There's truly something to be said about keeping your mind sharp by continuing to work, either through hobbies or otherwise. Just need to keep the stress in check.
Yea it ain't going to be pretty if you just slump on the couch and usually have sausage and beer for lunch you'll be dead or decrepit in a decade. Wouldn't advice working full time to your grave, there aren't many who regret spending fewer hours at the office but suddenly after five decades turn to couch potato is a death sentence.
Genetics play a significant role in this context. His current intellectual abilities serve as a testament to the remarkable intelligence he possessed during his prime.
Please have Dave Cutler back for another round or bring on more industry veterans to share parts of their life, this is some of the best content I've listened to!
Soooo much this! These stories of how the foundations of modern computation were developed are incredibly interesting! It's also a real trip to look back and realize how far tech has come since then!
Dave Cutler is a remarkable individual. I feel privileged to be his friend, and worked with him during the heyday at DEC in the 1970’s. I was on the product management side of the RSX-11M and VMS, VAX-11/750 era, tried to recruit him to Intel (48:52) , and later offered him VC financing (58:34). Dave is the most productive and dedicated individual in the technology field I have ever encountered. His set of accomplishments tells that story, and IMO he deserves more recognition. I am so thankful this history is being recorded.
*At a seminar decades ago, Mark Russinovich commented that Dave Cutler is last man in the world to have an entire OS (WinNT) in his head. After hearing this conversation I am inclined to believe it. Respect!*
What really shocks me to my core is not how much Dave Cutler has done and been a part of. No. What shocks me is that he was a boss that actually knew something. It must have been heavenly to have a boss like that.
Back then I think that was much more common. You almost had to know what you were doing. It was emerging technology. He started off doing what he was eventually hiring people to do.
@@bhollingsworth Was more common in all industries, there was a better blend of those that worked there way up and had practical knowledge and those with an academy background with underlying theory.(also the academics had less off-topic fluff) Stuff was really sliding downhill aound the time when the "personnel" depts became HR depts, wich was also associated with increased bureaucracy and middle managment bloat as everyone tries to game the system.
I don't know if it would have been that good. Because he can critique you more in depth and with him being obsessed with performance and optimization you would be having to put in 110% daily which would burn out. But trade off is a boss you can bounce ideas off of and really know. If my boss was Dave, I would love it. But bosses like that that also micro manage are hell. Dave appears he trusts his employees and if your good enough for him, he would let you do your work.
Me, for the past 3 days: “That’s a lot of excerpts, is he going to leave anything for the full video?” Me, right now looking at the full video length: “Oh…”
Dave C is proof that if you keep your brain active, even at over 80 years old you can still be just as capable as someone in their 20s, but with way more life experience!
Dave Cutler’s recall of his past experiences, the amount of detail he goes into when telling his stories, is so impressive. He is one of my role models, and I am so grateful to Dave Plummer for making this interview happen.
Timestamps of clips of this interview posted on Dave's Garage (some clips have stuff cut out and the order changed) 0:56 - 5:38 The time Microsoft sent coffins to competitors 5:38 - 14:23 I Could Have Been a COBOL Programmer! 26:15 - 31:01 Software with ZERO bugs 53:44 - 1:01:59 Microsoft's "Pathetic" Operating Systems - Steve Ballmer and Breakfast at Denny's 1:01:59 - 1:10:33 Linux-Xenix-Unix vs OS/2 and Windows 1:37:39 - 1:42:20 Windows Tukwila 3.99 and Windows Cairo 1:51:01 - 1:59:52 Windows Longhorn and the Worst Code I've Ever Seen 2:18:59 - 2:21:11 What Successful Programmers Do That Others Don't
This kind of people never retire. Considering it's the work of a lifetime it makes sense, and also it's pretty good to stay healthy, keep the brain working.
@@lppedd Actually it's usually the corporate who pushes people out as they age. But I can understand, a guy with these credentials, they wouldn't wanna lose an employee who is one of the few people who knows a lot about windows and its internals.
I grew up in a Central-East European small country, and my career has been based mostly on Microsoft's products from the early days of MS-DOS 3.x, Windows 3.x, NT4, 2k, 2k3 etc to Azure nowadays. I read about Dave Cutler sometime in the 90s, and I knew, he is the genius behind the scenes, and he is one of the people affecting and driving my career the most. Watching his interview is a very special experience, for which I am very thankful. If it was 10-20 hours long, I would still watch it. :)
It's awesome to hear an octogenarian is still coding! I'm in my 50s and I am refusing to leave the technical aspect. Most of my colleagues have gone on to pure management positions. I have chosen to stay in the technical realm and I'm still actively coding. I am a senior architect. I do have direct reports, but I'm still heavily entrenched in tech and I want to stay there. Programming is like solving puzzles, and they say the best way to keep your brain healthy - especially as you get older - is to solve puzzles. I owe everything I know to DOS Debug - so whomever wrote that you are a god :D. I taught myself assembler with DOS Debug when I was 13 and that paved the road for my life goals.
Management sucks for an engineer. Spent 12 years doing it - mostly that was a massive waste of time. Never actually accepted the rules of management - guessing instead of understanding the problem, blaming others for problems instead of just finding the root and fixing them, pretending to be infallible, so not willing to risk whatsoever, and thus not being able to create anything worthy. That's screwed up so badly. There are a few percent of population who are inherently good managers, there rest shouldn't even try. On the other hand, while I was attending bullshit meetings and writing bullshit reports, something bad happened to the technical expertise in the programming world. It's really difficult to hire a senior developer these days. The education mostly became oriented to learning patterns instead of thinking (and those patterns mostly concert integrating someone else's code, not writing your own, for some unimportant applications). So few years ago I switched to writing code as well. So long management nightmare.
@@InconspicuousChap Thank you for this comment. You really hit the nail on the head. Management is a massive waste of time for engineers. Many companies/IT departments these days are run by people, who don’t understand technology and got zero interest/passion for it. Sadly, too many developers are tempted by money/illusionary sense of power and join them. My 8 year management journey left me burnt out, disillusioned and cynical. I went back to programming and it took me awhile to recover. I never looked back.
Its funny what Cutler said about Gates saying they need to be nice to competitors. When I worked at Microsoft around 1997 I attended a sales meeting in Montreal and Ballmer showed a video of 2 kids in a dark alley wearing Oracle t-shirts vandalizing with spray paint a poster of either Windows NT or SQL Server. Then a super hero type guy came out wearing a Microsoft costume caught them and sprayed them with machine gun fire. It was pretty shocking.
Dave P you are a tremendous interviewer. Lots of space for the person to answer and knowledgable about the subject. I do hope you continue this as a series.
Agreed. It was refreshing to have the guest do nearly all the talking. So many other tubers do at least half if not more of the talking than the guest. Great questions, too.
I was also going to leave this comment. I’m so sick of interviewers who can’t avoid filling the interview with the sound of their own voice. The way to do it is what Dave did here: frame the story and then get out of the way and let the guy your interviewing tell it his way, and don’t feel like every pause is an opportunity to remind your viewers why you’re important. A couple seconds of dead air is OK, if it leads to another great story from the person you went through all this trouble to get on your show. DP knows his audience, and I suspect knows this great interview style from having to live it on a pretty much daily basis with his autism. I don’t know if I’m on the spectrum, but I am introverted, so I know all about verbally lining up the shot, and then stepping back so the other guy can take it and play the rest of the game.
Word of advice: if you find any valuable video on youtube that is worth rewatching from time to time, archive it yourself. Just remember: in 2023 december, google starts the inactive account purge project. If you know any channel, whose owner deceased (for example during covid), and hasnt logged into its google account in the past 1 year or so, now these videos will be the first candidates of the deletion. Search after it yourself, if you dont believe what I said. Thousands or millions of old videos will be gone forever very soon. Im not exaggerating.
Oh my god, Dave Cutler was my hero back in the days, when I got my hads on leaked windows nt kernel code, reading his design docs and seeing his name on top of source files felt like finding treasures!
As someone, who spent most of his career on "the other side" of Microsoft (first Novell, then the Linux world) and who would never consider working for Microsoft for some of the tricks the company pulled on products like DR DOS (an ex colleague of mine was a witness in the trial resulting from that) - I must still say: Dave Cutler has always been a name we all said with a great level of respect. In general, it seems that the engineering side of Microsoft was much better than what the product marketing teams made of it. On Dave's remarks on PowerPC and IBM making sure they don't create internal competition to AIX: I was at IBM in the 2000s and worked on the introduction of Power Linux systems - and went through the same pains then. Great to see this discussion! Thank you both for sharing.
