good at it. And, you know, I'd like I mean, my observation is that somebody like an IBM is good at it. I don't know how good they are now, but they traditionally have had a good strategy process. And they fund things for success, whereas some specialize at funding things not for success.
Speaker 1:Yes. Especially with us.
Speaker 5:You know, it was like we fund you, and if you're a hero, you will be successful.
Speaker 1:Well, we got it.
Speaker 5:And that proves that you're awesome, and the product is awesome.
Speaker 1:Well and I think that, Michelle, how many times did we see and surely this happened to them as well, where you had things killed at the moment that they're beginning to get traction. It's like you you find you know, you've stood by this thing for 2 or 3 years. It's finally beginning to get some traction, and now you put the poet the brave. It's like, what are you doing? This is
Speaker 5:like Exactly. Exactly. Just when you're getting to the finish line, they would pull the plug. Now, you know, Scott himself used to say that you gotta have the staying power, that, like, Java would never have happened if you didn't stay with it. But staying power is exactly what somebody like a Microsoft has, well, simply because they had the margin dollars to do it, and some other companies have.
Speaker 5:But, but it's but it's hard to do. It's really I I agree that it's hard to do. It's whole kind of the innovator's dilemma thing.
Speaker 1:Totally is. Tim, I saw you getting in here.
Speaker 9:Oh, no. Just listening.
Speaker 1:Oh, there you go. And, Ben, I know you had your hand up a bit ago, and I know that, obviously, certainly a veteran of at least a couple of these companies.
Speaker 12:Now the moment may have passed, but I was just one we've talked before, like, Google, you know, they bring up, the shiny object, new product, and then it falls by the wayside. And some speculate that that's because the whole incentive structure for employees is to show impact, and, you don't have much impact for maintaining something like your established customer base. And I was just wondering, do we know enough about the internal, culture of these companies to to contrast those? What was it like being a deck employee, versus, you know, what we think might be more supportive were the product, the customer experience, longevity of products.
Speaker 1:Yeah. That's a that's a great question. Certainly, my read on it is that and this is where I just and maybe I'm reading too much into what feels like the familiarity from Sun, but at the same kind of autonomy to a fault where you have different groups with different views on things. Good news is you can kinda do what you want to a degree. Bad news is that the company itself is not gonna be coherent, and you're gonna have to fight other groups to actually get something all the way over the line or or I mean, just the the witness the the this kind of the the approach they they took with the 3 PCs.
Speaker 1:But,
Speaker 9:you know, that's not necessarily always true because AWS is famously totally incoherent. The try and spot a pattern, you know, at any given reinvent or what gets announced. And that's because they they empower the teams, and they all charge off in every direction. And somehow, in that case, it seems to work fine. I'm not smart enough to figure out why sometimes it works and sometimes it doesn't.
Speaker 1:So did Tim, do they pit are there rival efforts inside of one another? I mean, I think Adam had said this earlier. I mean, is that like, how do they make that work? Because it is interesting and surprising.
Speaker 9:Yeah. There are there are some rival efforts, but here's the thing, in in that space there's so much territory to attack. I mean, you know, the cloud is huge, but still, you know, 8% of enterprise IT. So, you know, there there's there's empty ground in every direction to charge into. But, you know, the the policy there is don't hold people back, you know?
Speaker 9:If they can make a case, there's people out there who need this thing, let them build it.
Speaker 10:Yeah. Well, the obvious example that comes to my mind on parallel efforts within AWS is how many different ways are there to run a container? You have Amazon ECS, which is their proprietary Elastic Containers service versus EKS, the Elastic Kubernetes service, and more recently, Amazon App Runner, which I think is based on some of the same internal infrastructure as ECS, but is kind of different than a high level
Speaker 9:And and Fargate too and lots more. And then on the messaging side, there's SQS and SNS and MQ and, managed Kafka. And I'm I'm forgetting a few in here. But, you know, it all seems to work.
Speaker 1:Yeah. Right. Okay. I guess right. I I of course, now I'm regretting asking if there are rival efforts so I can answer that question myself with all those all the things you just read about, Tim.
