In the 1980s I had the pleasure of working for IBM in Austin on AIX, and it was an amazing experience. It started as a very small team of very motivated people, and I learned so much. Had some contact with Larry Loucks, Charlie Sauer, and Glen Henry, and these guys were on another dimension. I thought I was smart before I met them. I could not believe I was getting paid to do the work I was doing. I sure miss those days, and it's a delight to see some of the work that was done by this talented team of architects and coders lives on today in current versions of systems like VMware and Linux. We always had a friendly rivalry with the OS/2 guys, and wished that we had even 10% of the resources that they got to make AIX better. Thanks for bringing back some memories!
G. Glenn Henry was the software designer of the IBM S/38. The 38 then became the AS400. Great operating system. Highly secure. The system came with an integrated database.
I was at Ford motor a bit later where we ran AIX on Power5 - the shit was way ahead of anything I saw elsewhere at the time. The ability to carve up hardware, I thought, was ingenious. There was a number of IBM folks working with me to either port COBOL or make COBOL apps work on Linux and AIX on Power5. I think the name of the IBM group was called system7 or sector7 or something like that.
@@FLY1NGSQU1RR3L76 Definitely knew Phil Hester - he was the top guy the HW side, and I didn't know as many of those folks. The name Peter Woon sounds familiar, but I can't place him. There were a lot of very talented people on that team!
I built the London Stock Exchange system TOPIC2 on OS/2, with 4 IBM RIC co-processor cards which supported an incomming 64kb SDLC line with 32 * 19200 outgoing async lines to the Topic terminals. This was circa 1989-1990. I went to MicroSoft University at Seattle for a week to learn to write OS/2 device drivers etc. What pissed me off is that we were thrown out of the labs at 5PM. I was there from the UK, what else was I going to do. Got to see the aftermath of Mount St Helens which was pretty amazing. Writing server processes in C for OS/2 was great It worked well and was pretty fast 25Mhz 386, (for those days).
Back when Microsoft was calling Linux a “cancer”, it made a big noise about the London Stock Exchange choosing Windows Server over Linux. Even put out an ad featuring a fictitious newspaper, the “Highly Reliable Times”, touting the story. Then that big, embarrassing crash happened. The Windows Server systems were ripped out and replaced with ... Linux.
At a small St. Louis company, I had a similar experience with PS/2s, IBM Unix, and IBM RTIC (not RIC) cards to run inbound call center management software for BM/Rolm telephone switches. Got a commendation from our IBM customers for technical leadership for the design, and (I suspect) using all IBM components in the implementation. We didn't tell IBM that the RTIC card was running MP/M (multitasking version of CP/M) :-)
I worked on OS/2 for about five or six years from 1991 after a year or so working on Windows 2.1 & 3.0. OS/2 was massive in banking, as were things like MCA (microchannel architecture) and token ring because "nobody ever got fired for buying IBM". The OS/2 and Presentation Manager API was always far better than Windows from a programmer point of view just from a naming convention PoV. Then we migrated everything over to NT 3.51 over a couple of years. Biggest problem in the field in a data center was that you rebooted an OS/2 machine with CTRL-ALT-DEL, but you needed to use CTRL-ALT-DEL to get up the login screen in NT... what could possibly go wrong? So it wasn't unusual for someone to inadvertently reboot a production OS/2 server when they were actually trying to log in in a hybrid data center environment.
Yes, that was a big mistake in NT. If they were starting from scratch, then using CTRL-ALT-DEL to reset the login would be perfectly fine, but they ignored the legacy use.
@@wictimovgovonca320 wasn't it a deliberate choice to stop people exploiting dos/windows ctrlaltdel shanaggingins to exploit login bypassing? its 30 odd years ago, so i maybe remembering incorrectly.
I agree that the OS/2 APIs were much nicer than the Windows API. More consistent, more clearly named, grouped into subsystem functions by prefix. A bit more cumbersome to type the function calls but the code was clearer. You can see a bit of that in Windows today - ISTR that the registry functions in widows was made by an OS/2 team.
I was outfitted with a DEC Rainbow at work about 1984. The Rainbow came with CP/M and MS-DOS 2.11. I recall the Microsoft documentation talking quite a bit about, “MS-DOS is a stepping stone to Xenix.” There was quite a bit in the documentation admonishing to design application programs with user interaction in a way with transition to Xenix in mind. I guess that ship sailed pretty early on…
Back in the OS2 days, I was briefly responsible for a small IT group. At the time, our customer support teams had to have as many as 3 terminals to access the separate order, trouble and billing systems. The IT wizard came up with a way to have an OS2 platform to act as the access device for the three systems and allowed the service teams to access a comprehensive customer record with one terminal, one entry of the customer ID. The corporate IT people were incensed at being upstaged by a 3 person group in the field.
Ah, yes... the NIH syndrome: Not Invented Here. Many corporate IT shops develop a "vested interest" in their own products without regard to whether it was worth a toot.
@@sdrc92126 One nice thing about OS/2 was that it was great at multitasking, even with DOS & Windows 3.x apps. You could even install real version of DOS on it, of as many versions as you wanted. You could even install CP/M-86, IIRC. Even today, I find Windows 10 multitasking to be poorer than Linux.
I read Solomon's “Inside Windows NT" when it first came out. Cutler's original Windows NT was quite a 'purest' design and implementation with a small (but not micro) kernel that ran with ring 0 privs. Even the graphics sub-system ran in unprivileged space to protect the kernel. Unfortunately that hit performance and they rolled the graphics into the core in NT 4.0
I also liked it as it originally was. In 1993 I started writing device drivers for the NT beta versions, through to 3.1, 3.51 etc. and it was rock solid. I could reload crashed drivers without rebooting. Impressive. Then 4.0 came along and BSODs became part of life.
Yeah good read in the day... NT 3.5 (Daytona) moved some things into the kernel for speed thought graphic driver made the jump then. Dave's original design that is 3.1 was let down by hardware of the day all hardware via the HAL..
It was discovered, I think around the time of NT 3.5, that the difference between the “Workstation” and “Server” installs of NT came down to a simple registry key: change that in your “Workstation” install, and suddenly you had “Server” capabilities, without having to pay extra for them. Naturally, that no longer worked in later versions.
00:13 📊 Microsoft shipped various versions, but only 3.1 gained traction, selling 16 million copies in six months. 01:04 💰 Microsoft negotiated a volume license with AT&T for Xenix, selling sublicenses at a profit. 01:44 💻 DOS was chosen over CPM for IBM's first PC, later evolving into a program loader. 02:35 🤝 A joint design with IBM for OS/2 (16-bit) and HPFS file system faced bureaucratic hurdles. 03:33 💡 Bill Gates emphasized the need for a RISC-based processor for the future. 05:31 🌐 Microsoft aimed for a portable system to run multiple environments, including OS/2 and POSIX. 06:40 📚 Initial plans to customize system services for different environments faced challenges in implementation. 08:02 💻 Windows became the main environment for NT due to its widespread acceptance, surpassing OS/2's presentation manager.
DOS didn't evolve into a program loader, it was essentially that from the start and apps had to provide their own I/O code, as the support for it in DOS was essentially non-existent.
Great to see at 80+ Dave is still sharp. I spent many years working on his babies. When he left RSX-11M, I joined it. When he left VMS, I joined it. When he left Windows, I joined it. When he left Azure (Red Dog), I joined it.
I wqs a Warp addict from the start. It was so stable, compared to the early Windows distros. Then We all had to 'move on' as you say. Today I'm on Linux, Open SuSe leap 15.6 !!
I started GUI programming in 1993 and windows 3.11 and MFC and C++ it was a revelation! It was such a nice architecture and I coded and made the deadlines! Our boss wanted us to use NIHCL that was a class library that was free and used by the universities but I didn’t like the idea because integrating it to the Windows ecosystem would have taken ages. We invented something that was later to be called ”rapid prototyping”, we made a mockup of the GUI and showed it and got good fast feedback. I really miss those days.
Those were the days when one guy could write an application that made serious gains in productivity. I developed a web app in the late 90s that provided huge productivity gains to he employees because it integrated a few other systems and reduced triple data entry to single data entry as a byproduct. It generated all of the documents needed for the customers and fed data to the accounting system. It also used email, fax and printing for snail mail as an integrated delivery system. My plan was to integrate FedEx and UPS but I left the company for another job at that point.
OS/2 might have nose dived, but when OS/2 Warp came out, it was way superior than anything Microsoft produced technology-wise - and you could actually surf the Internet without crashing on it, while downloading something else - it was true pre-emptive multitasking (just like Desqview did a few years before). Now, IBM being IBM, marketing wise they didn't have it and that's why they nose dived. Many ATMs were still powered by OS/2 late in 2010...
@@MrRmeadowsOnly for OS/2 2.0, that came out in March 92. NT 3.1 came out in July 93. So that was only for about a year. I did love it though. Still do. Even though it wasn't perfect and NT is arguably a lot better in almost every way (except the user interface, Workplace Shell is still better than anything that MS has ever created).
Dace Cutler is right in the point that OS/2 was an absolutely terrible OS... but that changed with Warp. I'm still sad that IBM screwed that up but Dave is right, that company was a bureaucratic mess.
I used Warp at work for an embedded application. It took a few weeks to get it to work right, then it was bulletproof. Never crashed. We used to joke we got all our crashes done up front. I left a year later but they were still using it the last time I talked to them a dozen years later.
The problem with Apple A/UX was that it was rather constrained by the hardware. A/UX was pretty pokey even on what was then a fast 68030 based Macintosh II. Apple did a lot to make UNIX palatable and resilient for an end-user desktop aside from not being able to make it fast enough. I recall that all the AppleTalk stack as well as TCP/IP stack was there, so talking to Apple printers was automagically configured without needing to go in and make obscure changes to the /etc/lpd.conf (or whatever it was). A/UX was a curious mash-up of Sys V and BSD, kind of an opposite mash-up from SunOS. There were also HP/UX, AT&T’s 3B1 Safari PC snd DEC Ultrix that were hybrids which had pieces of both Sys V and BSD. None of those systems had quite the same hybrid mix, so porting applications between them was always a bother. As an aside, I got the place I worked interested in using a 3B1 and VoicePower board to implement a fairly complex IVR system. We did it all in Bourne shell with the user level VoicePower commands. It was pretty snappy in operation and ran for years with minimal need for administration. It is too bad the AT&T collaboration with Convergent Technologies died on the vine and essentially put an end to Convergent Technologies. The 3B1 was rather nice to work with.
Utterly typical of the chronic fragmentation in the Unix world. Basically, every system that licensed the “Unix” trademark is now either dead or dying. Instead of “Unix”, people are now running Linux.
@@wtmayhewI worked at Bell Labs and bought a 3B1 through the employee discount program. I had it running at home with UUCP to ihnp4 with a 2400 baud modem. It was incredible at that time (1987/88) to have email connectivity with graphical interface at home. Later I had an AT&T 630 graphics terminal and ISDN for 128kbps data. Now I have fiber and 5G cellular. Amazing how time flies and the world changes.