He has to diss PowerPC because he is a "Wintel" guy. With all due respect (which really is due for Cutler) saying an architecture is bad because IBM cannot produce a decent compiler is - well, what do you expect. Microsoft was always married to Intel with AMD allowed to play along. PowerPC was a neat RISC architecture beating Intel/AMD to 64 bit (which was much cleaner on PowerPC than the 64 bit hacks of AMD) and had a vector unit (Altivec/VMX) before Intel even thought about it (i did some Altivec stuff and would choose it over the horrible SSE any time). They could just have used GCC or the Freescale compiler (PowerPC was a collaboraton of IBM and Freescale and is used in embedded devices until today). It wasn't that IBM did not want to "cannibalize" anything, it was that any flavor of Windows shouid commercially run on Intel and Intel only. It's all politics with M$.
@@therealmccoy7221 well, as indicated in my original comment, I was at the receiving end of IBM protecting their AIX revenue at almost all cost (knowing the margin on those systems, that clearly is a sensible short term measure). I cannot say what they did with PowerPC, but I can certainly see them run much the same course there.
@@therealmccoy7221 sure, and the market overlap for an xbox and a power AIX system is exactly 0% - which is not true for the overlap between AIX systems and Windows server (or in my case, Linux servers). What I described and what I believe is similar to what Dave C described, is the power the AIX team held over the power hardware. I made no statement about the quality of the architecture.
I don't care whether it is the year of Linux or not, or if Windows wins the 'OS battle' or if 'MacOS remains superior'. I just have so much respect for legends like Cutler, Tevanian, Torvalds, Thompson... those folks who built that Apollo Guidance Computer... (ah yes you too Dave! Coding in ASM during winters is fun), that being alive at a time when I can watch interviews of these legends is a pleasure, is inspiring and makes me grateful for the foundation that has been laid down by these giants. Thanks Dave for interviewing Cutler and uploading it here on RUclips. The fact that this is 3 hrs long shows how your passion won't be subject to viewership statistics (a few months back CHM interviewed Ken Thompson and it was about his chess machines... no doubt an interesting interview... But I wished that it was as extensive as the one they did for Donald Knuth which was in two parts and each part being 3+ hrs long). Once again, thanks Dave!
It's interesting to see that his whole career started on essentially doing a job that he lacked the applicable formal education and experience to do, with the way it is now, he'd likely have never been given the opportunity to even start where he did at DuPont on the Scott Paper project, nowadays every HR department would look and go, "no relevant experience," "no relevant formal education," and promptly file your application into the rejected bin and wouldn't allow such a position change to occur.
Did you catch the part where they sent him off to IBM to learn GPSS III? There were no formal degrees in this stuff back then. I suspect he got sent because no one else wanted to go or be stuck with dealing with that end of the computing business. I assume that DaveC was noticed as a go-getter and was picked for that reason.
But your point that HR would filter him out has some merit. When I left MSFT, the bar was basicly a masters degree in CS which I think most of the people who worked on NT 3.1 wouldn't have met.
This is typically the advantage when things are brand new and essentially no one knows anything about them. There were plenty of jobs when he started that rejected people with no relevant experience. One of the problems faced today is everyone is familiar with software and all software is just software. To many there is no different between an iOS app and developing a new LLM.
While I agree with what you’re saying in regards to what would happen today, in fact back then he did have the requisite skill set. Math and physics were his minors or at least study subjects for him, I forget if he said he degreed in them or not. So he had what would have been the appropriate skill set back then. I doubt there were any computer science degrees being offered at the time.
How cool! I hired on to DuPont as a Chemical Engineer in 1968 and programmed on two of the machines that Dave Cutler mentions... Univac 1108 (I think he went from Bunker Ramo to DEC without mentioning that the 1107/8 were Univac computers) and the PDP-10 at the Ex Station. I never met Dave (but wish I had) as my programming was exclusively batch.
Univac 1108?? isnt that the one that executes fortran instructions directly from machine code? that was one of the all time best computers ever made in human history, a friend of mine used to repair those when they were at their peak.
@@bradleyr4451 Oh man... this is taxing my memory. The only thing I remember about the 1108 Fortran compiler was that it had very good diagnostics. I recall compiling Fortran programs on the 1108, correcting any syntax errors and then recompiling on the local IBM 1130 computer (which had very cryptic diagnostics). Also, it had drum magnetic storage (Fastrand, as I recall).
@@amendegw thats so cool! Id hear about that amazing machine, top of the line, incredible. I heard one where they were running a batch program and it was instantly done, like 300 milliseconds or? and they said there must be a mistake it should take longer to run, do it again... and again.. and the station controller came out yelling I have other batches to run what are you doing? and they said its finishing when it starts and he said it runs Fortran natively! I heard that a few times and used to wonder. Yes drum magnetic storage they worked on those, heard how one was so big and heavy it would move and they had to bolt it down? Young ppl would be in near shock to see what memory was. I worked on the CDC storage module drives and Gould/Encore superminis and later disk certifiers, so much precision electronics and science went in to all of it; Id really like to hear anything about those big systems
Fabulous and fascinating history! Thank you, Dave P for thinking of doing this and making it happen. You did a great job interviewing Dave C. Dave C’s departure from DEC, where I worked at the time, felt like an earthquake. He was, and is, one of the best.
brilliant interview, I have watched the clips and now watching the whole thing. amazing. I have programmed PDP-8, PDP-11, VAX, CP/M, DOS, Xenix, OS/2, Windows, and Linux and although 15 years shy of Dave Cutlers age I am still programming. So all this history really resonates with me.
@@JonathanMcCormackyes we did use pathworks but with DOS and Windows 3.11. A mainframe arrived later on and SNA server on NT 4 replaced the VAX and pathworks setup
From a technical standpoint my time with Digital, the VMS operating system and all the networking around that was the best time of my career. I just never learned so much so fast. When things expanded to Pathworks on the PC's, Alisa Systems on the Mac's (remember the Mac originally only had Appletalk), SNA connectivity with JNET on the VAX, Ultrix, DEC X.25 gateway to that thing called the Internet...it was just the best 10 years I ever spent.
Thank you for noticing! I was wondering if anyone would be able to spot the idle pattern from RSX11m, and I think you're the first to do so, or at least say so :-)
@@DavesGarage Dave, I also noticed the PiDP-11 running RSX-11M+ in the background behind Dave. It is likely running the RSX disk image that I provided to Oscar for the PiDP-11. There is an improved RSX disk image available from Johnny Billquist who created the TCP/IP stack and utilities for M+ in the past few years. I was wondering if Dave Cutler was aware of the renewed interested in RSX11M+ that's happening because of the more than 4,500 PiDP-11/70's that have been sold?
This is the only 3hrs video I watched full on RUclips. It’s better than Netflix. I remember to run the payroll systems of a company I worked on, on a Windows NT. That thing was never turned off I believe 😅. The amount of details Dave remembers is crazy, I am 48 and I don’t remember the model of the servers or systems I ran 20 years back. Thanks for such great interview.
I'm glad someone finally sat down with Dave Cutler and asked him in-depth operating system questions. It's really interesting to trace the origins of the thought process that went into WinNT's design.
OMG….. 3 hours this is awesome!!!!! Thank you both… I was in building 27 working on Win95, IE, Visual Studio and ran the internal IPTD web presence under BradSi (Silverberg). Best job of my life…. Was great hearing Jim’s name mentioned, I was the person who smart mouthed him back on email on our building 26/27 DL BEFORE checking who he reports too…. LOL… I had no idea Msft Security had flash bangs and OC spray, 😂 I also worked down the street in the DEC building for Compaq (I still drink coffee from my DEC Alpha coffee cup) so this entire video made me so homesick for two companies that USE to be the greatest places to work on the east side…. Miss that part of my life so much… well maybe not the OC. 😉
Would love a second round with Dave Cutler talking about his Xbox work and some of the stories around it. Maybe that’s not possible due to it being more recent work though.
Was he involved in XBox? The story I heard as that they 'took the Windows 2000 source code and hacked it' to make the original XBox OS. It sounds like the sort of thing Cutler would be very dismissive of.
Rumor has it that he was brought in, or maybe volunteered, to step into the Xbox project a while back to help save the Xbox One virtualization layer for Xbox and Xbox 360 backward compatibility. I’m not sure if he was heavily involved before that, but I believe he’s been working on Xbox stuff for a while now…but I could be totally off base.
He talks a bit about Xbox at around 2:00:00 in the video. Didn't know that all games are an actual VMs complete with its own OS. Thinking of that, if that's the architecture, it wouldn't be hard to run Xbox games on PCs. 🤔
@@rgl168it’s not quite it’s own full OS for each game, but the game application itself runs in a sandboxed virtualized environment, so it can only write to its own dedicated save slots and it’s memory space is isolated from the OS. It’s based on Microsoft’s Hyper-V technology.
