This is a great history so far, and the comment threads really enrich the content.
I was a teenager when all of this was going down, so some of it is a blur. But I remember Windows 3.0 being mentioned here and there by a few people that liked it, and then all of a sudden Windows 3.1 was out and it was ubiquitous. I had been using DOS for 3-4 years by that point and don't remember anyone talking about Windows before Windows 3.0.
A few other things I remember from that time:
1. As a Windows 3.1 user with a teenager's budget, I remember aspiring to run OS/2 2.0. It just seemed so much more advanced. I even bought 16MB of RAM to run it, which at the time was extremely expensive because of the chronic "memory factory fire" in far-off places that always seemed to explain why memory was so expensive. However, I think the technical sophistication sustained my interest more than what it could actually do. I remember so much of OS/2 feeling unpolished and underwhelming from an end user perspective, even if I felt assured that it was technically capable of doing more than Windows. I always got the sense that disk caching never worked properly on any of my systems, and what's the point of being able to use long filenames if those files then become invisible to your DOS and Windows applications?
2. I was at a consumer trade show in Canada where IBM was trying to sell OS/2 2.x to consumers. They did this by showing Space Quest 4 playing simultaneously with Flight Simulator while also searching for files in Windows File Manager - hey look, the Windows clock is still moving even though Windows only has co-operative multitasking! It was cool if you understood why it was cool, but in retrospect I find it funny that you'd use such a ridiculous use case to sell a product to home users / consumers who just want to get work done.
3. Windows 95 came along and OS/2 just disappeared from the map for me. It solved all of the issues I had with OS/2 and gave all of the benefits and more.
On the first Billion $ in revenue, I was not an employee at the time, but an intern in MSFT France, but I got to see Frank Gaudette's company meeting skit, as a Blues Brothers, singing "Coming to you on a dusty road, big numbers I got truck loads". This was MOST excellent.
"Windows 3.0 seemed to have everything that OS/2 did not. There were compatible PCs. There were new applications. There were supporting peripherals. It had the pricing and distribution too. The fact that it ran on as little as one megabyte of memory (though really two was better) and also ran all existing MS-DOS applications even better than they ran before made for an incredibly compelling launch."
You forgot "Windows 3.0 could print."
It's intriguing that y'all in Apps didn't see the Windows killing OS/2 writing on the wall. I was still in college that Spring (I didn't start at MS until July, 2 1990) and I had decided Windows was going to win the previous Fall. It was sooo obvious from the outside (at least to me).
We is me :) I think we were all devs and hung up on the fact that we were moving to a stable development environment finally with OS/2. Plus our group was the cross platform group :)
I arrived as an intern on the Windows dev team in March 1990, disappointed to have been assigned to the 'temporary' Windows platform, and not the future: OS/2. But that changed as soon as I started playing with Windows 3.0.
I did add optimizations designed for the OS/2 HP LaserJet drivers to the Windows drivers, before working on TrueType support for Windows 3.1 drivers. Banded printing was slow and such a pain, and it felt like every byte of PCL reduction was a major victory.
In apps we were pretty clear on Windows being the likely winner. There were still contracts and until the divorce was final, including the systems team's completion of handing off work on OS/2 2.0 Warp, we covered our bases (we never shipped apps for OS/2 2.0).
As for printing, a far too large a part of my life was spent on getting OS/2 to print. This was one of the few areas where it was as much Microsoft's fault as IBMs. The printing architecture was very complex. This complexity led to a lot of crashes and other errors if the configuration was not precisely correct. When I complained, Microsoft engineers would tell me that systems administrators would help users make it work. :-(
On the IBM side, related to their weird insistence on lines of code as a developer metric, they were equally rigid on their unit and system test results. In printing, none of these tests looked at the actual output, just the API status codes.