@@wtmayhewBTW I remember the VoicePower card and the high level shell features. It was one of the domain-specific language concepts that many years later inspired us to create what became the VoiceXML web standard for IVR applications.
@@kenrehor Thanks for the Reply. The VoicePower card seemed way ahead of its time. It was very productive to be able to script in Bourne shell. The UNIX Bourne and Korn shells are very over-looked by the younger crowd and have a lot more power, especially Korn shell, than most programmers realize.
I think a major problem with OS/2 was it needed stronger computers. I like everything is an object, I like extended attributes, even on floppies, and I liked attaching my modem and CD to machine A and using them on machine B. You might still find references to "john summerfield" "OS/2" I liked cumulative updates, creating a 5MB partition to boot and run the updates process was a dream incomparising with the Windows process. And IBM added support for new technology to old versions, USB for example. When Warp 4 surfaced, the MPTS stack was made available to OS/2 3.x users. Free.
Loving these interviews. Very interesting and influential guy. I was working on RSX way back when and I heard rumors about his legend at The Mill at DEC. I continued on VAX/VMS, MicroVAX/VAXELN, OS/2, and NT. Great to hear so many back stories
God this is wild to me. I've been using Windows since 3.0 days, and I remember the different OS/2 iterations. Always wondered at the people behind windows. I'm loving this series.
Back in the 1980's I worked at Bell Telephone Laboratories in Warrenville IL, we used UNIX for creating all the Circuit Board, NODAL and Schematic Drawings. Can still see the big 32" Green Screen Monochrome Monitors we used back then, was a great time. I also remember Windows 3.1, it was almost a direct copy of the GUI for the Apple Macintosh which I used at Bell Labs as well doing Graphic Arts design. I am a multi talented Computer user working with Almost every major Computer back in the day, Atari, Commodore, IBM, Tandy, Apple. Still own my original CoCo and VIC20.
Love the series. I worked for DEC for 15 years on RSX and VMS. Loved it! Would love to hear Cutler's views on the DEC presidents after Olsen. Seems like they destroyed the company, but I know things are complex and would love to hear Cutler's take on it.
Hey Dave, I love these videos. Do you have (or could you describe) the display and software for the LED display behind Dave Cutler? It looks like most of the bottom has a form of the LED Fire, with the top being a VU meter on Dave's audio. Just curious about the display HW and SW for that (given your love for ESP32 and addressable LEDs). Thanks, John
I missed the whole Windows and OS/2 thing. I started out in 1982 on a TRS-80 CoCo 1 and then switched to DOS in 1986. My first job out of university was on UNIX (SunOS, specifically) and once I decided I was fed up with DOS on my home computer, Linux had become practical to run on a 486. Been using Linux for 29+ years and never felt the need to try Windows.
Hey Dave, this stuff is great and though Cutler couldn't talk about some of the early stuff, there are people probably still close by you who can come talk. Henry Burgess for example can fill you in more on Xenix. I know you may not have direct contact with early folk who are probably recluses like me, but I am sure that you know someone who knows someone who does. If you could interview say David Weise, or Tim Paterson (who wrote Seattle DOS and went to my high school), that would be so interesting I think for a lot of your subscribers. Weise's cool memory trick is a story well known but having him talk about the Ballmer meetings and the situation with IBM I don't think is well known. Weise was a very helpful, bright and I found to be a fellow subversive type guy. I had a compiler whose memory requirements forced me to use OS/2 for builds, but he had a protect mode version of dos that I could make my compiler work on with minor changes and I never had to use OS/2 after it. This was guerrila programming- Cutler was on the "serious OS" side and every serious coder respects that sort of work if your hardware will support it, but MS had to have a market placeholder, and it would be years before NT was ready for consumer machines- and guys like David kept MS in the game. Also on the Apps side. Maybe Simonyi would be too difficult to talk to, but there are some cool stories- like the reason why Excel could run on the 128K Macs and didn't crash no matter what crazy thing Mac added to their operating system. Any Excel former leads can retell the lore. It's a cool story because MS hardly used any Mac system calls. Almost ignored it entirely. Which would give your subscribers some glimse into the competitve/ oftentimes mutually condescending relationship between Apps guys vs. Systems guys during the early years of MS. The excel attitude about the OS basically being a massive dependency to bypass and ignore whenever possible was not just towards MacOS but early Windows 2.x and 3.x.
I was beta on Windows 3.0/3/1 and OS/2 V4 Warp. I liked OS/2 a lot. What was interesting is that when helping people troubleshoot Windows 95, ME and XP I found OS2 commands worked. The divorce between MS and IBM was not nice, like some of the engagement ring was not returned. At the time my family computers were all RedHat and my kids used Win for class projects.
This is may be the most interesting short video about the rich history of operating systems. So much information to unpack. Brilliant. I reposted it to the "Science" page on MeWe.
I was an OS/2 beta-twster, from 1.1 all the way to Warp. I tended to concentrate in networking and network applications. I seen to remember OS/2 having significant performance and stability issues until 1.3 or maybe 2.0. I'm mostly an embedded guy these days, with Linux on my desktop. I really miss OS/2's animated backgrounds, though.
What about IBM charging for TCP/IP for OS/2 2.x? I couldn't believe it, so shortsighted for a great OS to be hobbled like that. Windows for Warehouses 3.11 had the 32-bit network stack from Chicago pulled forward, smoked OS/2 2.1. Such a bummer, and IBM treated 3rd party developers like a revenue stream, while Microsoft gave out developer docs and software by the pallet at trade shows. Cutler knew then what a graveyard IBM would become.
A few additions. 1. MS was a BASIC company. It licensed DOS to IBM without having an OS to sell. MS then bought DOS from Seattle Computer Products for ~$100K and used that as DOS. DOS was a rough copy of CP/M, with some people claiming early DOS versions had CP/M bugs in it (very bad). 2. MS got volume licensing for XENIX. But MS's licensing for DOS/Windows was that an OEM had to pay MS for every computer it sold whether or not it had MS's OS on it. 3. He did not want to work with IBM because IBM was an old established company. So was DEC, where he worked before he went to MS. 4. 32-bit OS/2 went nowhere because IBM was bad at marketing and because MS was anti-competitive. 5. He does not say which version of OS/2 he hated so much. OS/2 v2 was decent and more stable than MS's OSes. Versions 3 and 4 of OS/2 were very good, had a full object model, were extremely drag-and-drop oriented, could do multimedia when few companies were doing that, and had very good preemptive multitasking long before consumer Windows and better than NT. The amount of configurability on OS/2 is still among the best I have seen.
My experiences with OS/2 v2.1 1) countless trap errors 2) dirty power off took multiple diskettes to restart 3) installing took 20+ diskettes, downloading the service pack was 20+ diskette images, then had to copy image to diskette, then feed 20+ diskettes to load the update.
@@BobMar1964 I do not recall having those issues. Installation did not take 20+ diskettes if you had the CD version. The CD version of OS/2 2.1 came with two boot floppies. By comparison, "Windows NT 3.1 could be installed either by using the CD-ROM and a provided boot disk, or by utilizing a set of twenty-two 3.5" floppies (twenty-three floppies for Advanced Server)." So 20+ diskettes for OS/2 v 2.1 is bad, but 22 diskettes for Windows NT 3.1 is good? Got it.
@@georgeh6856 Just relaying my experience/inexperience at the time (1992) with the very smallish office I supported. We did not have any cd drives at the time. We had an IT manager who wanted IBM network.. right down to a OS/2 LanServer on a token ring network. After he left, new management switched to Netware/WfW since it was the market leader with a better support from the vendor. I got out of network administration.
On point 3 "IBM was an old established company. So was DEC". IBM was established in 1911. Digital was estalished in 1957. Bit of a difference. Digital was cool and groovy all through the 60s, 70s and early 80s. There was a world of difference in corporate culture between IBM and Digital; IBM was always rather business-oriented, Digitial was an engineering company, first and foremost. Sadly Digital started to go downhill from 1992 onwards, and eventually it too became a barureaucratic moster, only to be swallowed by Compaq. But that was long after the period Cutler is talking about.
@@Maclabhruinn By 1982, DEC had been around for 25 years. Trying to claim that a 25 year old company is not established is bizarre. By that estimation, Google which was founded in 1998, is still a startup. I would argue that 25 year old Google is much, much more established than Google was, say when it was 5 years old in 2003.
I wish he had asked Cutler if ever thought at the time about just saying to hell with the whole thing and just port VMS to x86 and slap a Windows shell on top of it. A lifetime ago while I was still in school Dave Cutler taught my VMS Crash Dump Analysis class at Digital, that got me started. :)
I was so sad when we walked away from the clean 32-bit Presentation Manager API & chose to go back to the 16/32-bit hybrid Windows API for WinNT. It was much easier to program. Eventually, tools like VB & MFC abstracted us away from it.
Speaking of OS/2.... I was a big fan back before MS essentially killed it. One thing I think was lost was a better file systems. Seems to me HPFS would hold up well compared to the modern Windows file system. Would love to see a deep dive into HPFS vs NTFS. Cheers!
What are you talking about? NTFS started out as an improved and extended HPFS, adding Journaling and other features from the UNIX world. To my knowledge, there's nothing HPFS can do that NTFS can't, and plenty of the reverse. Even OS/2 dropped it in favor of JFS, though that decision was mostly motivated by having to pay Microsoft royalties for the improved HPFS driver.
I worked as a dev on OS/2 for IBM and had MS devs on the team. We thought they were hackers and they thought we were bureaucrats as dave says. I dont agree with that sterotype, there were very good engineers at IBM. IBM saw all the issues in NT and decided that it was another DOS which was a bad mistake. I argued long and hard with senior people in IBM that we had to take NT seriously but got no traction so I joined MS. That was a huge shock.
"I dont agree with that sterotype, there were very good engineers at IBM. " - I don't think he doubted that. What he seemed to admonish was probably more the strangling atmosphere of a corporate bureaucracy that involves LOTS of non-technical people, particularly idiotic managers that insist on staying in the loop of things they barely know anything about (if at all). "I argued long and hard with senior people in IBM that we had to take NT seriously but got no traction" - Ironically enough this further proves that Dave Cutler's worries about IBM were rather well-founded..
I was working for a company that was under contract with IBM Austin porting the OSF Distributed Computing Environment to OS/2. Sometime in the middle of it we got copies of NT on CD. OS/2 was easy to get into a non-bootable state and since it was HPFS it was impossible to modify files outside of OS/2. So, you ended up booting OS/2 from floppies which took forever. I figured out that you could boot NT from CD quickly and get to a running environment where you could fix the OS/2 filesystem such that it would boot OS/2 again. I off-handedly conveyed this information to developers at IBM Austin. They pragmatically then began to use NT to fix their non-booting OS/2 machines. It was a big time-saver. An IBM executive found out and the proverbial stuff hit the fan. (As an aside, NT got it's RPC design from OSF DCE.)
Another great interview - thanks Dave. OS/2 had much better concurrency and I/O drivers than Windows of any flavor. I recall being frustrated by all of Windows locking up when accessing floppies, but OS/2 never had that issue, it would run as smooth with or without floppy accesses. IBM's problem is that it didn't have the marketing nous that Microsoft had, e.g. Microsoft would often give away valuable Windows development products for free at special events, but IBM would always charge an arm and two legs for all their developer toolsets.