This series is so damn awesome. Is there any chance you could get Mark Russinovich in on this OS history? From DEC times through scaring the integrity of Windows, to a MS employee to his work on VM's and containers, pre Docker, and onward into Azure?
Oh man what a privilege to be able to listen to this legend telling his stories. When I started my electronic engineering graduation, MS-DOS was at version 3.3 and I’ve had never heard of Windows. Still today I eventually need to launch a DOS VM in VMware to solve specifics problems on very old machinery. A few months ago I still had a Fidia (Italy) milling machine running on Win NT 4.0. This interview brought me a lot of memories. Thank you very much Dave, both of you, for sharing this with us.
Thank you Dave for just sitting back and letting this guy talk. I could listen to his stories for hours! I was so happy you didn't rush him along. You did an awesome job with this interview. I want more stories from this guy!
If age has taken a toll on Dave Cutler's mental faculties, it either doesn't show or he he's lost more than some people have at all. I'm 35 years younger than he is and probably not half as sharp.
Very precious, part II please! This is easily one of the most interesting and capable programmers on the planet, I wish I would see more content of him!
@@StuartWoodwardJPI’m pretty sure there was a scene in that book where he punched a wall at Microsoft and left a hole in it. He across as hardcore, but was 1000% committed to quality
Thank you Dave and Dave. I'm just a few years younger than Dave C, and I spent the first seven years of my professional life at IBM in their OS and complier System Test / Quality Assurance teams. So much of what Dave C said about fixing your own bugs or letting the next team "down the hall" do it are so true. Your discussions brought back so many great memories.
Listening to this interview was a blissful experience; having spend 5 years writing Assembler for PDP11 on RSX11S, having done some work on VMS and then substantial work in Windows NT and later versions this interview touched so many points and brought me back to many amazing memories. I am impressed about Mr Cutler and his amazing memory, recalling architectures, bit width and so on. Dave thank you for sharing this amazing content and, of course, for your whole channel
Having worked for CSC, I can't imagine the days when they had people who could write compilers, and I went to a college where a lecturer had written a FORTRAN compiler for a 1961 computer, just to prove it could be done.
Nope. 5-track paper tape. The machine was a Ferranti Sirius. Our college had 4 of them at one time, from a production run of 20. A store room also held a ZEBRA, with magnetic drum as main store.@@DavesGarage
It is very nice to meet Dave here in this meeting as I was AFAIK the first and only user of RSX-11a in Europe. Was some 50 years ago when I used a PDP-11/10 for controlling a high-voltage switchyard with automated switching sequences. With several modifications, this program ran under RSX-11m/s until all 'experts' got afraid of the millennium bug and did not know how to deal with a PDP-11 assembler program. BTW: I also modified the RSX-11m terminal driver for the VT30 color monitor controller. So I was programming on a color terminal already in the late 1970s. We needed it for the display of the high-voltage switchyards and controlled them by joystick and keyboard via (very slow) data connections to remote PDP-11s.
Having a boss like Dave Cutler must have been great, I totally get why you worked at Microsoft. Imagine working for someone who understands the industry hes managing people in. You probably didn't have to spend half your day trying to explain what you did in a way that a manger wouldn't think was you being a wizard with a magic wand.
@@MultiPetercool As long as they value honesty, and can admit faults when presented with the evidence, those are the best people to work with. For the simple reason that you know there is nothing stuffed under the carpet. If you don't get away with it, others don't get away with it. It keeps politicians and lazy people away. It can be tough sometimes, but it's worth it. The same with Steve Jobs, he was not always right, but the people who complain about him were mostly wrong.
I was already in the doghouse today, so I spent half the day listening to this and fixing all the broken things around the house. Loved this interview. Dave C is an absolute legend.
@@shallex5744 the guys at Stanford basically built this modern era computing infrastructure but there's always a moron , I bet you're a zealot of sorts.. tsc, tsc, monolith operating systems... Tanenbaum was right
I would honestly love to see Cutler do a super basic kernel/DOS on something retro and relatively simple like a C64. If for nothing else then to just see how he approaches it. You can learn so much just by watching how someone works and thinks about it.
I just finished reading "Showstopper!" recently, so this interview is totally on-time for me. I heard that Dave Cutler does not like giving interviews, so congrats for getting him to talk for 3 hours.
Almost half way through. Great history. I started my technology dependence with the Tandy Model 1, Level 2 16k with tape drive. When I started working for a Radio Shack Computer Center, 3.1 and 3.11 were the Windows I became adept at for my customers. I was even selling SCSI hard drives when they were nearly unheard of in the consumer market. Now I’m so old I’ve moved to the dark side for the simplicity. But for more than 30 years, Windows, applications and users, put food on the table.
A lot of DEC folks I knew said Windows NT or WNT was one letter better than VMS like IBM and HAL in 2001. DEC was pinning its success on the success of NT in the later days.
I love the talk at the end about the lack of precision on 24 bit mantissa's, i worked on the Ferranti Argus computers they ran the bloodhound missile guidance system and a lot of other safety critical systems, they had an 8 bit mantissa and 4 bit exponent ( memory was expensive ), the lack of precision was laughable, you would type 2.41666 into it and reading it back it would come back as 2.42000.
That is odd. The PDP-8 had 12-bit words and they used three of them for all floating-point numbers. Well, eventually, 48 bits for FORTRAN 4. 8 bits of precision does sound awfully meager.
Analogues were 16 bits in total, 8 bit mantissa, 4 bit exponent and 4 status bits. 4 status bits was extremely generous but when talking about real plant signals were critical.
I’m a DevOps guy. I’m adopting the term FCIB (described approximately at 2:00:30) and using it this very Monday. Might even make a meme out of it. Love it ❤
Dave is surprisingly smart and well spoken for someone in their 80s. Doesn't seem like he's slowed down cognitively much at all. I wish we asked more lifestyle questions of these great achievers... Perhaps we could all stand to learn a few things from the habits they employ.
Some people lose a bit of memory and mental sharpness through their late 60s and 70s, but quite a lot remain 100% sharp. I believe there's research showing that staying engaged with challenging projects bends the odds in your favor.
Thank you, both Daves for this fascinating review of compute history! Best spent 3 hours in a long time. In the early days I ran NT 3.1 Beta on a Dell server that was purchased to run Novell Netware 3.12 or so. The interview section on NT4.0 prompted me to build a new VM on my VMWare homelab, install NT 4.0 on it from the ISOs, apply SP6a, then VMWare Tools. All ran without a glitch. Network is up and running now, on less resources than my smartwatch. All this before the interview was over ;) Back in the days we ran out entire company on that. I agree with Dave C. that it's sad to see "Hello World" take up 1MB of code. Current software isn't coding, it's glueing libraries together. No wonder that bugs are harder and harder to track down.
He’s still actively contributing to the code, I’m honored to be able to work on the windows kernel which he firstly worked on decades ago when I havent even born:). All the design he made and the code he wrote is just a masterpiece, which is incredibly still working the same way that he did in 2024! (You will know what Im saying if you are the engineer, and understand how impressive is this). He’s always my role model in my whole career, and will always be, to help design a software as perfect as possible.
Absolutely amazing interview, I was mesmerised for the full three hours! He is such a legend and I would love Part 2 focusing more on his work at XBox!
This was a great interview. I've read Showstopper about 5 times I think and am fascinated with what Dave C & team was about to accomplish. And am blown away that it's still the basis of Windows today. I would've been interested to hear what Dave C's thoughts are on AI and writing code. Thank you for doing this interview!
DaveC’s thoughts about AI and ML would make a fantastic follow up to this interview, if he’s been keeping up! (Not easy for anyone to do given the constant development in the space)
Dave, the iAPX432 was Intel's biggest clusterfuck, ever. Eclipsing the i860, and that was a big one too. From a programmer's perspective, it was a CPU specifically designed for Ada. The application programmers hated it because they didn't want to deal with Ada. The systems engineers hated it because it required half a dozen chips to get even the most basic system running. The system programmers hated it because a subroutine call took hundreds of microseconds at a minimum. And what ultimately killed it off was the fact that the iAPX286 was available, which ran twice as fast, for half the cost. Dunno if people really understand this. Intel's track record for new processor architectures is beyond shit.
It's easy to say this with 50 years of hindsight, but you could argue that Intel were taking some of the directions the rest of the industry were making to their logical extreme. It borrowed heavily from the mainframe that programmers preferred to use - the Burroughs 6500 - and the direction the industry was going to impedance match with high level languages. Don't forget that the first RISC paper really tried to optimise call/return with that register windowing concept. It's more that Intel have repeatedly failed to understand market forces: nobody wants a multi-chip implementation of an old mainframe in the 70s, and when they finally got it right in the i960, they refused to sell the full featured version outside of the military. Itanium was a similar problem, it wasn't affordable enough to make up for the fact it was on a dated process node and didn't run existing code at a half-decent speed. The worst part was that Intel made it difficult to develop for, by failing to drop the price. If I'd been able to pick one up back then, I'd have been able to write some good quality jit. The architecture is not bad - it's no i960 - but it's still the nicest modern architecture we have.