I reported a bug on the main text output function ExtTextOut() (I forget the real OS/2 name, I think maybe GpiExtTextOut()) for a major printer. The tests said all was fine. The printer in question produced no output. Stepping into the printer driver in the debugger showed that the function set the success status code and returned, doing no work at all. The bug was resolved as by design and OS/2 1.1 shipped this way.
It took a while to get printing to actually work.
I spent a lot of time with the IBM printer people in Atlanta who eventually became LexMark. They were conscientious engineers who worked through all the issues and by the time OS/2 1.3 was released printing did actually work.
"To me the company still seemed so approachable. The Microsoft I knew was not much larger than my high school and I felt like I knew all the people in Apps." -- this was a great time at Microsoft, I remember the feeling of the company being so approachable as well. You could walk 5 minutes and see anyone in the company, you could get time with any exec, and yet the company was working on these incredibly impactful products -- MS/DOS, Windows, OS/2, Word, Excel, Powerpoint, etc.
Couldn't agree more. The energy, enthusiasm and can-do-anything attitude and confidence was palpable across the whole company for most of 90s. Most exciting time of my career/life !!
Yes Windows 3 was a great improvement over Windows 1 and 2. Having used a Mac, the Windows 1 & 2 UI with non overlapping windows felt primitive. Developing apps under Windows 1 and Windows 2 was interesting.
One of the great things about the NT project was how Paul Maritz basically kept it low key. I think it wasn’t until the summer of 1991 (so a year after this post) that Windows “New Technology” would start to make its way to the public. The team (Dave Cutler) joined Microsoft after DEC cancelled VMS 5.0 in late 1988. They spent 2+ years designing and building NT before even becoming known inside of Microsoft. It was another OS “skunkworks” in many ways—even BillG was kept at arms length :-) More on this coming.
My recollection was that razzle was just a name chosen for the first release of nt (3.1) but no one used that as dave hated code names. it was just NT. the build environment for the other 3.x 16bit releases had no connection to that (winall, snowball, pinball, etc). there was almost no coordination between those teams at that point.
I first booted NT sometime late 1990. I was in developer support at the time. It was on a Compaq 386e and it took over an hour to boot. Unlike, Steve, I only have vague memories of my first years at MS, but I remember the experience of getting NT booting like it was yesterday.
Yes. The internal builds started late summer early fall. You might remember that we did early MFC 1.0 on the PDC and later than these first builds the GDI subsystem of NT used the C++ prerelease. Stay tuned for more as we start to work on C++ in chapter 2.
Windows 3.0 came with Windows help WinHelp 1.0. It had almost everything that would later become the world wide web (except for being local, no servers). WinHelp 1.0 had text, graphics, even graphics with hyper link areas and text hyperlinks. Source code was RTF instead of HTML. Windows Help was complied into a .hlp file. For technical writers it was complicated to write help systems without tools due to the syntax.
Steven, I've been wondering about the symbiosis between Windows and PC gaming. I was never a big gamer myself but I know that gamers were always ahead of the curve, for hardware and I'm assuming for operating systems as well... did Microsoft make a special effort to ensure that Windows made game development easier than OS/2 or Macintosh did? I'm pretty sure even today gaming is all [console, mobile and] Windows - almost nothing on Mac or Linux
Gaming in Windows is a bit in the future. Microsoft had a long history of games. Gordon Letwin himself (the original MS-DOS architect and OS/2 architect) created a clone of the command line game Adventure that sold with the original PC (and of course BillG and someone you’ll meet next chapter, NeilK, created games with DOS 1.0).
At this time most of the games work was in the consumer division and multimedia where there was a huge effort around CDROM games. Most of this won’t show up for another year or two.
After that, after Windows 95 we start to see DirectX which is the really unique gaming work that led to Xbox.
This is a great history so far, and the comment threads really enrich the content.
I was a teenager when all of this was going down, so some of it is a blur. But I remember Windows 3.0 being mentioned here and there by a few people that liked it, and then all of a sudden Windows 3.1 was out and it was ubiquitous. I had been using DOS for 3-4 years by that point and don't remember anyone talking about Windows before Windows 3.0.