"OS/2 never had that issue" - Oh yeah, Windows is riddled with this madness up to this day: the moment you try to access data on ANY flimsier external devices (CD, DVD, USB flash drive, SD card, you name it), things start going downhill pretty fast. "IBM's problem is that it didn't have the marketing nous that Microsoft had, e.g. Microsoft would often give away valuable Windows development products for free at special events, but IBM would always charge an arm and two legs for all their developer toolsets." - More like IBM was WAY too greedy with everyone INCLUDING developers and that has greatly discouraged the adoption of its products among tech-savvy people. And then those tech-savvy people just kept recommending to laymen something else - DOS at first, then Windows later on.
I had watched the video twice, and the Linux was mentioned only in the title and in one comment down below where someone justified have missing the Windows x OS/2 thing because was into Unix and Linux. The word Linux wasn't even said in the video. If I have missed (twice), please, can someone reply with the time when it was mentioned?
I was a DOS guy back in those days despite the hassles of printer drivers, etc. I avoided Windows as long as I could but as Win 3.1 came out a lot of companies providing software for businesses started to embrace Windows. By the time Win 3.2 came out I had a major company that my company dealt with give me "An offer I couldn't refuse!" to start using Windows. The offer was simple; "Convert to Windows or you can't do business with us because we are only going to support Windows." I suspect this scenario played out with a lot of small companies and resulted in a lot of the success of Windows. It was understandable why software developers liked Windows since it greatly simplified compatiblility issues with printers, etc.
@@James_Knott Younger people today simply don't understand such conversations. Regarding your VAX/VMS, I actually built a PDP11 (the kit Heath/Zenith came out with if you remember it). I never got past a papertape reader for it but I did use it in my business. In the early 70's I was a student at the University of TN in Engineering and the Engineering Dept had their own computer system that ran an altered version of Princeton BASIC. We also had COBOL and FORTRAN. BASIC clicked with me and I really liked it. I had an engineering professor give a project and he said we would have to do it in FORTRAN since BASIC couldn't handle it. This was a Wednesday as I recall and the project was due the following Monday. I told him I would do it in BASIC and have it ready Monday morning. He laughed at me. My program worked as well as, or better, than any of the FORTRAN projects and the professor actually apologized for what he had said to me. There is nothing like slow processors and expensive memory to force you to learn how to actually program!!!! Years later I was designing and programming factory automation and those early years of learning how to actually program paid off.
It is both fascinating and tragic how many great ideas are in the both Microsoft's and Dave's past, and how little of those actually work or work well on present day Windows. Can you do a video on MS "Project Cairo" - the Wikipedia page on it gets shorter and shorter every time I look, and I distinctly remember how excited I was about it when I read articles on it (probably in McGraw-Hill's _Electronics_ magazine).
It's been a long time, and maybe my memory is faulty, but when I saw the names "Cutler" and "RSX-11/M" together it brought back memories of seeing "Dave Cutler" in code comments displayed on my VT-52 that was connected to a PDP-11/55 running RSX-11/M. That was in maybe 1977. Also had a GT40, which included a very nice "moon lander" game. Those were very nice machines for their time.
My company at the time placed it's bets on OS/2 PM, because we believed it was the future and would win out over Windows. Our stuff didn't run very well on Windows 3.x. It was much better on Windows 95 and NT. We later did a Java and a web-based version which was an easy solution for Mac and Linux.
Well, at least you joined the ranks of those giants of the MS-DOS world like Lotus and WordPerfect, who also bet on OS/2 and were completely blindsided when the customers went with Windows instead, and with Microsoft’s products for Windows instead of their own.
I love this series - Thank you both so very much for coming together like this; this kind of history and first hand exposure to an absolutely seminal piece of computing ... I was going to say history, but it's not is it? We're still living and breathing the ecosystem that you guys dreamed of and I don't think it's Hyperbole to say that you cannot understate the importance of what you all achieved here. I think the western world to an extent and technology to a great extent would be possibly unrecognisable (and certainly much worse) without the blood, sweat and tears that you all shed! Thank you, you inspired my entire adult life and career!
1000% Agree with Dave's comment on not being able to wait to get off OS/2 working for a major bank.. I was the sole Microsoft guy surrounded by the OS/2, Novell NetWare folks in team meetings they were ugly anytime I suck my head up and said NT would be ideal solution to be howled down as insecure, unreliable, etc, etc.. I recall about 2 years later those same peers who sledged Microsoft come up to me and asking what were the best courses to go on to get their heads around Windows NT as I had managed to get POC established and momentum building on Windows NT. Thanks Dave C for not blessing us with some sort of Frankenstein Xenix 🙂
I had a Netware 3.x CNE and really liked Netware. Then my company decided to roll out Windows and shut down IPX routing. Wanting to stay employed I learned windows. Windows 3.1 saved the day for MS. I went on to get my MCSE for Windows 3.51/4 and then 2000 AD. It gave me a good interesting and challenging career.
5:33 There is a TH-cam video somewhere that gives you a tour through the POSIX subsystem for Windows NT. Let’s just say it was more of a box-ticking exercise rather than something meant for serious use-it is truly horrible.
I ran Windows 1.0 on my IBM PC clone (8.54 Mhz, 640k RAM) in 1986. For all of about 30 minutes. Even then, it was an incredible kludge. Nine years later, armed with a Pentium computer from Gateway, I was enjoying Windows 3.51 well enough to be a (primarily) Windows user ever since, hopping over to Windows NT with 4.0 on up. I find Linux (Tails OS) to be incredibly handy as a portable OS I can run anywhere on most anyone else's computer, as well as any old laptop.
2:35 The trouble with OS/2 was that it was a 16-bit OS coming out in an era when the customers were already moving wholesale to 32-bit hardware. When IBM first released the AT back in 1983 or whenever, they made a promise to their customers that there would be a protected-mode OS that would take advantage of its capabilities. But by the time that promise was fulfilled with OS/2 in 1987/88, nobody cared any more.
wow. i remember in the 80s ibm hyping os/2, but also talking up the future as os/2 NT. so i assumed NT was a development from that after ms and ibm split. oh, this is geek heaven x
You might want to read the book "Show Stopper!: The Breakneck Race to Create Windows NT and the Next Generation at Microsoft" about the development of NT. Really fascinating stuff.
I loved Windows NT, got my MCSE in it. Ran IT for CNC shop where we ran a lot DEC alpha stations on NT, Great fast rock stable computers. We could run circles around SGI O2 stations, even running through DEC's FX!32 emulator programs ran faster than on the Pentiums of that time. Love hearing all this inside history!
My dad wrote a program for OS/2. It was called Tyme Myndr. He had this acrylic trophy from IBM on his desk thanking him for said program. When OS/2 Warp came out, my dad let me install it on my computer. I liked it. Perhaps the novelty of using something different made it feel special.
I love that PDP-11/70 in the background of Dave's office. The 11/70, or 11/45, is one of my "bucket list" computers. I have multiple bucket lists; there's the traditional BID (Before I Die); I checked off a thirty year old bucket list on July 8, 2011.
The title card says it comes out tomorrow, 10/21. If you need something to tide you over, the Computer History Museum did a long form interview with Cutler a few years ago, it's available on TH-cam.
To this day, my favorite OS is Windows NT 3.5. It was the first time I used a computer that refused to fail, but could still run a DOS box, Word Perfect, Lotus 123, Lotus Notes and an X-server for the Unix workstation. All of those programs would fail, but Windows NT would just keep going. The simplicity of the NT File Manager still calls to me and, in my view, Explorer tries to do too much. With it's "links" from drag and drop it can cause more problems than it solves. I recall losing almost all my documents because an update decided to "link" to nowhere. I had to restore from backup and lost about a week's work. Similarly, my biggest problem with having to use my Microsoft Account to setup is because I want to actually name my home folder so I know it's my actual home folder and I can use the command line without embedded spaces, dots, squiggles and numbers. I would appreciate it if Microsoft Windows Install would at least give me the option to name my home folder. Yes, I know how to bypass that requirement, but why should I have to jump through that hoop?
@@toby9999 I don't remember NT 4 as fondly as 3.5. It may be because NT 3.5 was simpler and the first. It may also be because NT 4.0 put graphics drivers back into the kernel to improve speed, but made it less stable. Those were simpler days and maybe that is what I really remember! I was fortunate to be able to use Windows NT at the time which was too expensive for consumer use. Of course, Windows 3.1 was a bit of a legacy hobbled 8-16-32 bit mess at the time.
I had forgotten all about Presentation Manager! I do however still remember reading detailed screenshoted articles in PC Magazine comparing Windows and OS/2 and everyone speculating which would become the dominant one 😜
I'm so stealing the idea for that screen with shelves for my own purpose.. genius, especially if you use it as a separator and want different stuff on either side..
3:25 - Boca Raton was where Armonk sent their "troublemakers" - it was a manufacturing systems innovation center, where they developed a machine that would automatically assemble those nice clickie keyboards they had - that's why the IBM PC came into existence because The Company stayed out of their way. Once the PC took off, Armonk took over, and that's why it eventually failed... Note the first year of IBM PC motherboards were assembled by hand by contractors in their garages in Boca. I had one of those boards once up on a time...
Kind of an oversimplification: "we shipped 3.0 and nobody cared". 3.0 was good...except that it crashed or BSOD at critical times. 3.1 was an attempt to fix the problems and was way more stable. That's why it took off. The users were there...the technology wasn't...until 3.0 finally shipped.
In all fairness to IBM, the _Warp_ iteration of OS/2 was insanely advanced compared to anything MS had at the time. We kept track of how many days, or weeks, or months, that it ran without fatally crashing. Compared to Windows at the time, which was measured in hours! But the lack of support for native applications, plus the fact that MS tried to make sure that every new computer had Windows, killed it.
I had to maintain an automated test system that ran Windows 2.0 No ducumentation, and very old hardware. IT decided to 'upgrade' the test system to Win 95, and wiped the hard drive. I had to go through about 200 old floppy disks to find the OS and test software in bits and pieces. That took over a week. I made five complete sets of restore disks, after tat. I had an unopened copy ow Windows 1.0, that disappeared from my computer shop
@@HunterShows What other option was their, when that test system was assembled? It was built by Scientific-Atlanta to do final test on their Telemetry receivers. We ended up with it after a patent infringement lawsuit. They took our design, repackaged it and added a very early LCD display Everything except their new front panel was an exact copy of the receiver they purchased from us. It was built when the 386 motherboards were quite expensive. It was buit about 30 years ago. The other SATE system I worked with used a Data General 4-bit mini computer, to test PRC-77 military radios. It used a KSR33 Teletype machine for input, and a huuge, 132 column drum printer for output. Each row had a print hammer,. When an error occurred, it printed 132 asterisks at once. That would shake the computer room. This was built in the days of 'Big Iron'. Data General kept a field engineer on site at least 40 hours a week too keep it running. Product couldn't ship if either of these weren't running.