I don't know about "nicest". In some theoretical sense maybe. But nobody has figured out how to write a compiler that generates good code that fills out those long words and runs very parallel. A few theorists think it's impossible to do so. @@capability-snob
@@georgegonzalez2476 the hard part of writing fast code on the itanium isn't filling out the bundles, it's using the prefetch and speculative machinery effectively across linkage boundaries. ia64 has a lot of great memory-latency-busting features, but since you've got some 1200 instructions between a load from main memory and the first time you can use it, you really want to move prefetch long before your function is ever invoked. That sort of supermodularity is possible - especially with whole-program compilers - it's just not how most compilers are written.
More please. Dave's reminiscing brings back so many memories from my own work history. I remember working with a number of the things he talks about ... but I was very young man at the time and he got there well ahead of me. Plus he's obviously held up better with age - I seriously doubt I'll remember all those details when I get to 81. He obviously deserves all the success he's had.
So much great and relevant history! That's what I love about this channel. Nothing compares to learning from the people who were actually there and built the new things at the time. Thank you!
Met Dave Cutler once at DEC. I joined in 1966 as a hardware techi from the UK. PDP 7-8 8i,8s-9. Fixed them and ran courses for customers. Thank him for the interview and wish him well. Have you any interview which would give a similar overview of apple software development? Visited the Swiss Super computer center last week. Quite an eye opener. Thanks Chris
No wonder we got the small bites. 3 hours! Haha that is amazing though THANK YOU for giving us this long form content. I watch a skateboard podcast that puts out interviews between 2 and 6 hours every week, and have been for years. There is still interest enough for it among the want for shorter and shorter stuff.
"I like to tell stories" - and doesn't he tell interesting stories, and tell them well! A great listen for those of us with any history in IT. Current programmers ought to listen to this; they will gain great insights into WHY things are the way they are.
This is an incredible interview! At first I kept reading 3 mins instead of 3 hours 😲. Thank you for this absolute treat! 🙏 I joined Microsoft the day Windows XP launched in the Windows Division and Dave was one of the first titans I visited at his office in Building 26, 3rd floor. Thank you Dave for everything you've done for Microsoft!
A software engineer's software engineer. His passion for the trade is reflected by his command of historical details. This interview will go down as a classic. Good job.
It was great to hear from Dave Cutler what really went on with the versions of windows from one to the next. Some of us in the industry were close with our guesses obviously but it is fantastic to hear the details from the expert himself. Thank you and I hope he comes back for another chat session. I hope the linux work he is doing also gets fed back into the GPL community.
In my CompSci coursework, he is referenced for VMS, but his wealth of knowledge is far more encompassing than that. This is a VERY valuable presentation of our field's beginning, and how far we've come...
Fantastic interview Dave, thank you. You mentioned toward the end a rotary phase changer your neighbor has. They are extremely simple, it’s a 3phase motor running on single phase with the extra leg from the motor generating the 3rd phase. In my younger years I built several of them for friends and relatives.
Seeing the amount of detail you have posted with regards to Windows and the O.S has had my attention. In watching this conversation it has been very cool just to hear. My question though, is why the mic was muted @ time 2:06:55 while talking about the X Box, virtual machines and the game packaging ?
3:30 “There’s a little side story here too. I hope this doesn’t go on too long ‘cause I like to tell stories”. Ummm…. Not a problem: we WANT the side stories and we will TAKE the time.
Amazing interview. It took me a few installments to watch the full thing, but it was worth every minute. The detail in which Dave C. recalls this stuff is extraordinary.
in the early 2000s I used to come into the office to the kitchenette and listen to DaveC and Jim Allchin argue about the topic of the day. If you are wondering WIM = Windows Improvement Meeting or Windows Integration Meeting depending on who you asked. Happened every week or so on Friday afternoons, especially when we were close to milestones. Lots of food, beer, and friends. Those were good days. RIP Bld 26 cafe...
Thank You Dave. I have been playing around with computers since the early 80's but I never got as good as you are. Please have more of the people of years gone by, while they are still here to teach.
Wow Dave. Thank you for creating this content and contribution to the OS’s I have used. Excellent interview. It really took me back to when I first got into IT. I have not heard the terms PDPx or VAX/VMS in 30 years. And thank you Dave C. for your contributions to all the operating systems you created or had influence on. This interview answered a lot of curiosity questions I had.
Reason why the internet is something great. Been a fan of the lore of DC since the 90’s but this is the first time I’ve heard him speak. Huge thank you. I played with Vax in the 90’s, kilostreams, pathworks and Novell, before getting my hands onto NT which changed everything for us and shaped my career. 35 years later - semi retired myself, yanked back now and again for problem solving for issues that your average ITIL and TOM structure prevents from being identified.
81y.o. and still fully sane and capable of intellectual work. I wish everyone such good health, thanks for the interview.
There's truly something to be said about keeping your mind sharp by continuing to work, either through hobbies or otherwise. Just need to keep the stress in check.
Yea it ain't going to be pretty if you just slump on the couch and usually have sausage and beer for lunch you'll be dead or decrepit in a decade.
Wouldn't advice working full time to your grave, there aren't many who regret spending fewer hours at the office but suddenly after five decades turn to couch potato is a death sentence.
Keep using your mind or you'll lose it!
Genetics play a significant role in this context. His current intellectual abilities serve as a testament to the remarkable intelligence he possessed during his prime.
I saw a photo, he was working on early days of Azure.
Please have Dave Cutler back for another round or bring on more industry veterans to share parts of their life, this is some of the best content I've listened to!
Raymond Chen and Mark Russinovick
BRING BILL@@monad_tcp
Yes same here.... And I would like to hear about Azure history, reddog from the guys that were there!
I nominate the Father of the Zen and Apple A architecture Jim Keller
Soooo much this! These stories of how the foundations of modern computation were developed are incredibly interesting! It's also a real trip to look back and realize how far tech has come since then!
Dave Cutler is a remarkable individual. I feel privileged to be his friend, and worked with him during the heyday at DEC in the 1970’s. I was on the product management side of the RSX-11M and VMS, VAX-11/750 era, tried to recruit him to Intel (48:52) , and later offered him VC financing (58:34). Dave is the most productive and dedicated individual in the technology field I have ever encountered. His set of accomplishments tells that story, and IMO he deserves more recognition. I am so thankful this history is being recorded.
How many Daves are responsible for today’s technology 🧐
Wow, thanks for such details 🙂
*At a seminar decades ago, Mark Russinovich commented that Dave Cutler is last man in the world to have an entire OS (WinNT) in his head. After hearing this conversation I am inclined to believe it. Respect!*
What really shocks me to my core is not how much Dave Cutler has done and been a part of. No. What shocks me is that he was a boss that actually knew something.
It must have been heavenly to have a boss like that.
Back then I think that was much more common. You almost had to know what you were doing. It was emerging technology. He started off doing what he was eventually hiring people to do.
@@bhollingsworth Was more common in all industries, there was a better blend of those that worked there way up and had practical knowledge and those with an academy background with underlying theory.(also the academics had less off-topic fluff)
Stuff was really sliding downhill aound the time when the "personnel" depts became HR depts, wich was also associated with increased bureaucracy and middle managment bloat as everyone tries to game the system.
@@mytech6779 I agree.
I don't know if it would have been that good. Because he can critique you more in depth and with him being obsessed with performance and optimization you would be having to put in 110% daily which would burn out. But trade off is a boss you can bounce ideas off of and really know. If my boss was Dave, I would love it. But bosses like that that also micro manage are hell. Dave appears he trusts his employees and if your good enough for him, he would let you do your work.
@@MikeHarris1984 agreed. It sounds miserable. People have said Gates was terrible to work for also for the same reason.
Oh my, it's three hours long. Exactly what I'd hoped for...
Haha just noticed, yes indeed I am in it for the long haul 3 hour cut.
So good
Me, for the past 3 days: “That’s a lot of excerpts, is he going to leave anything for the full video?”
Me, right now looking at the full video length: “Oh…”
Perfect length for a full interview!
I was wondering what the draw to the fell interview would be after so many clips.
Oh.
@@euromicelli5970 same here.
Dave C is proof that if you keep your brain active, even at over 80 years old you can still be just as capable as someone in their 20s, but with way more life experience!
