There is a long term relationship between Citrix and Microsoft. It started in 1989 when Citrix was founded. Without Microsoft’s backing in funds and licensing, it probably would not have been possible for Citrix to start. I have heard that this initial relationship was largely due to the reputation of Ed Iacobucci.
At the time, Microsoft was still a believer in OS/2 and I’m sure that Ed’s idea (multi-user OS/2) still made a lot of sense. I’m sure Microsoft would have been interested enough to eventually release this version of OS/2 if the relationship with IBM had not gone so badly. Of course, things did not go to plan and IBM and Microsoft did have a divorce which actually caused a problem for Citrix as well.
Because OS/2 was no longer a Microsoft focus, Citrix became less of a focus as well. During the time between 1991 and 1994, Citrix was really off the Microsoft radar screen. Funnyly enough, this gave Citrix some breathing space because we didn’t need to worry about Microsoft would think since we were based on OS/2. On the other side of the coin, it made it harder to convince customers that it was okay to base their businesses on an OS/2-based product. Over time, we learned to stop pushing the OS/2 message and instead focus on what our WinView product could do.
At one stage (1993) we were actually doing some life cycle maintenance of OS/2 1.3 for Microsoft. I believe we had some kind of contract to do work on the real OS/2 for Microsoft in order to receive our OS/2 license. I never saw the agreement but I did hear that Microsoft wanted two people assigned to working on OS/2 issues. I remember fixing a couple of problems for them.
The relationship started to change in 1993 when Citrix decided to make Windows NT multiuser. Obviously Microsoft was to become very interested in the outcome of this.
One thing that might not be common knowledge is that Citrix had full source access to OS/2 and NT during most of its early history. In fact, some people (Life Cycle Maintenance) still do have access to Windows code even today. In the early days, full access was assumed and it meant that Citrix could change anything it wanted to in order to make the OS multiuser.
The model was to ship a different version of OS/2 or NT code that would look and feel like the original OS but it would support multiple users at the same time.
This was a bit of a problem related to service packs and hotfixes. If Microsoft released a new service pack, it was a big deal to merge the two code bases together to produce a new update. It wouldn’t be until Windows 2000 that the multiuser extensions would be included in the mainstream product .
Anyways, the point is that Citrix had easy access to changing the operating systems and did so.
The tension began to appear in late 1995. When Citrix first released WinFrame 1.5, Microsoft almost instantly recognized the potential value. I think they saw their customers thinking about moving over to using WinFrame for certain tasks.
By 1997, things had changed considerably. Citrix had already done the work to convert NT 4.0 to be multiuser. The previous version was based on NT 3.51. Citrix did have rights to ship products based on NT 3.51 but not NT 4.0. Citrix made a mistake in not guaranteeing access to NT 4.0 before starting work on WinFrame 2.0. Microsoft denied access to the NT 4.0 market. Within a short time the relationship became very rocky. Microsoft was threatening to do the work itself and basically told Citrix that it was in its best interest to sell the multiuser code to them.
After a few months on a roller coaster ride in negotiations, Citrix agreed to sell the multiuser code to Microsoft. After this, Citrix would only have ICA to base its products on. It would no longer be able to change the core multiuser code.
Microsoft released WTS (Windows Terminal Server) in 1998 based on the Citrix code for NT 4.0. In one way, this was good for business since it sanctioned the market from Microsoft. In other ways, it was sad that Citrix no longer had this core technology.
I think I would summarize this as this: It is very difficult to trust Microsoft for any long term relationships. Eventually, Microsoft will show its true intentions and this is always a reflection of self interest.
The trend since 1997 has been further erosion of the Citrix market towards Microsoft. I will write more about this in a future post.
This is a great reminder of the early days of Citrix. I was there as an early adopter customer and then a VAR. Whatever challenges exist with Microsoft it is clear that Citrix has navigated them extremely well. Citrix is one of the very few ISV’s who have been able to play with Microsoft and not only survice, but prosper!
I jumped straight from DOS+Windows (actually DR-DOS, as I had dumped MS-DOS after 5.0), to IBM’s OS/2 2.0, the first IBM-only release, back in April 1992. Clearly, it was IBM who made OS/2 viable and technologically advanced, giving it the object oriented desktop, the WPS (Workplace Shell), DOS/Windows virtualization, a 32-bit GPI, long filenames, and more…
It’s kinda ironic that you said: “Over time, we learned to stop pushing the OS/2 message and instead focus on what our WinView product could do”, and then conclude “It is very difficult to trust Microsoft for any long term relationships. Eventually, Microsoft will show its true intentions and this is always a reflection of self interest.”
Gee, where did I hear that before… oh yeah… here:
“No company has ever done a deal with Microsoft that lasted. They’re just naive if they think they can” – Sun’s Jonathan Schwartz, 2002.”
Also: “One Microsoft internal memo suggested that the best way to “stick it” to rival Borland was to “pre-announce” a Microsoft program that was not ready. The judge said the document was “as close to a smoking gun as you can get.” (New York Times 1/21/95 P.17)”
After the split with Microsoft (which only produced a crappy 16-bit OS/1 1.x that wasn’t even nice to look at), IBM really did want to conquer the desktop OS marketplace with its own technically advanced, 32-bit OS/2 from 1992 until late 1995. That’s an about four-years edge.
Looking back, if Citrix had not gone towards the Evil Empire and instead released a multiuser 32-bit OS/2 instead of pursuing its own “Winframe” thing, perhaps the fate of IBM’s desktop OS would have been different and Microsoft wouldn’t have had a chance to “assimilate” Citrix… remember IBM only threw the towel on the desktop in mid to late 1996.