I think he didn’t work OS/2 long enough. OS/2 2.1 and warp ran fantasticly on my old Toshiba 486/SX laptop with 4 mb of ram and a 12 mb PCMCIA card. It was my portable development platform for several years before it finally got superseded by better laptops. Workplace shell did things that modern operating system still can’t do.
I saw on another channel that sales of Windows v3.0 picked up significantly over v2.x. I remember getting that version at work which included disks for Windows/286 and for Windows/386. I think I went from that to v3.1 and then Windows for Workgroups v3.11. I dual booted one of those with (the original) Red Hat Linux. I had a colleague across the hall who loved and used OS/2 up through one of the Warp versions.
Why did you choose Dave Cutler holding his head in most of your thumbnails, like he's fed up? 😂 Anyway, looking forward to the complete interview with this legend.
I would really like to hear the true story of why Microsoft, and who has worked there, hated OS/2 so much. For me it was a mature and stable system, which got me out on the internet!
I find this discussion just hilarious. Seriously, i had an Atari STE at the time which, with its Motorola MC68000 CPU was far superior technically to any PC. It also had some great professional applications and could address more than 640K of RAM ! Later on it also did pre-emptive Multitasking with MultiTOS. Not to patronize anyone, but Windows 3.1 was basically a DOS memory extender + a GUI "heavily inspired" by MacOS (Sculley from Apple later sold Apple's GUI to M$ in return for M$ continuing Office for the Mac) and GEM (which was the Window Manager on the Atari's). The success of M$ was always due to their ability to make deals with the big corporate users and their political Connections. They just knew how the game worked in the US of A. It took them over ten years to ship their First reasonable OS, which was NT 3.51 and later 4.0. but technology doesn't matter if you know how to play the game.
I worked with people at Motorola who told me that IBM really wanted to go with the 68000, but it was a couple dollars more than what Intel said they would sell them the 8080 for. Something really idiotly small like $4, but moto wouldn't budge. I really liked the 68k and PPC chips. Moto was run by retards
Amiga Dos was also a better solution with its taking TRIPOS for the minis and making it work for the Amiga and slapping a desktop on it. The 640k limit came from MS-Dos not going protected mode in Dos 3.x and beyond where other manufactures like Commodore and Apple had a legacy mode (the C128 booting as a C64 or the IIGS booting as a IIe) where nativity they broke compatibility.
The Motorola 68K family was very popular among Unix workstation vendors in the 1980s, until it got supplanted by higher-performance RISC CPUs from about 1989 onwards.
@@lawrencedoliveiro9104 i did some assembly on the M68K. It had a nice instruction set. Much nicer than the awful x86 CISC which still rules. But there is hope with ARM and maybe RISC-V. But smericans only accept what comes from the US. That's why Atari was never really succesful in the US even though much better than DOS and Windows PC's.
Gonna stick my oar in. Let me know if you disagree and why. As an OS from a security perspective and as a server, I loved Novell ( 4 & 5 were buggy though). If you weren't granted access you didn't see it at all. They screwed up in not developing user friendly WYSWIG applications. They tried too late with Zen, and died as a front end. Pity I still prefer their security model over Windows or Unix. Mind you I've been retired for a long time so maybe things have changed. The last I heard their paradigm was still active in the background even if the company died.
The original vision for Netware was pretty good, doing FILE sharing, instead of blck sharing like everyone else was doing beforehand. File sharing just made more sense. And Drew Major & the Superset guys really, really knew their code - things like disk IO were optimised to the nth degree. But as a general purpose OS, Netware was not good. It didn't have good memory protection, it didn't have a whole bunch of subsystems that a "real" OS needs, like GUI APIs etc. But for a while there ... late 80s, early 90s, yep: Netware was definitely king in the file sharing space. Ironically Novell had a winner later on with Directory services, even today I think Novell Directory Services could have been better than AD. But it was too late, Novell was already fading ... remember AppWare?
Awesome... I used Xenix on a TRS80 MODEL 2/12/16/6000 Series (I may have those numbers wrong)... but that thing ran Xenix on it's 68000 and cp/m, TRSDOS, L-DOS/LS-DOS on it's Z80... man I wish I still had that thing
MS became such a Monopoly…screwing companies over. Geo Works became publisher. Lotus /Excel so many more. Netscape incredible browser. Hardware companies got screwed over… Linux was the way to go…but open source was a mystery toThe masses. The biggest sufferer was the consumer. Crap Windows OS. Crap updates. Windows XP SP2 stable as can be..,ruined with future Windows OS. Gates has got to be one of the …….well add your own comment.
I have some books on this. One of my senior friends, used to have to look after a network, using it. Imagine MS, cloning something. Did they do their usual trick.?
Back in the late 90s, I worked at IBM Canada, doing 3rd level OS/2 support, but also supported some apps on Windows 95 and NT. In terms of reliability and stability, OS/2 was the best. NT wasn't too bad, but W95 was a disgrace in comparison. One thing I could be sure of is if there was a problem I had to deal with, it wasn't with OS/2. It just worked! Back then with Windows, you had to be careful about the order of installing things, as conflicts could arise. The issue was with DLLs, IIRC and different versions of them. These days, I run Linux on my home computers and have for over 20 years. I still find some things in Windows annoying, though it's not as bad as it was. BTW, I have a virtual machine of OS/2 Warp 4, along with VMs of Windows 10 and Windows 98 on my ThinkPad. It's amazing to see the changes that have happened over a quarter century.
1: IBM seemed to think they had the best ideas. For example, at one point in OS/2, to delete something from the desktop you would click the item and then drag it to the trash bin. At the time they felt that dragging the mouse around like this was better than a right mouse click plus delete. 2: Prior to Windows 3.1, if you launched app A, launched app B, then closed app A, the system would crash. You had to close the apps in the reverse order that they were launched, which suggested a stack management problem. 3: I think Windows NT was the OS that won people over.
@@jackgerberuaeI believe it was the standard OS in many universities throughout the US in the latter part of the 90s. Also offered was Solaris running on Sun Sparc workstations. It was a great time to be going for a masters in CompSci.
One thing I miss about OS/2 is the smart folder that would reopen the apps and files that were active when it (the folder) was closed. Nobody seems to want to do that on a desktop.
Every clip you've posted with him has been incredibly fascinating to me, can't wait to see the full thing! Great stuff! As a fellow neurodivergent/autistic programmer, thank you so much for your channel and what you do.
I wish the big guys can allow the brought back of the OS/2, not as a mainstream system or to compete or anything, but as a pet project and just for the good times sake. (Or completely open source it if they don’t want to take care of it themselves)
As an end user for a large IBM shop, I got mired in this OS2/Windows schism, and as a software developer it was beyond awful. We were “attempting” to develop/run on OS/2 platform, and it was just atrocious. I’d blame someone, but it really doesn’t do any good to blame. In my mind, it was both Microsoft AND IBM’s fault, and this cluster almost got me fired from a massive project with high visibility!
I remember, when Unix first raised its head, it was being touted as the industry 'Open System', with the intention of it becoming the first system, where users could port their systems onto any Unix based system. Obviously, with the intention of moving away from proprietary systems. It wasn't long, before manufacturers released their own versions of Unix, then followed by Linux etc. And of course, users soon found out, they couldn't port onto them, without extensive redevelopment.
The systems that were officially “Unix” all had some element of proprietary lock-in in them. This fragmented the market, and made them weak against the onslaught of Windows NT. Nowadays, you have Linux and the BSDs-none officially “Unix”, but nevertheless they are the kind of system people think of when they think of “Unix”. There are maybe half a dozen BSD variants, versus about 50 times that number of Linux distros. Yet it is easier to move among Linux distros than it is to move among BSD variants.
@@lawrencedoliveiro9104 the linux distros share the same (upstream) kernel, and some degree of binary compatibility, but the BSDs don't. I think a Microsoft Windows flavor of xenix would have looked a lot like OSX, where there's a stratified layer of legacy with new stuff on top.
@@poofygoof I wonder why the BSDs can’t agree on a common kernel architecture? Now that Microsoft is adding a Linux kernel to Windows with WSL2, I think the logical next step is to let the existing Windows kernel atrophy away, leaving just a GUI layer on top of Linux.
In the 1980s I had the pleasure of working for IBM in Austin on AIX, and it was an amazing experience. It started as a very small team of very motivated people, and I learned so much. Had some contact with Larry Loucks, Charlie Sauer, and Glen Henry, and these guys were on another dimension. I thought I was smart before I met them. I could not believe I was getting paid to do the work I was doing. I sure miss those days, and it's a delight to see some of the work that was done by this talented team of architects and coders lives on today in current versions of systems like VMware and Linux. We always had a friendly rivalry with the OS/2 guys, and wished that we had even 10% of the resources that they got to make AIX better. Thanks for bringing back some memories!
G. Glenn Henry was the software designer of the IBM S/38. The 38 then became the AS400. Great operating system. Highly secure. The system came with an integrated database.
I was at Ford motor a bit later where we ran AIX on Power5 - the shit was way ahead of anything I saw elsewhere at the time. The ability to carve up hardware, I thought, was ingenious. There was a number of IBM folks working with me to either port COBOL or make COBOL apps work on Linux and AIX on Power5. I think the name of the IBM group was called system7 or sector7 or something like that.
I bet you knew my dad (Peter Woon) or Phil Hester if you were in Austin Research Labs?
@@FLY1NGSQU1RR3L76 Definitely knew Phil Hester - he was the top guy the HW side, and I didn't know as many of those folks. The name Peter Woon sounds familiar, but I can't place him. There were a lot of very talented people on that team!
I tripped over a power cord and had to reinstall AIX. Good job :P (joking great OS)
I built the London Stock Exchange system TOPIC2 on OS/2, with 4 IBM RIC co-processor cards which supported an incomming 64kb SDLC line with 32 * 19200 outgoing async lines to the Topic terminals. This was circa 1989-1990. I went to MicroSoft University at Seattle for a week to learn to write OS/2 device drivers etc.
What pissed me off is that we were thrown out of the labs at 5PM. I was there from the UK, what else was I going to do. Got to see the aftermath of Mount St Helens which was pretty amazing.
Writing server processes in C for OS/2 was great It worked well and was pretty fast 25Mhz 386, (for those days).
Back when Microsoft was calling Linux a “cancer”, it made a big noise about the London Stock Exchange choosing Windows Server over Linux. Even put out an ad featuring a fictitious newspaper, the “Highly Reliable Times”, touting the story.
Then that big, embarrassing crash happened. The Windows Server systems were ripped out and replaced with ... Linux.
At a small St. Louis company, I had a similar experience with PS/2s, IBM Unix, and IBM RTIC (not RIC) cards to run inbound call center management software for BM/Rolm telephone switches. Got a commendation from our IBM customers for technical leadership for the design, and (I suspect) using all IBM components in the implementation. We didn't tell IBM that the RTIC card was running MP/M (multitasking version of CP/M) :-)
No you didnt
@@unknownkingdom no you didnt what ?.
Thanks for making this interview!
This guy is a gold mine of good stories!
More we need more!
I worked on OS/2 for about five or six years from 1991 after a year or so working on Windows 2.1 & 3.0.