Speaker 1:Well, I I I think we wanna wrap up here. I know we I'm trying to keep these to an hour, but, of course, this was a, a topic that I think is pretty mesmerizing. I'm sorry. Jason, I know you've had your hand up. You wanna get the the kind of the last word in here?
Speaker 6:Oh, well, it's just I guess it's kinda I mean, too late now. I just think when you're talking about the efforts within Sun, it just reminded me of an experience as the customer where we were asking for better patch management. And, basically, the answer was, well, there's 2 groups that wrote 2 different engines, and they're fighting each other. And they can't agree on which one will actually be the solution to ship to customers, so you just gotta wait.
Speaker 1:Well That
Speaker 2:sounds right.
Speaker 1:Yeah. That that exactly. That's exactly the the no lies detected. I feel that and and, you know, it it's kind of the same too about these 3 rival efforts and that that they kind of the 3 out of all PC efforts and and putting them all in the customer base. I think, you know, I I used to joke that Sun's motto should be work in fuse so you don't have to be.
Speaker 1:But the reality is a lot of that confusion is exported to to the customer. And, you know, it's a challenge to how do you and I I I do think that, you know, Amazon has pulled us off to a better do degree. But, yeah, there's there's a balance there between, giving folks what they need and but not allowing rival efforts to stymie one another or compete with one another. I mean, it's it's it's it's challenging. I think, Shailene, I go back to what what you said about about losing track of the customer.
Speaker 1:And I I I do feel that is the thing that Amazon does exceedingly well, that Sun did well at its height, that I think Deck did well at its height, but I think it's easy to lose track of, which is always putting yourself in your customer's shoes. You know, let let me get in
Speaker 9:with with one more word story that I think was really the the death signal for me around deck. And I'm I'm see you see it happen in technology companies systematically. So I already said the sales organization was obsessed with IBM, and they just needed to be more businesslike. Like. If they were more business like than it'd be like IBM.
Speaker 9:And the story I'm gonna tell is when they launched the VAC 750, which was, you know, sold a ton of of computers and was was a very successful product. And, what happened was they had this this launch event, and they had a room full of people, and this is not a headquarters. This is out in the field. Right? To announce this new computer.
Speaker 9:And they showed, like, 25 slides all about business. Now we understand your business model. We're all about doing business, and the crowd's getting madder and madder because they wanna find out how much memory it has, you know, what the clock speed
Speaker 3:and all
Speaker 4:this stuff.
Speaker 9:And they just would not do that and what the price was. And they got, like, 45 minutes of this business y crap. And then, like, in the last 5 minutes, they said, oh, it doesn't cost, you know, whatever it cost and and so on. And, you know, as I one of the as I was getting ready to leave Amazon, I was starting to see that actually come in a bit at AWS, saying, you know, we we shouldn't be so bits and bytes y. We should be more, you know, business y because, you know, there's people who make the decisions.
Speaker 9:Although at both DEC and AWS, the way they got to where they are was by getting the technologists on board and routing around management. Yeah. So so there's there's a big danger there. That's all.
Speaker 1:That that is a that's a great story, Tim, and a a a great one to end on, I think. Certainly, I I I know as we I find it super interesting to go through these because there's a lot to learn, and I think that's something we all learn is the the technologists that make the decision and understand put yourself constantly in their shoes. Don't lose their perspective, and don't, please, don't, bury them with bullshit.
Speaker 2:I would I would add one more little thing here, which is we all kinda owe Digital Equipment Corporation a debt of gratitude. I mean, if it had not been for Deck Machines, most of the technology that we take for granted today would have evolved in a very different way. And and I, you know, that's the Internet, UNIX, CE, you name it.
Speaker 1:Yeah. Totally. Totally. Right. Yeah.
Speaker 1:So I there a lot to be said for Deck and, Adam, I to the degree that you end up getting, these books end up drawing you in. I apologize advance versus flooding your queue. But there's a lot of good stuff out there. Really interesting company in history. Yeah.
Speaker 1:Awesome stuff. Alright. Thanks everybody. Great great to hear from so many people, but and familiar voices and and some new ones too. So great to hear from everyone.
Speaker 1:Take care.