Demonstration of what is probably OS/2's key selling point to the PC market: the capacity to seamlessly multitask native and Windows APIs.
From what I can assess IBM remains a success today performing the role that made them successful for a century: centralized data infrastructure. Technology has marched on, the micros of today exceeding the mainframes of yesterday and the mainframes of today are primarily built with information security rather than storage space (which nowadays is ample and easily upgradeable) in mind. Demand for qualified personnel to maintain that information has grown but so has the usability of interfaces for non-technical information workers (of which I am one) to access and update it. Indeed the past 4 decades have seen a computer revolution!
As part of the Millennial generation I had the honor (whether positive or negative is contingent on one's perspective) of witnessing the consolidation of the computer system standards that power our everyday lives today. Petty rivalries aside (Windows dominates market share while Apple's obscene profit margins keep them in the top spot), in general usability for everyday people is as good as it's ever been and it is only helpful when learning curves are eliminated by home use of the same systems as on the job. It isn't easy to conceive today when we have it so much simpler but there was a time of no dominant standards and the competition to produce a winner yielded some interesting results.
I'll try not to focus too much on the history of OS/2 because others much more immersed have done so better than I could. Multiple sources can be found easily with a quick internet search but for a technical overview I will shout out Michael Necasek's OS/2 Museum (https://www.os2museum.com/wp/) which as a bonus gives the same treatment to the first several versions of DOS, thereby tying together the whole IBM/Microsoft saga. But while he proceeds from a professional perspective I wish to explore from the consumer perspective I actually had growing up.
Beginnings
I started using DOS at quite a young age...and wouldn't have gotten far if not for a menu program Dad had installed because I wouldn't have made heads or tails of command line operation. That these file and application management programs were purely text-based was a non-issue if they adequately did the job thus the market for them continued until the end of DOS itself. A mouse was a specialized tool only needed for the select programs that worked with it; otherwise keyboard and joystick were just fine for games and most other routine uses.
A graphical user interface would first go mainstream with Apple's Macintosh in 1984. It was far from the first GUI however the design innovations of what would later be known as MacOS had such impact that for the rest of the decade Apple would try to assert ownership of the GUI standard itself, cowing pioneer Digital Research and pressuring Microsoft (despite the latter being one of Macintosh's key developers of productivity applications). To be while at the same time not be MacOS would shape the IBM PC realm's transition from text menus to iconic desktop interface.
In the mid-80s both Macintosh and IBM PC compatibles were seen as business and enthusiast machines and priced accordingly; the still-small segment of home computer users had their own selection of 8-bit Apple II, Atari and Commodore machines that were less capable but still adequately covered their basic productivity (and gaming) needs. IBM (as ill-equipped to do so) tried breaking into that market with PCjr which was a notable failure in that it prompted gaming-minded innovation on the professional-oriented PC platform, yielding Tandy's 1000 line which is a good history in itself. Overall however the paltry color palette and pathetic sound that defined the PC made it clear what the design priorities were.
That was not to say the PC was supposed to be a specialist machine any more than the Macintosh; even then technical specialists made up a shrinking fraction of computer operators, giving way to information workers who used microcomputers as aids to accomplish more in less time and would've been helpless if expected to do the same via the coded interface of a mainframe or minicomputer terminal. For these workers usability was of utmost importance; that improvements in that area would also benefit computer users at home wouldn't hurt either. Particularly as the new Atari ST and Amiga systems followed Apple's approach by stocking windowed GUIs with the systems both IBM and Microsoft knew the heyday of the command line as defined by CP/M was past and the PC needed its own flagship equivalent to the Xerox Alto interface that had impacted both Steve Jobs and Bill Gates.
Predecessors
Jobs manifested the idea in his Apple Lisa, the failed predecessor to Macintosh that used a very similar windowing and menu system, essentially a pre-release Mac OS. Gates took the same idea and ordered development of what would become Windows 1.0. Despite the close working relationship between the two companies IBM refused to get on board with Windows, and soon enough it was revealed why.
Remember that in 1984 IBM and Microsoft were development partners for DOS, allies but still separate companies with their own interests (as mentioned already Microsoft developed for Apple too). In what would become a prophetic pattern they each took their own approaches to addressing the limitations of DOS which was already being outgrown by advances in PC hardware, that year seeing launch of the 286-based PC/AT and introduction of protected mode which had great multitasking potential...if it could be exploited well by the applications themselves. The attempts by IBM and Microsoft to accomplish this is detailed further below, but suffice to say the need for backwards compatibility with existing applications proved a tripwire for them and the protected mode memory paging only got used really well by Digital Research in their DOS-supporting multiuser operating systems (which, of course, are strictly professional and not consumer systems).