OS/2 was massive in banking, as were things like MCA (microchannel architecture) and token ring because "nobody ever got fired for buying IBM".
The OS/2 and Presentation Manager API was always far better than Windows from a programmer point of view just from a naming convention PoV.
Then we migrated everything over to NT 3.51 over a couple of years.
Biggest problem in the field in a data center was that you rebooted an OS/2 machine with CTRL-ALT-DEL, but you needed to use CTRL-ALT-DEL to get up the login screen in NT... what could possibly go wrong? So it wasn't unusual for someone to inadvertently reboot a production OS/2 server when they were actually trying to log in in a hybrid data center environment.
Yes, that was a big mistake in NT. If they were starting from scratch, then using CTRL-ALT-DEL to reset the login would be perfectly fine, but they ignored the legacy use.
@@wictimovgovonca320 wasn't it a deliberate choice to stop people exploiting dos/windows ctrlaltdel shanaggingins to exploit login bypassing? its 30 odd years ago, so i maybe remembering incorrectly.
@@julianhaddrill4286 it was deliberate, but most people are idiots.
@@dameanvildoesnt that just describe IT in general 😂
I agree that the OS/2 APIs were much nicer than the Windows API. More consistent, more clearly named, grouped into subsystem functions by prefix. A bit more cumbersome to type the function calls but the code was clearer. You can see a bit of that in Windows today - ISTR that the registry functions in widows was made by an OS/2 team.
I was outfitted with a DEC Rainbow at work about 1984. The Rainbow came with CP/M and MS-DOS 2.11. I recall the Microsoft documentation talking quite a bit about, “MS-DOS is a stepping stone to Xenix.” There was quite a bit in the documentation admonishing to design application programs with user interaction in a way with transition to Xenix in mind. I guess that ship sailed pretty early on…
This interview is pure gold. Thank you Dave. Can't wait to see the final cut. Cheers!
Back in the OS2 days, I was briefly responsible for a small IT group. At the time, our customer support teams had to have as many as 3 terminals to access the separate order, trouble and billing systems. The IT wizard came up with a way to have an OS2 platform to act as the access device for the three systems and allowed the service teams to access a comprehensive customer record with one terminal, one entry of the customer ID. The corporate IT people were incensed at being upstaged by a 3 person group in the field.
how dare you usurp the high priests!
I remember running multiple dos windows with their own instances of Wolfenstein. That's what really got me interested in operating systems.
Ah, yes... the NIH syndrome: Not Invented Here. Many corporate IT shops develop a "vested interest" in their own products without regard to whether it was worth a toot.
@@sdrc92126 One nice thing about OS/2 was that it was great at multitasking, even with DOS & Windows 3.x apps. You could even install real version of DOS on it, of as many versions as you wanted. You could even install CP/M-86, IIRC. Even today, I find Windows 10 multitasking to be poorer than Linux.
@@kegginstructure Sad to say that your statement could just as well apply to the current and most recent NIH (National Institutes of Health) debacle.
I read Solomon's “Inside Windows NT" when it first came out. Cutler's original Windows NT was quite a 'purest' design and implementation with a small (but not micro) kernel that ran with ring 0 privs. Even the graphics sub-system ran in unprivileged space to protect the kernel. Unfortunately that hit performance and they rolled the graphics into the core in NT 4.0
I also liked it as it originally was. In 1993 I started writing device drivers for the NT beta versions, through to 3.1, 3.51 etc. and it was rock solid. I could reload crashed drivers without rebooting. Impressive. Then 4.0 came along and BSODs became part of life.
Yeah good read in the day... NT 3.5 (Daytona) moved some things into the kernel for speed thought graphic driver made the jump then. Dave's original design that is 3.1 was let down by hardware of the day all hardware via the HAL..
It was discovered, I think around the time of NT 3.5, that the difference between the “Workstation” and “Server” installs of NT came down to a simple registry key: change that in your “Workstation” install, and suddenly you had “Server” capabilities, without having to pay extra for them.
Naturally, that no longer worked in later versions.
Microkernels are the roman numerals of operating systems.
00:13 📊 Microsoft shipped various versions, but only 3.1 gained traction, selling 16 million copies in six months.
01:04 💰 Microsoft negotiated a volume license with AT&T for Xenix, selling sublicenses at a profit.
01:44 💻 DOS was chosen over CPM for IBM's first PC, later evolving into a program loader.
02:35 🤝 A joint design with IBM for OS/2 (16-bit) and HPFS file system faced bureaucratic hurdles.
03:33 💡 Bill Gates emphasized the need for a RISC-based processor for the future.
05:31 🌐 Microsoft aimed for a portable system to run multiple environments, including OS/2 and POSIX.
06:40 📚 Initial plans to customize system services for different environments faced challenges in implementation.
08:02 💻 Windows became the main environment for NT due to its widespread acceptance, surpassing OS/2's presentation manager.
Significant points, nice!
DOS didn't evolve into a program loader, it was essentially that from the start and apps had to provide their own I/O code, as the support for it in DOS was essentially non-existent.
Great to see at 80+ Dave is still sharp. I spent many years working on his babies. When he left RSX-11M, I joined it. When he left VMS, I joined it. When he left Windows, I joined it. When he left Azure (Red Dog), I joined it.
Used and loved Warp. Fast and stable was so refreshing. It was such a downer when I had to move on.
but it was still dragging this ancient history, from mainframes, and minicomputers, and such... drive letters, ugh!
I wqs a Warp addict from the start. It was so stable, compared to the early Windows distros. Then We all had to 'move on' as you say. Today I'm on Linux, Open SuSe leap 15.6 !!
I started GUI programming in 1993 and windows 3.11 and MFC and C++ it was a revelation! It was such a nice architecture and I coded and made the deadlines! Our boss wanted us to use NIHCL that was a class library that was free and used by the universities but I didn’t like the idea because integrating it to the Windows ecosystem would have taken ages. We invented something that was later to be called ”rapid prototyping”, we made a mockup of the GUI and showed it and got good fast feedback. I really miss those days.
Those were the days when one guy could write an application that made serious gains in productivity. I developed a web app in the late 90s that provided huge productivity gains to he employees because it integrated a few other systems and reduced triple data entry to single data entry as a byproduct. It generated all of the documents needed for the customers and fed data to the accounting system. It also used email, fax and printing for snail mail as an integrated delivery system. My plan was to integrate FedEx and UPS but I left the company for another job at that point.
Thanks Dave. I started my career on RSX11M and VMS. It's nice to hear from someone that came out of that environment.
OS/2 might have nose dived, but when OS/2 Warp came out, it was way superior than anything Microsoft produced technology-wise - and you could actually surf the Internet without crashing on it, while downloading something else - it was true pre-emptive multitasking (just like Desqview did a few years before). Now, IBM being IBM, marketing wise they didn't have it and that's why they nose dived. Many ATMs were still powered by OS/2 late in 2010...
I ran 32 bit OS2. There wasn't anything else 32 bit at the time. Worked really good.
@@MrRmeadowsOnly for OS/2 2.0, that came out in March 92. NT 3.1 came out in July 93. So that was only for about a year. I did love it though. Still do. Even though it wasn't perfect and NT is arguably a lot better in almost every way (except the user interface, Workplace Shell is still better than anything that MS has ever created).
Same for Win2000. So if you’re comparing a subsequent update use apples and oranges
Dace Cutler is right in the point that OS/2 was an absolutely terrible OS... but that changed with Warp. I'm still sad that IBM screwed that up but Dave is right, that company was a bureaucratic mess.
I used Warp at work for an embedded application. It took a few weeks to get it to work right, then it was bulletproof. Never crashed. We used to joke we got all our crashes done up front. I left a year later but they were still using it the last time I talked to them a dozen years later.
My memory of A/UX and AIX was, A/UX stood for "Almost UNIX", and AIX stood for "Almost imitates UNIX"
The problem with Apple A/UX was that it was rather constrained by the hardware. A/UX was pretty pokey even on what was then a fast 68030 based Macintosh II. Apple did a lot to make UNIX palatable and resilient for an end-user desktop aside from not being able to make it fast enough. I recall that all the AppleTalk stack as well as TCP/IP stack was there, so talking to Apple printers was automagically configured without needing to go in and make obscure changes to the /etc/lpd.conf (or whatever it was). A/UX was a curious mash-up of Sys V and BSD, kind of an opposite mash-up from SunOS. There were also HP/UX, AT&T’s 3B1 Safari PC snd DEC Ultrix that were hybrids which had pieces of both Sys V and BSD. None of those systems had quite the same hybrid mix, so porting applications between them was always a bother.
As an aside, I got the place I worked interested in using a 3B1 and VoicePower board to implement a fairly complex IVR system. We did it all in Bourne shell with the user level VoicePower commands. It was pretty snappy in operation and ran for years with minimal need for administration. It is too bad the AT&T collaboration with Convergent Technologies died on the vine and essentially put an end to Convergent Technologies. The 3B1 was rather nice to work with.
Utterly typical of the chronic fragmentation in the Unix world.
Basically, every system that licensed the “Unix” trademark is now either dead or dying. Instead of “Unix”, people are now running Linux.
@@wtmayhewI worked at Bell Labs and bought a 3B1 through the employee discount program. I had it running at home with UUCP to ihnp4 with a 2400 baud modem. It was incredible at that time (1987/88) to have email connectivity with graphical interface at home. Later I had an AT&T 630 graphics terminal and ISDN for 128kbps data. Now I have fiber and 5G cellular. Amazing how time flies and the world changes.
@@wtmayhewBTW I remember the VoicePower card and the high level shell features. It was one of the domain-specific language concepts that many years later inspired us to create what became the VoiceXML web standard for IVR applications.
@@kenrehor Thanks for the Reply. The VoicePower card seemed way ahead of its time. It was very productive to be able to script in Bourne shell. The UNIX Bourne and Korn shells are very over-looked by the younger crowd and have a lot more power, especially Korn shell, than most programmers realize.
I think a major problem with OS/2 was it needed stronger computers. I like everything is an object, I like extended attributes, even on floppies, and I liked attaching my modem and CD to machine A and using them on machine B.
You might still find references to "john summerfield" "OS/2"
I liked cumulative updates, creating a 5MB partition to boot and run the updates process was a dream incomparising with the Windows process.
And IBM added support for new technology to old versions, USB for example. When Warp 4 surfaced, the MPTS stack was made available to OS/2 3.x users. Free.
The major problem with OS/2 was it marketed by Microsoft.🤣
Windows still has all of these things
Loving these interviews. Very interesting and influential guy. I was working on RSX way back when and I heard rumors about his legend at The Mill at DEC. I continued on VAX/VMS, MicroVAX/VAXELN, OS/2, and NT. Great to hear so many back stories
I could listen to this guy talk all day long. He's got so many good stories from his career.
God this is wild to me. I've been using Windows since 3.0 days, and I remember the different OS/2 iterations. Always wondered at the people behind windows. I'm loving this series.
This is gold, bother Dave. Gold.
⭐️
Really loving these internals interviews, Dave Cutler is fascinating.