He is amazing!! 80 years and that eloquence is fantastic
I was thinking the same, its so cool to watch. hahahaha
Really unbelievable how switched on he is. He sounds 40
Dave Cutler’s recall of his past experiences, the amount of detail he goes into when telling his stories, is so impressive. He is one of my role models, and I am so grateful to Dave Plummer for making this interview happen.
😊😊😊😊😊
😊😊0000000000000⁰000000
His memory is ridiculously good, this guy is sharp.
I'm half his age and feel like a dimwitted sausage after watching this, can barely recall what I did a month ago at work. 🤯
I wish I was that sharp when I was in my 20's.
Timestamps of clips of this interview posted on Dave's Garage
(some clips have stuff cut out and the order changed)
0:56 - 5:38 The time Microsoft sent coffins to competitors
5:38 - 14:23 I Could Have Been a COBOL Programmer!
26:15 - 31:01 Software with ZERO bugs
53:44 - 1:01:59 Microsoft's "Pathetic" Operating Systems - Steve Ballmer and Breakfast at Denny's
1:01:59 - 1:10:33 Linux-Xenix-Unix vs OS/2 and Windows
1:37:39 - 1:42:20 Windows Tukwila 3.99 and Windows Cairo
1:51:01 - 1:59:52 Windows Longhorn and the Worst Code I've Ever Seen
2:18:59 - 2:21:11 What Successful Programmers Do That Others Don't
2:52:17 Dave C. eating something.
2:59:50 While talking about X Cloud, Dave C. bumps the microphone and expertly realigns it again.
2:06:55 apparently what are your choices on updating games vms in xbox it's been censored
He is still coding!? I thought he would have retired by now. Seriously, what a legend!
This kind of people never retire. Considering it's the work of a lifetime it makes sense, and also it's pretty good to stay healthy, keep the brain working.
@@lppedd Actually it's usually the corporate who pushes people out as they age. But I can understand, a guy with these credentials, they wouldn't wanna lose an employee who is one of the few people who knows a lot about windows and its internals.
You never give up coding.
I’m feeling my own age a bit today as I celebrate a milestone birthday, and hearing Dave Cutler say he’s still coding is so inspirational for me.
He cant. He knows to much lol.
I grew up in a Central-East European small country, and my career has been based mostly on Microsoft's products from the early days of MS-DOS 3.x, Windows 3.x, NT4, 2k, 2k3 etc to Azure nowadays. I read about Dave Cutler sometime in the 90s, and I knew, he is the genius behind the scenes, and he is one of the people affecting and driving my career the most. Watching his interview is a very special experience, for which I am very thankful. If it was 10-20 hours long, I would still watch it. :)
Witaj! 😉
It's awesome to hear an octogenarian is still coding! I'm in my 50s and I am refusing to leave the technical aspect. Most of my colleagues have gone on to pure management positions. I have chosen to stay in the technical realm and I'm still actively coding. I am a senior architect. I do have direct reports, but I'm still heavily entrenched in tech and I want to stay there. Programming is like solving puzzles, and they say the best way to keep your brain healthy - especially as you get older - is to solve puzzles. I owe everything I know to DOS Debug - so whomever wrote that you are a god :D. I taught myself assembler with DOS Debug when I was 13 and that paved the road for my life goals.
Ditto.
Management sucks for an engineer. Spent 12 years doing it - mostly that was a massive waste of time. Never actually accepted the rules of management - guessing instead of understanding the problem, blaming others for problems instead of just finding the root and fixing them, pretending to be infallible, so not willing to risk whatsoever, and thus not being able to create anything worthy. That's screwed up so badly. There are a few percent of population who are inherently good managers, there rest shouldn't even try. On the other hand, while I was attending bullshit meetings and writing bullshit reports, something bad happened to the technical expertise in the programming world. It's really difficult to hire a senior developer these days. The education mostly became oriented to learning patterns instead of thinking (and those patterns mostly concert integrating someone else's code, not writing your own, for some unimportant applications). So few years ago I switched to writing code as well. So long management nightmare.
@@InconspicuousChap Thank you for this comment. You really hit the nail on the head. Management is a massive waste of time for engineers. Many companies/IT departments these days are run by people, who don’t understand technology and got zero interest/passion for it. Sadly, too many developers are tempted by money/illusionary sense of power and join them. My 8 year management journey left me burnt out, disillusioned and cynical. I went back to programming and it took me awhile to recover. I never looked back.
Its funny what Cutler said about Gates saying they need to be nice to competitors. When I worked at Microsoft around 1997 I attended a sales meeting in Montreal and Ballmer showed a video of 2 kids in a dark alley wearing Oracle t-shirts vandalizing with spray paint a poster of either Windows NT or SQL Server. Then a super hero type guy came out wearing a Microsoft costume caught them and sprayed them with machine gun fire. It was pretty shocking.
Dave P you are a tremendous interviewer. Lots of space for the person to answer and knowledgable about the subject. I do hope you continue this as a series.
Agreed. It was refreshing to have the guest do nearly all the talking. So many other tubers do at least half if not more of the talking than the guest. Great questions, too.
I was also going to leave this comment. I’m so sick of interviewers who can’t avoid filling the interview with the sound of their own voice. The way to do it is what Dave did here: frame the story and then get out of the way and let the guy your interviewing tell it his way, and don’t feel like every pause is an opportunity to remind your viewers why you’re important. A couple seconds of dead air is OK, if it leads to another great story from the person you went through all this trouble to get on your show.
DP knows his audience, and I suspect knows this great interview style from having to live it on a pretty much daily basis with his autism. I don’t know if I’m on the spectrum, but I am introverted, so I know all about verbally lining up the shot, and then stepping back so the other guy can take it and play the rest of the game.
This is such a historical interview and I hope it is saved and available for future generations to watch for decades to come! Thank you!
Just because you said it, I'm ripping and achieving it myself.
I am making dvd with it😅
Word of advice: if you find any valuable video on youtube that is worth rewatching from time to time, archive it yourself. Just remember: in 2023 december, google starts the inactive account purge project. If you know any channel, whose owner deceased (for example during covid), and hasnt logged into its google account in the past 1 year or so, now these videos will be the first candidates of the deletion. Search after it yourself, if you dont believe what I said. Thousands or millions of old videos will be gone forever very soon. Im not exaggerating.
Agree.
Oh my god, Dave Cutler was my hero back in the days, when I got my hads on leaked windows nt kernel code, reading his design docs and seeing his name on top of source files felt like finding treasures!
As someone, who spent most of his career on "the other side" of Microsoft (first Novell, then the Linux world) and who would never consider working for Microsoft for some of the tricks the company pulled on products like DR DOS (an ex colleague of mine was a witness in the trial resulting from that) - I must still say: Dave Cutler has always been a name we all said with a great level of respect.
In general, it seems that the engineering side of Microsoft was much better than what the product marketing teams made of it.
On Dave's remarks on PowerPC and IBM making sure they don't create internal competition to AIX: I was at IBM in the 2000s and worked on the introduction of Power Linux systems - and went through the same pains then.
Great to see this discussion! Thank you both for sharing.
He sounds like he was the antithesis of Allchin... but i'm just guessing.
He has to diss PowerPC because he is a "Wintel" guy. With all due respect (which really is due for Cutler) saying an architecture is bad because IBM cannot produce a decent compiler is - well, what do you expect. Microsoft was always married to Intel with AMD allowed to play along. PowerPC was a neat RISC architecture beating Intel/AMD to 64 bit (which was much cleaner on PowerPC than the 64 bit hacks of AMD) and had a vector unit (Altivec/VMX) before Intel even thought about it (i did some Altivec stuff and would choose it over the horrible SSE any time). They could just have used GCC or the Freescale compiler (PowerPC was a collaboraton of IBM and Freescale and is used in embedded devices until today). It wasn't that IBM did not want to "cannibalize" anything, it was that any flavor of Windows shouid commercially run on Intel and Intel only. It's all politics with M$.
@@therealmccoy7221 well, as indicated in my original comment, I was at the receiving end of IBM protecting their AIX revenue at almost all cost (knowing the margin on those systems, that clearly is a sensible short term measure).
I cannot say what they did with PowerPC, but I can certainly see them run much the same course there.
@@pjakobs well they ironically delivered the PowerPC CPU's (Xeon) for M$ Xbox 360 later on.
@@therealmccoy7221 sure, and the market overlap for an xbox and a power AIX system is exactly 0% - which is not true for the overlap between AIX systems and Windows server (or in my case, Linux servers). What I described and what I believe is similar to what Dave C described, is the power the AIX team held over the power hardware.
I made no statement about the quality of the architecture.
This channel is ending up as an important history museum in its own right for a particular era of computing, love it.