A few other things I remember from that time:
1. As a Windows 3.1 user with a teenager's budget, I remember aspiring to run OS/2 2.0. It just seemed so much more advanced. I even bought 16MB of RAM to run it, which at the time was extremely expensive because of the chronic "memory factory fire" in far-off places that always seemed to explain why memory was so expensive. However, I think the technical sophistication sustained my interest more than what it could actually do. I remember so much of OS/2 feeling unpolished and underwhelming from an end user perspective, even if I felt assured that it was technically capable of doing more than Windows. I always got the sense that disk caching never worked properly on any of my systems, and what's the point of being able to use long filenames if those files then become invisible to your DOS and Windows applications?
2. I was at a consumer trade show in Canada where IBM was trying to sell OS/2 2.x to consumers. They did this by showing Space Quest 4 playing simultaneously with Flight Simulator while also searching for files in Windows File Manager - hey look, the Windows clock is still moving even though Windows only has co-operative multitasking! It was cool if you understood why it was cool, but in retrospect I find it funny that you'd use such a ridiculous use case to sell a product to home users / consumers who just want to get work done.
3. Windows 95 came along and OS/2 just disappeared from the map for me. It solved all of the issues I had with OS/2 and gave all of the benefits and more.
On the first Billion $ in revenue, I was not an employee at the time, but an intern in MSFT France, but I got to see Frank Gaudette's company meeting skit, as a Blues Brothers, singing "Coming to you on a dusty road, big numbers I got truck loads". This was MOST excellent.
I recall a cartoonish photo of Frank Gaudette shovelling dollar bills into a cart at the company meeting
"Windows 3.0 seemed to have everything that OS/2 did not. There were compatible PCs. There were new applications. There were supporting peripherals. It had the pricing and distribution too. The fact that it ran on as little as one megabyte of memory (though really two was better) and also ran all existing MS-DOS applications even better than they ran before made for an incredibly compelling launch."
You forgot "Windows 3.0 could print."
It's intriguing that y'all in Apps didn't see the Windows killing OS/2 writing on the wall. I was still in college that Spring (I didn't start at MS until July, 2 1990) and I had decided Windows was going to win the previous Fall. It was sooo obvious from the outside (at least to me).
Good stuff, Steve!
We is me :) I think we were all devs and hung up on the fact that we were moving to a stable development environment finally with OS/2. Plus our group was the cross platform group :)
I arrived as an intern on the Windows dev team in March 1990, disappointed to have been assigned to the 'temporary' Windows platform, and not the future: OS/2. But that changed as soon as I started playing with Windows 3.0.
I did add optimizations designed for the OS/2 HP LaserJet drivers to the Windows drivers, before working on TrueType support for Windows 3.1 drivers. Banded printing was slow and such a pain, and it felt like every byte of PCL reduction was a major victory.
In apps we were pretty clear on Windows being the likely winner. There were still contracts and until the divorce was final, including the systems team's completion of handing off work on OS/2 2.0 Warp, we covered our bases (we never shipped apps for OS/2 2.0).
As for printing, a far too large a part of my life was spent on getting OS/2 to print. This was one of the few areas where it was as much Microsoft's fault as IBMs. The printing architecture was very complex. This complexity led to a lot of crashes and other errors if the configuration was not precisely correct. When I complained, Microsoft engineers would tell me that systems administrators would help users make it work. :-(
On the IBM side, related to their weird insistence on lines of code as a developer metric, they were equally rigid on their unit and system test results. In printing, none of these tests looked at the actual output, just the API status codes.
I reported a bug on the main text output function ExtTextOut() (I forget the real OS/2 name, I think maybe GpiExtTextOut()) for a major printer. The tests said all was fine. The printer in question produced no output. Stepping into the printer driver in the debugger showed that the function set the success status code and returned, doing no work at all. The bug was resolved as by design and OS/2 1.1 shipped this way.