Back in the 1980's I worked at Bell Telephone Laboratories in Warrenville IL, we used UNIX for creating all the Circuit Board, NODAL and Schematic Drawings. Can still see the big 32" Green Screen Monochrome Monitors we used back then, was a great time. I also remember Windows 3.1, it was almost a direct copy of the GUI for the Apple Macintosh which I used at Bell Labs as well doing Graphic Arts design. I am a multi talented Computer user working with Almost every major Computer back in the day, Atari, Commodore, IBM, Tandy, Apple. Still own my original CoCo and VIC20.
Love the series. I worked for DEC for 15 years on RSX and VMS. Loved it! Would love to hear Cutler's views on the DEC presidents after Olsen. Seems like they destroyed the company, but I know things are complex and would love to hear Cutler's take on it.
This is a great series of interviews, thanks Dave & Dave.
Hey Dave, I love these videos. Do you have (or could you describe) the display and software for the LED display behind Dave Cutler? It looks like most of the bottom has a form of the LED Fire, with the top being a VU meter on Dave's audio. Just curious about the display HW and SW for that (given your love for ESP32 and addressable LEDs). Thanks, John
Cutler is really is above average. I think anything he would try to do he would accomplish. He is so lucid and on point at anything he says.
I missed the whole Windows and OS/2 thing. I started out in 1982 on a TRS-80 CoCo 1 and then switched to DOS in 1986. My first job out of university was on UNIX (SunOS, specifically) and once I decided I was fed up with DOS on my home computer, Linux had become practical to run on a 486. Been using Linux for 29+ years and never felt the need to try Windows.
You are the first modern day dinosaur I had the pleasure of meeting on the internet!
Nice story you have to tell.
🎯
Hey Dave, this stuff is great and though Cutler couldn't talk about some of the early stuff, there are people probably still close by you who can come talk. Henry Burgess for example can fill you in more on Xenix. I know you may not have direct contact with early folk who are probably recluses like me, but I am sure that you know someone who knows someone who does. If you could interview say David Weise, or Tim Paterson (who wrote Seattle DOS and went to my high school), that would be so interesting I think for a lot of your subscribers. Weise's cool memory trick is a story well known but having him talk about the Ballmer meetings and the situation with IBM I don't think is well known. Weise was a very helpful, bright and I found to be a fellow subversive type guy. I had a compiler whose memory requirements forced me to use OS/2 for builds, but he had a protect mode version of dos that I could make my compiler work on with minor changes and I never had to use OS/2 after it. This was guerrila programming- Cutler was on the "serious OS" side and every serious coder respects that sort of work if your hardware will support it, but MS had to have a market placeholder, and it would be years before NT was ready for consumer machines- and guys like David kept MS in the game.
Also on the Apps side. Maybe Simonyi would be too difficult to talk to, but there are some cool stories- like the reason why Excel could run on the 128K Macs and didn't crash no matter what crazy thing Mac added to their operating system. Any Excel former leads can retell the lore. It's a cool story because MS hardly used any Mac system calls. Almost ignored it entirely.
Which would give your subscribers some glimse into the competitve/ oftentimes mutually condescending relationship between Apps guys vs. Systems guys during the early years of MS. The excel attitude about the OS basically being a massive dependency to bypass and ignore whenever possible was not just towards MacOS but early Windows 2.x and 3.x.
Great back stories. @Dave’s Garage should take note
I was beta on Windows 3.0/3/1 and OS/2 V4 Warp. I liked OS/2 a lot. What was interesting is that when helping people troubleshoot Windows 95, ME and XP I found OS2 commands worked. The divorce between MS and IBM was not nice, like some of the engagement ring was not returned. At the time my family computers were all RedHat and my kids used Win for class projects.
@@furry_homunculus
Didn't see the Z16 IBM video Dave made?
This is may be the most interesting short video about the rich history of operating systems. So much information to unpack. Brilliant. I reposted it to the "Science" page on MeWe.
The word of the day is pontificate. Great wordplay thanks Mr
--Dave Cutler
I was an OS/2 beta-twster, from 1.1 all the way to Warp. I tended to concentrate in networking and network applications.
I seen to remember OS/2 having significant performance and stability issues until 1.3 or maybe 2.0.
I'm mostly an embedded guy these days, with Linux on my desktop. I really miss OS/2's animated backgrounds, though.
What about IBM charging for TCP/IP for OS/2 2.x? I couldn't believe it, so shortsighted for a great OS to be hobbled like that. Windows for Warehouses 3.11 had the 32-bit network stack from Chicago pulled forward, smoked OS/2 2.1. Such a bummer, and IBM treated 3rd party developers like a revenue stream, while Microsoft gave out developer docs and software by the pallet at trade shows. Cutler knew then what a graveyard IBM would become.
A few additions.
1. MS was a BASIC company. It licensed DOS to IBM without having an OS to sell. MS then bought DOS from Seattle Computer Products for ~$100K and used that as DOS. DOS was a rough copy of CP/M, with some people claiming early DOS versions had CP/M bugs in it (very bad).
2. MS got volume licensing for XENIX. But MS's licensing for DOS/Windows was that an OEM had to pay MS for every computer it sold whether or not it had MS's OS on it.
3. He did not want to work with IBM because IBM was an old established company. So was DEC, where he worked before he went to MS.
4. 32-bit OS/2 went nowhere because IBM was bad at marketing and because MS was anti-competitive.
5. He does not say which version of OS/2 he hated so much. OS/2 v2 was decent and more stable than MS's OSes. Versions 3 and 4 of OS/2 were very good, had a full object model, were extremely drag-and-drop oriented, could do multimedia when few companies were doing that, and had very good preemptive multitasking long before consumer Windows and better than NT. The amount of configurability on OS/2 is still among the best I have seen.
My experiences with OS/2 v2.1
1) countless trap errors
2) dirty power off took multiple diskettes to restart
3) installing took 20+ diskettes, downloading the service pack was 20+ diskette images, then had to copy image to diskette, then feed 20+ diskettes to load the update.
@@BobMar1964 I do not recall having those issues. Installation did not take 20+ diskettes if you had the CD version. The CD version of OS/2 2.1 came with two boot floppies. By comparison, "Windows NT 3.1 could be installed either by using the CD-ROM and a provided boot disk, or by utilizing a set of twenty-two 3.5" floppies (twenty-three floppies for Advanced Server)."
So 20+ diskettes for OS/2 v 2.1 is bad, but 22 diskettes for Windows NT 3.1 is good? Got it.
@@georgeh6856 Just relaying my experience/inexperience at the time (1992) with the very smallish office I supported. We did not have any cd drives at the time. We had an IT manager who wanted IBM network.. right down to a OS/2 LanServer on a token ring network. After he left, new management switched to Netware/WfW since it was the market leader with a better support from the vendor. I got out of network administration.
On point 3 "IBM was an old established company. So was DEC". IBM was established in 1911. Digital was estalished in 1957. Bit of a difference. Digital was cool and groovy all through the 60s, 70s and early 80s. There was a world of difference in corporate culture between IBM and Digital; IBM was always rather business-oriented, Digitial was an engineering company, first and foremost. Sadly Digital started to go downhill from 1992 onwards, and eventually it too became a barureaucratic moster, only to be swallowed by Compaq. But that was long after the period Cutler is talking about.
@@Maclabhruinn By 1982, DEC had been around for 25 years. Trying to claim that a 25 year old company is not established is bizarre. By that estimation, Google which was founded in 1998, is still a startup. I would argue that 25 year old Google is much, much more established than Google was, say when it was 5 years old in 2003.
Thank you sirs for this timeline memories! Finally got to my PC and watched a few videos of this interview.
I wish he had asked Cutler if ever thought at the time about just saying to hell with the whole thing and just port VMS to x86 and slap a Windows shell on top of it. A lifetime ago while I was still in school Dave Cutler taught my VMS Crash Dump Analysis class at Digital, that got me started. :)
Brilliant interview. I was basically weaned on 3.1 after vague memories of Amiga 600 and DOS so this is quite something.
I wonder if Cutler will disown the debacles of WIndows 8, Windows Mobile 10, Windows RT and throw Sinofsky under a PCI Bus?
Dave only cared about the kernel - he cared little about the personalities it ran like Win32 or the Shell or Windows UI.
I was so sad when we walked away from the clean 32-bit Presentation Manager API & chose to go back to the 16/32-bit hybrid Windows API for WinNT. It was much easier to program. Eventually, tools like VB & MFC abstracted us away from it.
I'm no developer (just a Helpdesk guy) but it is always fascinating to listen to people who developed everything I've had to work with.
helpdesk is very important work.
Speaking of OS/2.... I was a big fan back before MS essentially killed it. One thing I think was lost was a better file systems. Seems to me HPFS would hold up well compared to the modern Windows file system. Would love to see a deep dive into HPFS vs NTFS. Cheers!
What are you talking about? NTFS started out as an improved and extended HPFS, adding Journaling and other features from the UNIX world. To my knowledge, there's nothing HPFS can do that NTFS can't, and plenty of the reverse. Even OS/2 dropped it in favor of JFS, though that decision was mostly motivated by having to pay Microsoft royalties for the improved HPFS driver.
I worked as a dev on OS/2 for IBM and had MS devs on the team. We thought they were hackers and they thought we were bureaucrats as dave says. I dont agree with that sterotype, there were very good engineers at IBM. IBM saw all the issues in NT and decided that it was another DOS which was a bad mistake. I argued long and hard with senior people in IBM that we had to take NT seriously but got no traction so I joined MS. That was a huge shock.
Huge mistake. NT rocked. I miss it. Simpler times for sure. I threw away $75,000 in Novell licenses for NT 4.0
"I dont agree with that sterotype, there were very good engineers at IBM. " - I don't think he doubted that. What he seemed to admonish was probably more the strangling atmosphere of a corporate bureaucracy that involves LOTS of non-technical people, particularly idiotic managers that insist on staying in the loop of things they barely know anything about (if at all).
"I argued long and hard with senior people in IBM that we had to take NT seriously but got no traction" - Ironically enough this further proves that Dave Cutler's worries about IBM were rather well-founded..
I was working for a company that was under contract with IBM Austin porting the OSF Distributed Computing Environment to OS/2. Sometime in the middle of it we got copies of NT on CD. OS/2 was easy to get into a non-bootable state and since it was HPFS it was impossible to modify files outside of OS/2. So, you ended up booting OS/2 from floppies which took forever. I figured out that you could boot NT from CD quickly and get to a running environment where you could fix the OS/2 filesystem such that it would boot OS/2 again. I off-handedly conveyed this information to developers at IBM Austin. They pragmatically then began to use NT to fix their non-booting OS/2 machines. It was a big time-saver. An IBM executive found out and the proverbial stuff hit the fan. (As an aside, NT got it's RPC design from OSF DCE.)
By the time NT came around, you could boot OS/2 off CD as well.
Another great interview - thanks Dave.
OS/2 had much better concurrency and I/O drivers than Windows of any flavor. I recall being frustrated by all of Windows locking up when accessing floppies, but OS/2 never had that issue, it would run as smooth with or without floppy accesses. IBM's problem is that it didn't have the marketing nous that Microsoft had, e.g. Microsoft would often give away valuable Windows development products for free at special events, but IBM would always charge an arm and two legs for all their developer toolsets.
