Was OS/2 a viable alternative for daily DOS and Windows tasks?

Will they finally open the Workplace Shell? It's a truly object-oriented desktop design that's still superior - a decade later - to anything Windows has to offer. Looking back it's hard to believe a lot of the early FUD from MS against OS/2 was aimed at scaring people away because, hey, 2 megabytes of memory was just an absurd requiremet! They also claimed multithreaded programming was no big deal. If they open up the Workplace Shell maybe OS/2 could preserve some of its legacy. It would rock on Linux.

From: https://tech.slashdot.org/comments.pl?sid=155912&cid=1307025...

What does this even mean?

Inheritance. In windows (or Linux) a file is an extension and is associated with an application or collection of applications. In OS/2 a file can inherit from various parents. So for example you could have a file xyz.mp3 of a lecture:

1) Since its an .mp3 the sound programs work on it and "edit" opens it up in a sound editor

2) Since its part of a lecture it will inherit from the word transcript and when you "open for transcription" it opens with another (say word doc) transcript

3) Since the lecture an be associated with other lectures which have video with them when you "open for viewing" it can pull up the associated powerpoint

etc...

Now the important thing is that these behaviors are inherited because the xyz.mp3 is a member of class lecture in addition to being a member of the class .mp3. Members of the class .mp3 are sound files while members of the class lecture have associated transcripts and lectures.

It also made some attempt at polymorphic behavior (i.e. edit, open, etc..). Finally information was encapsulated at the lowest level (the file, the folder, etc..)

We wrote a large body of building automation software subsystems in OS/2. There was no easy way to provide the same functionality in Windows, so it was never cost effective to port it.

To this day, we keep the central routing server and all the subsystems in OS/2 boxes that are treated like embedded control systems, and have written Windows 2K-based interface code that proxies everything as BACnet devices.

OS/2 was a good combination of modern OS services (named pipes, threads, etc.) and easy development. Given how simple it was to access serial ports, we could easily interface via DigiBoard multiplexers and such, and could write a new system driver (including reverse engineering time) in less than six months.

I'm the primary contact for IBM in our office, so they've been flooding me with information about porting these apps to Linux, which sadly, may never be cost effective.

I am very sorry to see this event, even though I fully understand and appreciate all the factors that led to OS/2's demise. It's like watching a very dependable ship being sent to the bottom of the ocean because it's too expensive to keep it afloat.

Oh well...

Tim

Good luck with that son, but I'm sorry to tell you that Microsft did NOT help IBM code OS/2 so it would run Windows. As a matter of fact, Microsoft did far more to STOP OS/2 from running Windows and Windows applications. When Microsoft was releasing betas of Chicago( Win95 ), IBM had Chicago apps running under OS/2. When Microsoft found out, they changed the OS so that a very small portion of the Win32 resources loaded up at the 1GB memory address. This was so OS/2 could not run ANY Chicago applications or the OS. It worked because OS/2 supported virtual memory up to 512MB.

So you got that WAY WRONG. The bit about Microsoft licensing issues preventing opensourcing OS/2 is correct.

LoB

For DOS and Windows, it could run legacy apps but not the newer Win32 (i.e. Windows 95) apps and if you needed those versions it wouldn't be viable. That said, there were just enough apps to make it work if you cared, similar to the current situation with Linux. I remember the Describe word processor which was the first non-desktop-publishing editor that handled styles properly. Development tools and hardware support continually improved. What was lacking was interoperability with folks using Windows 95 (e.g. Office).

After the whole IBM/MS OS/2 v2/v3 (aka NT) ordeal, we all switched to Windows NT and everything was fine (except for IBM).

If you want to understand the mindset at the time, it was all about hardware drivers. The only thing that mattered was hardware drivers, and the hardware manufacturers created drivers for dos/windos and nothing else at all.

Drivers was one of the primary challenges holding back any competition to MS. True also of OS/2.

OS/2 was seen as a solid architecture, but it arrived without a GUI, right at the time that the GUI became the key feature needed in any OS. It didn't take long for the Presentation Manager GUI to arrive, but it was nothing exciting.

The other thing seen as an issue at the time was that OS/2 was built for the rather terrible 80286 when really what was needed was full 32 bit 80386 support. It eventually got the 32 bit support but this was a bad start.

Microsoft released Word and Excel for OS/2 then discontinued them. That was the finish of OS/2.

Could you buy a white-box PC with it already installed, for the same price as one with Windos on it? Were there drivers for all the hardware likely to be needed? No and no.

Sure it would run the programs, way more reliably than Windos could.

The playing field was sharply tilted, so there was never a possibility of fair competition.