I don't care whether it is the year of Linux or not, or if Windows wins the 'OS battle' or if 'MacOS remains superior'. I just have so much respect for legends like Cutler, Tevanian, Torvalds, Thompson... those folks who built that Apollo Guidance Computer... (ah yes you too Dave! Coding in ASM during winters is fun), that being alive at a time when I can watch interviews of these legends is a pleasure, is inspiring and makes me grateful for the foundation that has been laid down by these giants.
Thanks Dave for interviewing Cutler and uploading it here on RUclips. The fact that this is 3 hrs long shows how your passion won't be subject to viewership statistics (a few months back CHM interviewed Ken Thompson and it was about his chess machines... no doubt an interesting interview... But I wished that it was as extensive as the one they did for Donald Knuth which was in two parts and each part being 3+ hrs long).
Once again, thanks Dave!
I also say thanks to the men at bell labs who invented unix , without said unix they wouldn't have been able to create Linux
It's interesting to see that his whole career started on essentially doing a job that he lacked the applicable formal education and experience to do, with the way it is now, he'd likely have never been given the opportunity to even start where he did at DuPont on the Scott Paper project, nowadays every HR department would look and go, "no relevant experience," "no relevant formal education," and promptly file your application into the rejected bin and wouldn't allow such a position change to occur.
Did you catch the part where they sent him off to IBM to learn GPSS III? There were no formal degrees in this stuff back then. I suspect he got sent because no one else wanted to go or be stuck with dealing with that end of the computing business. I assume that DaveC was noticed as a go-getter and was picked for that reason.
But your point that HR would filter him out has some merit. When I left MSFT, the bar was basicly a masters degree in CS which I think most of the people who worked on NT 3.1 wouldn't have met.
This is typically the advantage when things are brand new and essentially no one knows anything about them. There were plenty of jobs when he started that rejected people with no relevant experience.
One of the problems faced today is everyone is familiar with software and all software is just software. To many there is no different between an iOS app and developing a new LLM.
While I agree with what you’re saying in regards to what would happen today, in fact back then he did have the requisite skill set. Math and physics were his minors or at least study subjects for him, I forget if he said he degreed in them or not. So he had what would have been the appropriate skill set back then. I doubt there were any computer science degrees being offered at the time.
his recollection is so good; he can remember details from so long ago easily... he's truly remarkable.
How cool! I hired on to DuPont as a Chemical Engineer in 1968 and programmed on two of the machines that Dave Cutler mentions... Univac 1108 (I think he went from Bunker Ramo to DEC without mentioning that the 1107/8 were Univac computers) and the PDP-10 at the Ex Station. I never met Dave (but wish I had) as my programming was exclusively batch.
Univac 1108?? isnt that the one that executes fortran instructions directly from machine code? that was one of the all time best computers ever made in human history, a friend of mine used to repair those when they were at their peak.
@@bradleyr4451 Oh man... this is taxing my memory. The only thing I remember about the 1108 Fortran compiler was that it had very good diagnostics. I recall compiling Fortran programs on the 1108, correcting any syntax errors and then recompiling on the local IBM 1130 computer (which had very cryptic diagnostics). Also, it had drum magnetic storage (Fastrand, as I recall).
@@amendegw thats so cool! Id hear about that amazing machine, top of the line, incredible. I heard one where they were running a batch program and it was instantly done, like 300 milliseconds or? and they said there must be a mistake it should take longer to run, do it again... and again.. and the station controller came out yelling I have other batches to run what are you doing? and they said its finishing when it starts and he said it runs Fortran natively! I heard that a few times and used to wonder. Yes drum magnetic storage they worked on those, heard how one was so big and heavy it would move and they had to bolt it down? Young ppl would be in near shock to see what memory was. I worked on the CDC storage module drives and Gould/Encore superminis and later disk certifiers, so much precision electronics and science went in to all of it; Id really like to hear anything about those big systems
Fabulous and fascinating history! Thank you, Dave P for thinking of doing this and making it happen. You did a great job interviewing Dave C. Dave C’s departure from DEC, where I worked at the time, felt like an earthquake. He was, and is, one of the best.
brilliant interview, I have watched the clips and now watching the whole thing. amazing. I have programmed PDP-8, PDP-11, VAX, CP/M, DOS, Xenix, OS/2, Windows, and Linux and although 15 years shy of Dave Cutlers age I am still programming. So all this history really resonates with me.
It was posted 30 mins ago?
Started my career with VAX/VMS in 1990 before moving into LAN networked PCs with NT 3.51 & 4.0. Probably the most enjoyable decade of my life!
Did you use to use Pathworks like we did to get the PCs talking to the VMS machines?
@@JonathanMcCormack omg, had forgotten about pathworks, installed on clients from floppies. was a bit kludgy but did the job and it seemed magical.
@@JonathanMcCormackNot OP, but I did!
@@JonathanMcCormackyes we did use pathworks but with DOS and Windows 3.11. A mainframe arrived later on and SNA server on NT 4 replaced the VAX and pathworks setup
From a technical standpoint my time with Digital, the VMS operating system and all the networking around that was the best time of my career. I just never learned so much so fast. When things expanded to Pathworks on the PC's, Alisa Systems on the Mac's (remember the Mac originally only had Appletalk), SNA connectivity with JNET on the VAX, Ultrix, DEC X.25 gateway to that thing called the Internet...it was just the best 10 years I ever spent.
I do love the touch of having the PiDP11 in the background running RSX11M and showing the idle light pattern animation
Thank you for noticing! I was wondering if anyone would be able to spot the idle pattern from RSX11m, and I think you're the first to do so, or at least say so :-)
@@DavesGarage Dave, I also noticed the PiDP-11 running RSX-11M+ in the background behind Dave. It is likely running the RSX disk image that I provided to Oscar for the PiDP-11. There is an improved RSX disk image available from Johnny Billquist who created the TCP/IP stack and utilities for M+ in the past few years. I was wondering if Dave Cutler was aware of the renewed interested in RSX11M+ that's happening because of the more than 4,500 PiDP-11/70's that have been sold?
Dave Cutler is the preson who can use "Don't cite the deep magic to me, I was there when it was written" to more scenarios than any mortal man.
* any _other..._
(Maybe.)
This is the only 3hrs video I watched full on RUclips. It’s better than Netflix. I remember to run the payroll systems of a company I worked on, on a Windows NT. That thing was never turned off I believe 😅. The amount of details Dave remembers is crazy, I am 48 and I don’t remember the model of the servers or systems I ran 20 years back. Thanks for such great interview.
I'm glad someone finally sat down with Dave Cutler and asked him in-depth operating system questions. It's really interesting to trace the origins of the thought process that went into WinNT's design.
Dave Cutler looks so young for his age. Really glad to see he's doing well!
OMG….. 3 hours this is awesome!!!!! Thank you both… I was in building 27 working on Win95, IE, Visual Studio and ran the internal IPTD web presence under BradSi (Silverberg). Best job of my life…. Was great hearing Jim’s name mentioned, I was the person who smart mouthed him back on email on our building 26/27 DL BEFORE checking who he reports too…. LOL… I had no idea Msft Security had flash bangs and OC spray, 😂
I also worked down the street in the DEC building for Compaq (I still drink coffee from my DEC Alpha coffee cup) so this entire video made me so homesick for two companies that USE to be the greatest places to work on the east side…. Miss that part of my life so much… well maybe not the OC. 😉
I had a copy of Visual Studio 6 standard when I was a teen, lol. Also had VB4. :p
Why was the C++ dialog editor so awful? Lol
Would love a second round with Dave Cutler talking about his Xbox work and some of the stories around it. Maybe that’s not possible due to it being more recent work though.
Was he involved in XBox? The story I heard as that they 'took the Windows 2000 source code and hacked it' to make the original XBox OS. It sounds like the sort of thing Cutler would be very dismissive of.
Rumor has it that he was brought in, or maybe volunteered, to step into the Xbox project a while back to help save the Xbox One virtualization layer for Xbox and Xbox 360 backward compatibility. I’m not sure if he was heavily involved before that, but I believe he’s been working on Xbox stuff for a while now…but I could be totally off base.
He talks a bit about Xbox at around 2:00:00 in the video. Didn't know that all games are an actual VMs complete with its own OS.
Thinking of that, if that's the architecture, it wouldn't be hard to run Xbox games on PCs. 🤔
@@rgl168it’s not quite it’s own full OS for each game, but the game application itself runs in a sandboxed virtualized environment, so it can only write to its own dedicated save slots and it’s memory space is isolated from the OS. It’s based on Microsoft’s Hyper-V technology.
This series is so damn awesome. Is there any chance you could get Mark Russinovich in on this OS history? From DEC times through scaring the integrity of Windows, to a MS employee to his work on VM's and containers, pre Docker, and onward into Azure?
Now that would be an awesome episode!