"OS/2 never had that issue" - Oh yeah, Windows is riddled with this madness up to this day: the moment you try to access data on ANY flimsier external devices (CD, DVD, USB flash drive, SD card, you name it), things start going downhill pretty fast.
"IBM's problem is that it didn't have the marketing nous that Microsoft had, e.g. Microsoft would often give away valuable Windows development products for free at special events, but IBM would always charge an arm and two legs for all their developer toolsets." - More like IBM was WAY too greedy with everyone INCLUDING developers and that has greatly discouraged the adoption of its products among tech-savvy people. And then those tech-savvy people just kept recommending to laymen something else - DOS at first, then Windows later on.
When he mentioned HPFS. it brought back memories of being a tech in the 1990's. It was always known as "half assed file system" amongst my peers.
That would be HAFS, not HPFS.
I had watched the video twice, and the Linux was mentioned only in the title and in one comment down below where someone justified have missing the Windows x OS/2 thing because was into Unix and Linux. The word Linux wasn't even said in the video. If I have missed (twice), please, can someone reply with the time when it was mentioned?
Great interview...you let Cutler talk with minimal interruptions. Then you ask a question and he talks.again.
I was a DOS guy back in those days despite the hassles of printer drivers, etc. I avoided Windows as long as I could but as Win 3.1 came out a lot of companies providing software for businesses started to embrace Windows. By the time Win 3.2 came out I had a major company that my company dealt with give me "An offer I couldn't refuse!" to start using Windows. The offer was simple; "Convert to Windows or you can't do business with us because we are only going to support Windows." I suspect this scenario played out with a lot of small companies and resulted in a lot of the success of Windows. It was understandable why software developers liked Windows since it greatly simplified compatiblility issues with printers, etc.
I came to DOS from VAX/VMS. What a let down! Then, in April 1992, I started with OS/2 2.0.
@@James_Knott Younger people today simply don't understand such conversations. Regarding your VAX/VMS, I actually built a PDP11 (the kit Heath/Zenith came out with if you remember it). I never got past a papertape reader for it but I did use it in my business. In the early 70's I was a student at the University of TN in Engineering and the Engineering Dept had their own computer system that ran an altered version of Princeton BASIC. We also had COBOL and FORTRAN. BASIC clicked with me and I really liked it. I had an engineering professor give a project and he said we would have to do it in FORTRAN since BASIC couldn't handle it. This was a Wednesday as I recall and the project was due the following Monday. I told him I would do it in BASIC and have it ready Monday morning. He laughed at me. My program worked as well as, or better, than any of the FORTRAN projects and the professor actually apologized for what he had said to me. There is nothing like slow processors and expensive memory to force you to learn how to actually program!!!! Years later I was designing and programming factory automation and those early years of learning how to actually program paid off.
It is both fascinating and tragic how many great ideas are in the both Microsoft's and Dave's past, and how little of those actually work or work well on present day Windows.
Can you do a video on MS "Project Cairo" - the Wikipedia page on it gets shorter and shorter every time I look, and I distinctly remember how excited I was about it when I read articles on it (probably in McGraw-Hill's _Electronics_ magazine).
It's been a long time, and maybe my memory is faulty, but when I saw the names "Cutler" and "RSX-11/M" together it brought back memories of seeing "Dave Cutler" in code comments displayed on my VT-52 that was connected to a PDP-11/55 running RSX-11/M. That was in maybe 1977. Also had a GT40, which included a very nice "moon lander" game. Those were very nice machines for their time.
My company at the time placed it's bets on OS/2 PM, because we believed it was the future and would win out over Windows. Our stuff didn't run very well on Windows 3.x. It was much better on Windows 95 and NT.
We later did a Java and a web-based version which was an easy solution for Mac and Linux.
Well, at least you joined the ranks of those giants of the MS-DOS world like Lotus and WordPerfect, who also bet on OS/2 and were completely blindsided when the customers went with Windows instead, and with Microsoft’s products for Windows instead of their own.
@@lawrencedoliveiro9104 We did the same. We had a version of our product running on OS/2 with PM. It never shipped We had to re-tool for Windows.
These stories are utterly fascinating! Thank you for sharing these.
Where was the bit about Linux?
Excelent! Simply Excelent Guest and Host! GREAT CONTENT and History!
I love this series - Thank you both so very much for coming together like this; this kind of history and first hand exposure to an absolutely seminal piece of computing ... I was going to say history, but it's not is it? We're still living and breathing the ecosystem that you guys dreamed of and I don't think it's Hyperbole to say that you cannot understate the importance of what you all achieved here.
I think the western world to an extent and technology to a great extent would be possibly unrecognisable (and certainly much worse) without the blood, sweat and tears that you all shed!
Thank you, you inspired my entire adult life and career!
1000% Agree with Dave's comment on not being able to wait to get off OS/2 working for a major bank.. I was the sole Microsoft guy surrounded by the OS/2, Novell NetWare folks in team meetings they were ugly anytime I suck my head up and said NT would be ideal solution to be howled down as insecure, unreliable, etc, etc.. I recall about 2 years later those same peers who sledged Microsoft come up to me and asking what were the best courses to go on to get their heads around Windows NT as I had managed to get POC established and momentum building on Windows NT. Thanks Dave C for not blessing us with some sort of Frankenstein Xenix 🙂
They weren't wrong to howl. Novell had a brilliant system. Too bad it rode on top of Windows.
I had a Netware 3.x CNE and really liked Netware. Then my company decided to roll out Windows and shut down IPX routing. Wanting to stay employed I learned windows. Windows 3.1 saved the day for MS. I went on to get my MCSE for Windows 3.51/4 and then 2000 AD. It gave me a good interesting and challenging career.
5:33 There is a TH-cam video somewhere that gives you a tour through the POSIX subsystem for Windows NT. Let’s just say it was more of a box-ticking exercise rather than something meant for serious use-it is truly horrible.
I ran Windows 1.0 on my IBM PC clone (8.54 Mhz, 640k RAM) in 1986.
For all of about 30 minutes. Even then, it was an incredible kludge.
Nine years later, armed with a Pentium computer from Gateway, I was enjoying Windows 3.51 well enough to be a (primarily) Windows user ever since, hopping over to Windows NT with 4.0 on up.
I find Linux (Tails OS) to be incredibly handy as a portable OS I can run anywhere on most anyone else's computer, as well as any old laptop.
2:35 The trouble with OS/2 was that it was a 16-bit OS coming out in an era when the customers were already moving wholesale to 32-bit hardware. When IBM first released the AT back in 1983 or whenever, they made a promise to their customers that there would be a protected-mode OS that would take advantage of its capabilities. But by the time that promise was fulfilled with OS/2 in 1987/88, nobody cared any more.
But it was still ahead of DOS and Windows in every aspect you described.
wow. i remember in the 80s ibm hyping os/2, but also talking up the future as os/2 NT. so i assumed NT was a development from that after ms and ibm split. oh, this is geek heaven x
You might want to read the book "Show Stopper!: The Breakneck Race to Create Windows NT and the Next Generation at Microsoft" about the development of NT. Really fascinating stuff.
I loved Windows NT, got my MCSE in it. Ran IT for CNC shop where we ran a lot DEC alpha stations on NT, Great fast rock stable computers. We could run circles around SGI O2 stations, even running through DEC's FX!32 emulator programs ran faster than on the Pentiums of that time. Love hearing all this inside history!
My dad wrote a program for OS/2. It was called Tyme Myndr. He had this acrylic trophy from IBM on his desk thanking him for said program.
When OS/2 Warp came out, my dad let me install it on my computer. I liked it. Perhaps the novelty of using something different made it feel special.
I just tried googling Tyme Myndr and got nothing. What did it do?
I love that PDP-11/70 in the background of Dave's office. The 11/70, or 11/45, is one of my "bucket list" computers. I have multiple bucket lists; there's the traditional BID (Before I Die); I checked off a thirty year old bucket list on July 8, 2011.
You think that's fun? The same designer behind the PiDP-11 is doing a PiDP-10. 36-bit mainframe simulator, in all its blinky glory.
Is there the full interview of this uploaded somewhere? I'd be very interested so see this in it's entirety.
The title card says it comes out tomorrow, 10/21. If you need something to tide you over, the Computer History Museum did a long form interview with Cutler a few years ago, it's available on TH-cam.
To this day, my favorite OS is Windows NT 3.5. It was the first time I used a computer that refused to fail, but could still run a DOS box, Word Perfect, Lotus 123, Lotus Notes and an X-server for the Unix workstation. All of those programs would fail, but Windows NT would just keep going.
The simplicity of the NT File Manager still calls to me and, in my view, Explorer tries to do too much. With it's "links" from drag and drop it can cause more problems than it solves. I recall losing almost all my documents because an update decided to "link" to nowhere. I had to restore from backup and lost about a week's work.
Similarly, my biggest problem with having to use my Microsoft Account to setup is because I want to actually name my home folder so I know it's my actual home folder and I can use the command line without embedded spaces, dots, squiggles and numbers. I would appreciate it if Microsoft Windows Install would at least give me the option to name my home folder.
Yes, I know how to bypass that requirement, but why should I have to jump through that hoop?
What about NT4.0? Did you like it? I skipped 3.5.
@@toby9999 I don't remember NT 4 as fondly as 3.5. It may be because NT 3.5 was simpler and the first. It may also be because NT 4.0 put graphics drivers back into the kernel to improve speed, but made it less stable.
Those were simpler days and maybe that is what I really remember! I was fortunate to be able to use Windows NT at the time which was too expensive for consumer use. Of course, Windows 3.1 was a bit of a legacy hobbled 8-16-32 bit mess at the time.
I had forgotten all about Presentation Manager! I do however still remember reading detailed screenshoted articles in PC Magazine comparing Windows and OS/2 and everyone speculating which would become the dominant one 😜
I'm so stealing the idea for that screen with shelves for my own purpose.. genius, especially if you use it as a separator and want different stuff on either side..
3:25 - Boca Raton was where Armonk sent their "troublemakers" - it was a manufacturing systems innovation center, where they developed a machine that would automatically assemble those nice clickie keyboards they had - that's why the IBM PC came into existence because The Company stayed out of their way. Once the PC took off, Armonk took over, and that's why it eventually failed... Note the first year of IBM PC motherboards were assembled by hand by contractors in their garages in Boca. I had one of those boards once up on a time...
Kind of an oversimplification: "we shipped 3.0 and nobody cared". 3.0 was good...except that it crashed or BSOD at critical times. 3.1 was an attempt to fix the problems and was way more stable. That's why it took off. The users were there...the technology wasn't...until 3.0 finally shipped.
Would love to hear more on his thoughts regarding OS/2
He hates it with every fiber of his being.
If I remember correctly the head OS/2 architect was from IBM - he later started Citrix - Ed Iacobucci ???.
Even at this age I really like his attitude. A true legend.
In all fairness to IBM, the _Warp_ iteration of OS/2 was insanely advanced compared to anything MS had at the time. We kept track of how many days, or weeks, or months, that it ran without fatally crashing. Compared to Windows at the time, which was measured in hours!