It took a while to get printing to actually work.
I spent a lot of time with the IBM printer people in Atlanta who eventually became LexMark. They were conscientious engineers who worked through all the issues and by the time OS/2 1.3 was released printing did actually work.
"To me the company still seemed so approachable. The Microsoft I knew was not much larger than my high school and I felt like I knew all the people in Apps." -- this was a great time at Microsoft, I remember the feeling of the company being so approachable as well. You could walk 5 minutes and see anyone in the company, you could get time with any exec, and yet the company was working on these incredibly impactful products -- MS/DOS, Windows, OS/2, Word, Excel, Powerpoint, etc.
Couldn't agree more. The energy, enthusiasm and can-do-anything attitude and confidence was palpable across the whole company for most of 90s. Most exciting time of my career/life !!
Yes Windows 3 was a great improvement over Windows 1 and 2. Having used a Mac, the Windows 1 & 2 UI with non overlapping windows felt primitive. Developing apps under Windows 1 and Windows 2 was interesting.
At this time was NT anywhere on the radar as far as Apps or the press is concerned?
One of the great things about the NT project was how Paul Maritz basically kept it low key. I think it wasn’t until the summer of 1991 (so a year after this post) that Windows “New Technology” would start to make its way to the public. The team (Dave Cutler) joined Microsoft after DEC cancelled VMS 5.0 in late 1988. They spent 2+ years designing and building NT before even becoming known inside of Microsoft. It was another OS “skunkworks” in many ways—even BillG was kept at arms length :-) More on this coming.
Brownie points - name NT's code name prior to NT?
Razzle
Is this why the build environement for XP also refers to itself as razzle then?
I'm assuming it just considers everything to be a version of NT, irrespective of what it is intended for or what branding it will use
My recollection was that razzle was just a name chosen for the first release of nt (3.1) but no one used that as dave hated code names. it was just NT. the build environment for the other 3.x 16bit releases had no connection to that (winall, snowball, pinball, etc). there was almost no coordination between those teams at that point.
Interesting, thanks for taking the time to reply, I just joined the list but I am LOVING the way you tell the story
I first booted NT sometime late 1990. I was in developer support at the time. It was on a Compaq 386e and it took over an hour to boot. Unlike, Steve, I only have vague memories of my first years at MS, but I remember the experience of getting NT booting like it was yesterday.
Yes. The internal builds started late summer early fall. You might remember that we did early MFC 1.0 on the PDC and later than these first builds the GDI subsystem of NT used the C++ prerelease. Stay tuned for more as we start to work on C++ in chapter 2.
Windows 3.0 came with Windows help WinHelp 1.0. It had almost everything that would later become the world wide web (except for being local, no servers). WinHelp 1.0 had text, graphics, even graphics with hyper link areas and text hyperlinks. Source code was RTF instead of HTML. Windows Help was complied into a .hlp file. For technical writers it was complicated to write help systems without tools due to the syntax.
Steven, I've been wondering about the symbiosis between Windows and PC gaming. I was never a big gamer myself but I know that gamers were always ahead of the curve, for hardware and I'm assuming for operating systems as well... did Microsoft make a special effort to ensure that Windows made game development easier than OS/2 or Macintosh did? I'm pretty sure even today gaming is all [console, mobile and] Windows - almost nothing on Mac or Linux
Hello Rohit,
Gaming in Windows is a bit in the future. Microsoft had a long history of games. Gordon Letwin himself (the original MS-DOS architect and OS/2 architect) created a clone of the command line game Adventure that sold with the original PC (and of course BillG and someone you’ll meet next chapter, NeilK, created games with DOS 1.0).
At this time most of the games work was in the consumer division and multimedia where there was a huge effort around CDROM games. Most of this won’t show up for another year or two.
After that, after Windows 95 we start to see DirectX which is the really unique gaming work that led to Xbox.