+1 for this. Azure is my favorite Public Cloud platform and I would love to see the CTO of it sharing his old days.
+1 for Mark interview. Though I think Mark's work is not as close to Dave's interests.
Add plus one for russnovich interview.
Oh man what a privilege to be able to listen to this legend telling his stories. When I started my electronic engineering graduation, MS-DOS was at version 3.3 and I’ve had never heard of Windows. Still today I eventually need to launch a DOS VM in VMware to solve specifics problems on very old machinery. A few months ago I still had a Fidia (Italy) milling machine running on Win NT 4.0. This interview brought me a lot of memories. Thank you very much Dave, both of you, for sharing this with us.
Great interview. Please thank Mr. Cutler for being so generous with his time and sharing his memories.
Thank you Dave for just sitting back and letting this guy talk. I could listen to his stories for hours! I was so happy you didn't rush him along. You did an awesome job with this interview. I want more stories from this guy!
Thanks! I'm a good listener, and he was interesting, so it was a good mix :-)
If age has taken a toll on Dave Cutler's mental faculties, it either doesn't show or he he's lost more than some people have at all. I'm 35 years younger than he is and probably not half as sharp.
Very precious, part II please! This is easily one of the most interesting and capable programmers on the planet, I wish I would see more content of him!
Nice, the full episode! Thank you Dave for interviewing Dave Cutler!!
You guys are amazing
two programming giants in a room talking shop has to be my favourite format for content. please keep this up dave!
please have him back and talk more about windows internals. More about azure internals. More about the kernel. More about drivers.
I read "Showstopper" many years ago, and I really enjoyed it. And here's the man himself. Thanks for a fascinating interview.
That was a great book.
Excellent Book. It's where I first read about Dave Cutler in the 2000s.
I got the impression that Dave Cutller was a kind of a scary character from that book but he seems so approachable and a fun guy in this interview.
@@StuartWoodwardJP - I agree the book did certainly give the impression he was scary. I am sure he was when deadlines were looming!
@@StuartWoodwardJPI’m pretty sure there was a scene in that book where he punched a wall at Microsoft and left a hole in it. He across as hardcore, but was 1000% committed to quality
Thank you Dave and Dave. I'm just a few years younger than Dave C, and I spent the first seven years of my professional life at IBM in their OS and complier System Test / Quality Assurance teams. So much of what Dave C said about fixing your own bugs or letting the next team "down the hall" do it are so true. Your discussions brought back so many great memories.
Love this! Would be incredible for it to become a series; maybe Raymond Chen next? ❤
Listening to this interview was a blissful experience; having spend 5 years writing Assembler for PDP11 on RSX11S, having done some work on VMS and then substantial work in Windows NT and later versions this interview touched so many points and brought me back to many amazing memories. I am impressed about Mr Cutler and his amazing memory, recalling architectures, bit width and so on. Dave thank you for sharing this amazing content and, of course, for your whole channel
Having worked for CSC, I can't imagine the days when they had people who could write compilers, and I went to a college where a lecturer had written a FORTRAN compiler for a 1961 computer, just to prove it could be done.
On punch cards likely too!
Nope. 5-track paper tape. The machine was a Ferranti Sirius. Our college had 4 of them at one time, from a production run of 20. A store room also held a ZEBRA, with magnetic drum as main store.@@DavesGarage
It is very nice to meet Dave here in this meeting as I was AFAIK the first and only user of RSX-11a in Europe. Was some 50 years ago when I used a PDP-11/10 for controlling a high-voltage switchyard with automated switching sequences. With several modifications, this program ran under RSX-11m/s until all 'experts' got afraid of the millennium bug and did not know how to deal with a PDP-11 assembler program.
BTW: I also modified the RSX-11m terminal driver for the VT30 color monitor controller. So I was programming on a color terminal already in the late 1970s. We needed it for the display of the high-voltage switchyards and controlled them by joystick and keyboard via (very slow) data connections to remote PDP-11s.
Thank you Dave, this did not disappoint. Great insights and interview.
Glad you enjoyed it!
@@DavesGarage This was simply phenomenal.
Having a boss like Dave Cutler must have been great, I totally get why you worked at Microsoft. Imagine working for someone who understands the industry hes managing people in. You probably didn't have to spend half your day trying to explain what you did in a way that a manger wouldn't think was you being a wizard with a magic wand.
I was blessed in that I came in at a time and place where I was surrounded by people like him. The learning curve was steep!
@@DavesGarage I have dramatically changed my opinion of Microsoft and its management for the better from the two Daves.
@@robsyoutube Cutler had a reputation for a nasty temper and punching holes in drywall at DECwest.
@@robsyoutube honestly don't judge a book by its cover.
@@MultiPetercool As long as they value honesty, and can admit faults when presented with the evidence, those are the best people to work with. For the simple reason that you know there is nothing stuffed under the carpet. If you don't get away with it, others don't get away with it. It keeps politicians and lazy people away. It can be tough sometimes, but it's worth it.
The same with Steve Jobs, he was not always right, but the people who complain about him were mostly wrong.
I was already in the doghouse today, so I spent half the day listening to this and fixing all the broken things around the house. Loved this interview. Dave C is an absolute legend.
Love how these guys contributed so much to our society
like what
@@shallex5744 Listen to the video
@@shallex5744operating systems, hardware, interfaces…what else you want?
@@shallex5744 the guys at Stanford basically built this modern era computing infrastructure
but there's always a moron , I bet you're a zealot of sorts..
tsc, tsc, monolith operating systems...
Tanenbaum was right
it is so heart-warming watching how much YOU enjoy doing this interview. You are rocking the the subs and likes Dave. Thank you so much.
Holy moly. Didn't expect a 3 hour interview! Ok, off to make popcorn.
I would honestly love to see Cutler do a super basic kernel/DOS on something retro and relatively simple like a C64. If for nothing else then to just see how he approaches it. You can learn so much just by watching how someone works and thinks about it.
I just finished reading "Showstopper!" recently, so this interview is totally on-time for me. I heard that Dave Cutler does not like giving interviews, so congrats for getting him to talk for 3 hours.
Almost half way through. Great history. I started my technology dependence with the Tandy Model 1, Level 2 16k with tape drive. When I started working for a Radio Shack Computer Center, 3.1 and 3.11 were the Windows I became adept at for my customers. I was even selling SCSI hard drives when they were nearly unheard of in the consumer market. Now I’m so old I’ve moved to the dark side for the simplicity. But for more than 30 years, Windows, applications and users, put food on the table.
A lot of DEC folks I knew said Windows NT or WNT was one letter better than VMS like IBM and HAL in 2001. DEC was pinning its success on the success of NT in the later days.
I love the talk at the end about the lack of precision on 24 bit mantissa's, i worked on the Ferranti Argus computers they ran the bloodhound missile guidance system and a lot of other safety critical systems, they had an 8 bit mantissa and 4 bit exponent ( memory was expensive ), the lack of precision was laughable, you would type 2.41666 into it and reading it back it would come back as 2.42000.
That is odd. The PDP-8 had 12-bit words and they used three of them for all floating-point numbers. Well, eventually, 48 bits for FORTRAN 4. 8 bits of precision does sound awfully meager.
Analogues were 16 bits in total, 8 bit mantissa, 4 bit exponent and 4 status bits. 4 status bits was extremely generous but when talking about real plant signals were critical.
This interviewing was wonderful. Dave Cutler is such a legend.
I could have listened to 3 more hours of this! Fascinating! I agree with the other comments. Bring him back for another round!
Incredible inspiration. Thanks for the interview Dave. Would really like more of this stuff by people that defined our early computer lives.
This was a fantastic interview. Cutler is so frank, honest, and has many great stories to tell. Thanks to both the Daves for this one!
Glad you enjoyed it!
I’m a DevOps guy. I’m adopting the term FCIB (described approximately at 2:00:30) and using it this very Monday. Might even make a meme out of it. Love it ❤
Table keeper guys then
Cutler never seems to age.
3 hours long - thank you !!!!
Dave Cutler is on par with Linus Torvalds.
Dave is surprisingly smart and well spoken for someone in their 80s. Doesn't seem like he's slowed down cognitively much at all. I wish we asked more lifestyle questions of these great achievers... Perhaps we could all stand to learn a few things from the habits they employ.
Some people lose a bit of memory and mental sharpness through their late 60s and 70s, but quite a lot remain 100% sharp. I believe there's research showing that staying engaged with challenging projects bends the odds in your favor.
Thank you, both Daves for this fascinating review of compute history! Best spent 3 hours in a long time.