But the lack of support for native applications, plus the fact that MS tried to make sure that every new computer had Windows, killed it.
Putting a PDP-11 behind Dave Cutler is pure modern art :D
Xenix was actually pretty good in the day. SCO UNIX (OpenServer) was also a thing for a while.
Incredible guys. Both of you.
This is pure Gold
I had to maintain an automated test system that ran Windows 2.0 No ducumentation, and very old hardware. IT decided to 'upgrade' the test system to Win 95, and wiped the hard drive. I had to go through about 200 old floppy disks to find the OS and test software in bits and pieces. That took over a week. I made five complete sets of restore disks, after tat.
I had an unopened copy ow Windows 1.0, that disappeared from my computer shop
I'm amazed someone used it for something.
@@HunterShows What other option was their, when that test system was assembled? It was built by Scientific-Atlanta to do final test on their Telemetry receivers. We ended up with it after a patent infringement lawsuit. They took our design, repackaged it and added a very early LCD display Everything except their new front panel was an exact copy of the receiver they purchased from us.
It was built when the 386 motherboards were quite expensive. It was buit about 30 years ago.
The other SATE system I worked with used a Data General 4-bit mini computer, to test PRC-77 military radios. It used a KSR33 Teletype machine for input, and a huuge, 132 column drum printer for output. Each row had a print hammer,. When an error occurred, it printed 132 asterisks at once. That would shake the computer room. This was built in the days of 'Big Iron'.
Data General kept a field engineer on site at least 40 hours a week too keep it running. Product couldn't ship if either of these weren't running.
Where was the Linux vs OS/2 & windows? I watched the whole thing and he only mentioned Xenix briefly and how MS just licensed it out.
I think he didn’t work OS/2 long enough. OS/2 2.1 and warp ran fantasticly on my old Toshiba 486/SX laptop with 4 mb of ram and a 12 mb PCMCIA card. It was my portable development platform for several years before it finally got superseded by better laptops. Workplace shell did things that modern operating system still can’t do.
Such as what?
I saw on another channel that sales of Windows v3.0 picked up significantly over v2.x. I remember getting that version at work which included disks for Windows/286 and for Windows/386. I think I went from that to v3.1 and then Windows for Workgroups v3.11. I dual booted one of those with (the original) Red Hat Linux. I had a colleague across the hall who loved and used OS/2 up through one of the Warp versions.
Can the light fixture to Cutler's right be purchased somewhere? It's very cool!
Why did you choose Dave Cutler holding his head in most of your thumbnails, like he's fed up? 😂
Anyway, looking forward to the complete interview with this legend.
Can you blame Cutler. Imagine how many headaches he has had
😵💫😵💫😵💫
@@jackgerberuaeHaha! Good point!
I would really like to hear the true story of why Microsoft, and who has worked there, hated OS/2 so much.
For me it was a mature and stable system, which got me out on the internet!
I find this discussion just hilarious. Seriously, i had an Atari STE at the time which, with its Motorola MC68000 CPU was far superior technically to any PC. It also had some great professional applications and could address more than 640K of RAM ! Later on it also did pre-emptive Multitasking with MultiTOS. Not to patronize anyone, but Windows 3.1 was basically a DOS memory extender + a GUI "heavily inspired" by MacOS (Sculley from Apple later sold Apple's GUI to M$ in return for M$ continuing Office for the Mac) and GEM (which was the Window Manager on the Atari's). The success of M$ was always due to their ability to make deals with the big corporate users and their political Connections. They just knew how the game worked in the US of A. It took them over ten years to ship their First reasonable OS, which was NT 3.51 and later 4.0. but technology doesn't matter if you know how to play the game.
I worked with people at Motorola who told me that IBM really wanted to go with the 68000, but it was a couple dollars more than what Intel said they would sell them the 8080 for. Something really idiotly small like $4, but moto wouldn't budge. I really liked the 68k and PPC chips. Moto was run by retards
Amiga Dos was also a better solution with its taking TRIPOS for the minis and making it work for the Amiga and slapping a desktop on it. The 640k limit came from MS-Dos not going protected mode in Dos 3.x and beyond where other manufactures like Commodore and Apple had a legacy mode (the C128 booting as a C64 or the IIGS booting as a IIe) where nativity they broke compatibility.
The Motorola 68K family was very popular among Unix workstation vendors in the 1980s, until it got supplanted by higher-performance RISC CPUs from about 1989 onwards.
@@lawrencedoliveiro9104 i did some assembly on the M68K. It had a nice instruction set. Much nicer than the awful x86 CISC which still rules. But there is hope with ARM and maybe RISC-V. But smericans only accept what comes from the US. That's why Atari was never really succesful in the US even though much better than DOS and Windows PC's.
@@therealmccoy7221 ARM already ships more units per year than the entire population of the Earth. I think RISC-V has entered that league as well.
OS/2 Warp 4 was the nicest OS I've used until I stopped distro hopping and settled on MX Linux a few years back.
Gonna stick my oar in. Let me know if you disagree and why. As an OS from a security perspective and as a server, I loved Novell ( 4 & 5 were buggy though). If you weren't granted access you didn't see it at all. They screwed up in not developing user friendly WYSWIG applications. They tried too late with Zen, and died as a front end. Pity I still prefer their security model over Windows or Unix.
Mind you I've been retired for a long time so maybe things have changed. The last I heard their paradigm was still active in the background even if the company died.
The original vision for Netware was pretty good, doing FILE sharing, instead of blck sharing like everyone else was doing beforehand. File sharing just made more sense. And Drew Major & the Superset guys really, really knew their code - things like disk IO were optimised to the nth degree. But as a general purpose OS, Netware was not good. It didn't have good memory protection, it didn't have a whole bunch of subsystems that a "real" OS needs, like GUI APIs etc. But for a while there ... late 80s, early 90s, yep: Netware was definitely king in the file sharing space. Ironically Novell had a winner later on with Directory services, even today I think Novell Directory Services could have been better than AD. But it was too late, Novell was already fading ... remember AppWare?
@@Maclabhruinn I totally agree with you. It was Novell's security concept that I admired not its inability to grow to the newer paradigms.
Awesome... I used Xenix on a TRS80 MODEL 2/12/16/6000 Series (I may have those numbers wrong)... but that thing ran Xenix on it's 68000 and cp/m, TRSDOS, L-DOS/LS-DOS on it's Z80... man I wish I still had that thing
MS became such a Monopoly…screwing companies over. Geo Works became publisher. Lotus /Excel so many more. Netscape incredible browser. Hardware companies got screwed over…
Linux was the way to go…but open source was a mystery toThe masses.
The biggest sufferer was the consumer. Crap Windows OS. Crap updates. Windows XP SP2 stable as can be..,ruined with future Windows OS.
Gates has got to be one of the …….well add your own comment.
Certified OS/2 Engineer here ^^ I switched the company I was at from OS/2 to NT4.0 though ;)
I have some books on this. One of my senior friends, used to have to look after a network, using it. Imagine MS, cloning something. Did they do their usual trick.?
Back in the late 90s, I worked at IBM Canada, doing 3rd level OS/2 support, but also supported some apps on Windows 95 and NT. In terms of reliability and stability, OS/2 was the best. NT wasn't too bad, but W95 was a disgrace in comparison. One thing I could be sure of is if there was a problem I had to deal with, it wasn't with OS/2. It just worked! Back then with Windows, you had to be careful about the order of installing things, as conflicts could arise. The issue was with DLLs, IIRC and different versions of them. These days, I run Linux on my home computers and have for over 20 years. I still find some things in Windows annoying, though it's not as bad as it was.
BTW, I have a virtual machine of OS/2 Warp 4, along with VMs of Windows 10 and Windows 98 on my ThinkPad. It's amazing to see the changes that have happened over a quarter century.
1: IBM seemed to think they had the best ideas. For example, at one point in OS/2, to delete something from the desktop you would click the item and then drag it to the trash bin. At the time they felt that dragging the mouse around like this was better than a right mouse click plus delete.
2: Prior to Windows 3.1, if you launched app A, launched app B, then closed app A, the system would crash. You had to close the apps in the reverse order that they were launched, which suggested a stack management problem.
3: I think Windows NT was the OS that won people over.
But no common PC user got to use NT, so how did this happen?
@@jackgerberuaeI believe it was the standard OS in many universities throughout the US in the latter part of the 90s. Also offered was Solaris running on Sun Sparc workstations. It was a great time to be going for a masters in CompSci.
pure GOLD. Thank you.
One thing I miss about OS/2 is the smart folder that would reopen the apps and files that were active when it (the folder) was closed. Nobody seems to want to do that on a desktop.
Great stuff! 'Showstoppers' is one of the most interesting books I've read!
Is Gordon letting still around? He might be interestiNG to hear about
what is that cool strato-globe on the bottom left of dave?
Every clip you've posted with him has been incredibly fascinating to me, can't wait to see the full thing! Great stuff! As a fellow neurodivergent/autistic programmer, thank you so much for your channel and what you do.
Glad you enjoyed it!
I wish the big guys can allow the brought back of the OS/2, not as a mainstream system or to compete or anything, but as a pet project and just for the good times sake. (Or completely open source it if they don’t want to take care of it themselves)
Actually, I believe it is, though I don't recall it's current name.
As an end user for a large IBM shop, I got mired in this OS2/Windows schism, and as a software developer it was beyond awful. We were “attempting” to develop/run on OS/2 platform, and it was just atrocious. I’d blame someone, but it really doesn’t do any good to blame. In my mind, it was both Microsoft AND IBM’s fault, and this cluster almost got me fired from a massive project with high visibility!
I remember, when Unix first raised its head, it was being touted as the industry 'Open System', with the intention of it becoming the first system, where users could port their systems onto any Unix based system. Obviously, with the intention of moving away from proprietary systems. It wasn't long, before manufacturers released their own versions of Unix, then followed by Linux etc. And of course, users soon found out, they couldn't port onto them, without extensive redevelopment.
The systems that were officially “Unix” all had some element of proprietary lock-in in them. This fragmented the market, and made them weak against the onslaught of Windows NT.
Nowadays, you have Linux and the BSDs-none officially “Unix”, but nevertheless they are the kind of system people think of when they think of “Unix”. There are maybe half a dozen BSD variants, versus about 50 times that number of Linux distros. Yet it is easier to move among Linux distros than it is to move among BSD variants.
@@lawrencedoliveiro9104 the linux distros share the same (upstream) kernel, and some degree of binary compatibility, but the BSDs don't.
I think a Microsoft Windows flavor of xenix would have looked a lot like OSX, where there's a stratified layer of legacy with new stuff on top.
@@poofygoof I wonder why the BSDs can’t agree on a common kernel architecture?
Now that Microsoft is adding a Linux kernel to Windows with WSL2, I think the logical next step is to let the existing Windows kernel atrophy away, leaving just a GUI layer on top of Linux.
@@lawrencedoliveiro9104 Not how that works. WSL2 runs Linux under Hyper-V. The rest of the system is running on the NT Kernel.