By the time OS/2 Warp version 4.0 was released (once again, an innovative product, with built-in voice dictation and browser navigation, something that MSFT is copying 10 years later with Vista), Microsoft had a beach head with NT Workstation 4.0, and the availability of products like Citrix only made the Microsoft assault on desktops faster.
So… do I feel bad about Citrix? Not a bit. They choose to go with the Evil Empire, and now looking back, it seems that you regret that. I’d frankly say you got what you deserved. If you do business with wolves… don’t complain if you get hurt.
Oh that brings back memories, remember WinView and Winframe and doing my first course and getting a lovely printed certificate in Winframe 🙂
Ed Iacobucci that’s a name from the past retired in 2000 and started DayJet.
Can everyone remember Virtigo and the Viper give away ?
Nice summary of the history with microsoft. I was there in the old days 1995/1996 and my first experience was with tektronix Windd based on nt 3.51. They actually improved winframe with features like client drive access etc. etc. Windd completely dissapeared. I guess mostly because the additions to winframe they made were integrated in winframe 1.6 or 1.7 (i don’t exatly remember).
Kind off like microsoft did with winview technology i guess 😉
What’s the history with tektronix anyway? did they have source code access?
Interesting trip down memory lane! I seem to remember somewhere I have a copy of the 2.0 Release of MetaFrame. I seem to remember this was code named PICAsso or something like that but the name was later changed due to copyright issues.
When I used to run the WinFrame training course so many people complained that they wanted a Windows 4.0 interface. And then Windows 2000…
At the time of WinFrame we had a lot of customers who used WinView for a very long time. At the time it seemed to be alot more robust that the Windows equivalent.
Thanks for the post – enjoyed reading it!
These comments are very interesting. I honestly didn’t expect to hear from such a diehard OS/2 fan.
The comments about OS/2 are even more interesting because I worked at IBM on OS/2. From 1989 to 1993 I worked for IBM Boca Raton. I think the last release I worked on with OS/2 2.1.
So, believe it or not, I fully understand what happened with OS/2.
Deep down, I think can confess to the world that IBM didn’t know how to do something like OS/2, especially on its own.
It had this big thing about taking on Microsoft after the divorce but then proceeded to completely mess it up.
I think it is almost funny to blame Citrix for the lack of success with OS/2. That was just weird actually. Citrix could have helped make OS/2 more popular but IBM really wasn’t interested. IBM was deeply hurt that developers from OS/2 “defected” to Citrix and therefore excluded Citrix from the OS/2 “party” for a number of years. Citrix actually had to license everything through Microsoft for OS/2.
As for Microsoft, they are not innocent either. The divorce meant that even though OS/2 was supposed to be a joint IBM/Microsoft project, IBM was to get sole custody of OS/2 and Microsoft would focus on Windows.
At this point, Windows 3.0/3.1 were doing really well and I think it would have been stupid for Microsoft to ignore where the money was. Regardless of the issues with IBM (which were many), I’m sure that Microsoft saw an opportunity to focus on the Windows family on their own. IBM had no stake in Windows and obviously it would be a lot more interesting to do the Windows work solo.
The only lose end was what to do with the high end market. By the point, OS/2 3.0 was being worked on by the Cutler team. IBM had already invested in it but had very few IBM people involved in it. Microsoft renamed this project and claimed it for what is now called Windows NT.
It should have been obvious to IBM that producing system software for the general market was not going to work.
Why would one hardware manufacturer trust another? Why would Compaq want to ship OS/2 when it can get unbiased support from Microsoft. There was a complete lack of trust between the hardware vendors even though IBM tried so hard to convince them otherwise.
And, probably more significantly, IBM tried to use a model based on “big iron” software to produce OS/2. The bureaucratic layers above the programmers where terrible. Even worse, we not trusted as much as the Microsoft programmers even though we often knew more than them.
It really was a terrible situation to work in. I certainly did not want OS/2 to fail. By the end of 1992 I had enough and when the phone call came in from Citrix, I thought I should give it a chance.
Even though Citrix was going through some of its roughest times, it was a dream job compared to what was going on at IBM.
I remember that IBM had this philosophy that anyone could be a programmer. They would actually take people from manufacturing lines and send them to a training course for six weeks to become programmers. I found this to be unfair to everyone involved. Believe it or not, some of these people actually did okay. Most struggled and pretended. The point is that you cannot expect quality teams if you don’t invest in having quality teams.
I’m glad I left when I did.
By 1996, the IBM Boca site was closed and the remaining people scattered. Most ended up in Austin Texas working on other projects.
Don’t forget, the Citrix team of the early 90’s was essentially the same team that designed and delivered the original OS/2 vision and the early(1.x) products — in the days that MS and IBM could actually work together.
Ironically, it was IBM executive leadership’s inability to embrace an independent partner (MS) in a “strategic IBM product” that was the ultimate cause of serious morale issues within *both* companies. Good technical people cannot do their best work when their own management does not tell them when to be “competitors” and when to be “partners”. In a project as complicated as CP/DOS, issues would arise in nearly every design discussion.
As Jeff clearly points out, the morale problems led to an irreconcileable tension at the technical working levels that fueled mass defections of talented programmers to Citrix.
What goes around …
[…] Alas, other deadlines loom. But if Citrix is going to make any big partner announcements, next week seems like perfect timing. Some channel publications have taken Citrix to task over its channel program. But The VAR Guy can’t afford to point fingers or throw any stones at the competition. In the 1990s, he predicted that Microsoft’s own multi-user software would trample Citrix in the market place. Boy was he wrong. […]
Can you remember the websites of old… .