In the early days I ran NT 3.1 Beta on a Dell server that was purchased to run Novell Netware 3.12 or so. The interview section on NT4.0 prompted me to build a new VM on my VMWare homelab, install NT 4.0 on it from the ISOs, apply SP6a, then VMWare Tools. All ran without a glitch. Network is up and running now, on less resources than my smartwatch. All this before the interview was over ;)
Back in the days we ran out entire company on that. I agree with Dave C. that it's sad to see "Hello World" take up 1MB of code. Current software isn't coding, it's glueing libraries together. No wonder that bugs are harder and harder to track down.
He’s still actively contributing to the code, I’m honored to be able to work on the windows kernel which he firstly worked on decades ago when I havent even born:). All the design he made and the code he wrote is just a masterpiece, which is incredibly still working the same way that he did in 2024! (You will know what Im saying if you are the engineer, and understand how impressive is this).
He’s always my role model in my whole career, and will always be, to help design a software as perfect as possible.
Wow. I’m more of a UNIX guy but love this interview, great to just hear Dave Cutler talk!
Odd, some audio is missing, such as 2:06:58?
Looks like my video recorded drops a second of audio when restarting a new file every hour or so... sorry about that!
Thanks for the explanation. And here I thought it was censoring top-secret Xbox trade secrets 😅
Absolutely amazing interview, I was mesmerised for the full three hours!
He is such a legend and I would love Part 2 focusing more on his work at XBox!
I could listen to you guys talk forever, you have to have him back on again!
This was a great interview. I've read Showstopper about 5 times I think and am fascinated with what Dave C & team was about to accomplish. And am blown away that it's still the basis of Windows today. I would've been interested to hear what Dave C's thoughts are on AI and writing code. Thank you for doing this interview!
DaveC’s thoughts about AI and ML would make a fantastic follow up to this interview, if he’s been keeping up! (Not easy for anyone to do given the constant development in the space)
This is the most amazing IT podcast I've personally ever heard.
Dave, the iAPX432 was Intel's biggest clusterfuck, ever. Eclipsing the i860, and that was a big one too.
From a programmer's perspective, it was a CPU specifically designed for Ada.
The application programmers hated it because they didn't want to deal with Ada.
The systems engineers hated it because it required half a dozen chips to get even the most basic system running.
The system programmers hated it because a subroutine call took hundreds of microseconds at a minimum.
And what ultimately killed it off was the fact that the iAPX286 was available, which ran twice as fast, for half the cost.
Dunno if people really understand this. Intel's track record for new processor architectures is beyond shit.
It's easy to say this with 50 years of hindsight, but you could argue that Intel were taking some of the directions the rest of the industry were making to their logical extreme. It borrowed heavily from the mainframe that programmers preferred to use - the Burroughs 6500 - and the direction the industry was going to impedance match with high level languages. Don't forget that the first RISC paper really tried to optimise call/return with that register windowing concept.
It's more that Intel have repeatedly failed to understand market forces: nobody wants a multi-chip implementation of an old mainframe in the 70s, and when they finally got it right in the i960, they refused to sell the full featured version outside of the military.
Itanium was a similar problem, it wasn't affordable enough to make up for the fact it was on a dated process node and didn't run existing code at a half-decent speed. The worst part was that Intel made it difficult to develop for, by failing to drop the price. If I'd been able to pick one up back then, I'd have been able to write some good quality jit. The architecture is not bad - it's no i960 - but it's still the nicest modern architecture we have.
I don't know about "nicest". In some theoretical sense maybe. But nobody has figured out how to write a compiler that generates good code that fills out those long words and runs very parallel. A few theorists think it's impossible to do so. @@capability-snob
@@georgegonzalez2476 the hard part of writing fast code on the itanium isn't filling out the bundles, it's using the prefetch and speculative machinery effectively across linkage boundaries. ia64 has a lot of great memory-latency-busting features, but since you've got some 1200 instructions between a load from main memory and the first time you can use it, you really want to move prefetch long before your function is ever invoked. That sort of supermodularity is possible - especially with whole-program compilers - it's just not how most compilers are written.
More please.
Dave's reminiscing brings back so many memories from my own work history. I remember working with a number of the things he talks about ... but I was very young man at the time and he got there well ahead of me. Plus he's obviously held up better with age - I seriously doubt I'll remember all those details when I get to 81.
He obviously deserves all the success he's had.
So much great and relevant history! That's what I love about this channel. Nothing compares to learning from the people who were actually there and built the new things at the time. Thank you!
Met Dave Cutler once at DEC. I joined in 1966 as a hardware techi from the UK. PDP 7-8 8i,8s-9. Fixed them and ran courses for customers. Thank him for the interview and wish him well.
Have you any interview which would give a similar overview of apple software development?
Visited the Swiss Super computer center last week. Quite an eye opener.
Thanks
Chris
I loved Dave Cutler's Language (DCL) and VAX/VMS. Microsoft's batch files were such a letdown until PowerShell came out.
No wonder we got the small bites. 3 hours! Haha that is amazing though THANK YOU for giving us this long form content. I watch a skateboard podcast that puts out interviews between 2 and 6 hours every week, and have been for years. There is still interest enough for it among the want for shorter and shorter stuff.
Thank both Daves for the video.
"I like to tell stories" - and doesn't he tell interesting stories, and tell them well! A great listen for those of us with any history in IT. Current programmers ought to listen to this; they will gain great insights into WHY things are the way they are.
This is an incredible interview! At first I kept reading 3 mins instead of 3 hours 😲. Thank you for this absolute treat! 🙏
I joined Microsoft the day Windows XP launched in the Windows Division and Dave was one of the first titans I visited at his office in Building 26, 3rd floor. Thank you Dave for everything you've done for Microsoft!
A software engineer's software engineer. His passion for the trade is reflected by his command of historical details. This interview will go down as a classic. Good job.
It was great to hear from Dave Cutler what really went on with the versions of windows from one to the next. Some of us in the industry were close with our guesses obviously but it is fantastic to hear the details from the expert himself. Thank you and I hope he comes back for another chat session. I hope the linux work he is doing also gets fed back into the GPL community.
Thank you Dave, for an AMAZING episode. Dave Cutler's passion and insatiable curiosity are so obvious. What a treat of an interview!
Odd end to the video.
Also a couple of times the audio cut out just as Dave was about to comment or reveal something.
Interesting career history.
I think he cut out cuss words 😅
In my CompSci coursework, he is referenced for VMS, but his wealth of knowledge is far more encompassing than that.
This is a VERY valuable presentation of our field's beginning, and how far we've come...
Fantastic interview Dave, thank you. You mentioned toward the end a rotary phase changer your neighbor has.
They are extremely simple, it’s a 3phase motor running on single phase with the extra leg from the motor generating the 3rd phase.
In my younger years I built several of them for friends and relatives.
I was surprised it was so "easy", but it worked great!
Seeing the amount of detail you have posted with regards to Windows and the O.S has had my attention. In watching this conversation it has been very cool just to hear.
My question though, is why the mic was muted @ time 2:06:55 while talking about the X Box, virtual machines and the game packaging ?
3:30 “There’s a little side story here too. I hope this doesn’t go on too long ‘cause I like to tell stories”. Ummm…. Not a problem: we WANT the side stories and we will TAKE the time.
Amazing interview. It took me a few installments to watch the full thing, but it was worth every minute. The detail in which Dave C. recalls this stuff is extraordinary.
in the early 2000s I used to come into the office to the kitchenette and listen to DaveC and Jim Allchin argue about the topic of the day.
If you are wondering WIM = Windows Improvement Meeting or Windows Integration Meeting depending on who you asked. Happened every week or so on Friday afternoons, especially when we were close to milestones. Lots of food, beer, and friends. Those were good days.
RIP Bld 26 cafe...
Thank You Dave. I have been playing around with computers since the early 80's but I never got as good as you are.
Please have more of the people of years gone by, while they are still here to teach.
Wow Dave. Thank you for creating this content and contribution to the OS’s I have used. Excellent interview. It really took me back to when I first got into IT. I have not heard the terms PDPx or VAX/VMS in 30 years. And thank you Dave C. for your contributions to all the operating systems you created or had influence on. This interview answered a lot of curiosity questions I had.
One of the best tech interviews to ever happen on this platform. Incredible wealth of knowledge present in only these two humans.
Thank you Dave for this video.
Reason why the internet is something great.
Been a fan of the lore of DC since the 90’s but this is the first time I’ve heard him speak.
Huge thank you.
I played with Vax in the 90’s, kilostreams, pathworks and Novell, before getting my hands onto NT which changed everything for us and shaped my career.
35 years later - semi retired myself, yanked back now and again for problem solving for issues that your average ITIL and TOM structure prevents from being identified.
What a legend and coincidentally the same age as Joe Biden. I hope young kids watch this and choose a career in engineering and not in politics :-)
It's great to see these stories preserved for posterity. So many amazing insights into the history, business, and engineering of computing system.