IBM and Microsoft made efforts to address both the multitasking and user interface shortcomings in the same program, their respective labors bearing fruit in 1985. Microsoft Windows was a true full-fledged GUI (and colorful, something no Macintosh could claim at the time) however its delayed production and strict real mode support all but obsoleted its first version at launch since multitasking more than the few bundled applets could quickly hit a memory wall. Text-based IBM TopView did support protected mode and introduced tailoring DOS application attributes through the Program Information File (PIF), however its multitasking capability (and success) was likewise limited though for different reasons than Windows.
Joining Forces
As 1987 approached and it was clear both approaches made strides but not successfully enough the two companies would both converge and diverge. 386-based PCs were about to launch (this time not from IBM first) and that left IBM and Microsoft in a real conundrum between the need for legacy support and pressing forward before ending up left behind, with insufficient time and resources to cover both those bases well. So begins one of those fascinating tales of how necessity breeds innovation.
Anticipating this problem back in 1985 had led to the IBM/Microsoft Joint Developer Agreement (JDA) which officially shifted the norms for the first time since IBM launched the PC and Microsoft took charge of its companion operating system DOS. Microsoft, who continued improving Windows for the low end PC market, was tasked with producing the replacement to DOS, Windows AND TopView. IBM would in the meantime assert refining DOS and TopView for the new PC-based hardware it had in production.
Microsoft's first result was a multitasking edition of command line DOS, however limited release and lack of protected mode support leave this as a computing history footnote. The new Microsoft OS/2 wasn't ready for the launch of IBM's PS/2 line so IBM had DOS 3.3 and TopView 1.12 for it instead. IBM would also be responsible for DOS 4 though that would be perhaps the most poorly-received version due to its high end orientation causing the bulk of users to stick with the less memory-hungry DOS 3.
OS/2 1.0
The one edition lacking a GUI altogether has it considered by some as a publicly-released beta (the later PowerPC edition may also qualify but that won't be covered in this post). Not unlike Windows NT in the 90s OS/2 started with so much potential but demanded too much of most to reach that potential. This can be seen as useful in the professional realm for protected mode multitasking of Microsoft's line of text-mode development tools, a task OS/2 was particularly suited for provided they were Family API and therefore not subject to the restriction of OS/2 1.x supporting only a single DOS session.
OS/2 1.1
The first to offer Presentation Manager and support for a companion networking package, both still only useful in the professional realm in 1988. Its companion is Windows 2 with which it shares the windowing interface (naturally, both were made by Microsoft) and can even run in the OS/2 DOS session which was more useful than it sounds because by version 2 Windows had much better application (plus protected mode) support. Still, only a fraction of professional operators would need these extended capabilities of OS/2 since a dearth of native application support left that as its primary asset.
OS/2 1.2
The Presentation Manager was upgraded to what would later be integrated in Windows 3, but the latter wasn't launched yet so this was the pinnacle Microsoft's side of OS/2 development. It was supposed to continue from here but the low success of OS/2 prompted Microsoft to keep on their own way with DOS 5 and Windows 3. That being said, the application of domain networking in business was understood so for better or worse Microsoft would, in one of its more interesting twists, be compelled to support OS/2 for years after quitting development because it was their only operating system capable of covering that ground until Windows NT was ready to do the same.
OS/2 1.3
Windows 3.0 had launched and was selling fast, ending the JDA so at this point IBM took over refining OS/2 like they had with DOS starting with 3.3. In retrospect the biggest drawback of all the 1.x versions is that despite their elite system requirements they only supported 286 protected mode via a Microsoft reset tweak that enabled it to work more (but not perfectly) seamlessly with real mode applications, save for the server editions supporting 32-bit access through HPFS386. It's a glaring difference that the more accessible Windows 2 & 3, while less stable due to running on top of DOS, didn't have this weakness.
OS/2 2.0
Microsoft gave IBM more than just a bunch of viable desktop operating systems: they were also in tune with the growing non-specialized end user market and their departure left IBM with what would turn out to be an unrecoverable appeal deficit. That point made, IBM was not lacking in technical proficiency or knowledge of what makes systems usable and it shows by how much their independent versions of OS/2 exceeded their Microsoft-produced predecessors. While how much this affected third-party application support is a subject for debate IBM was onto something in integrating a fully-operational Windows 3.0 kernel since they knew that gave OS/2 automatic mass market appeal.
OS/2 2.1
That appeal only grew as Windows 3.1 proved the most successful Windows yet so OS/2 2.1 integrated that kernel upgrade. It must be understood that before Windows NT launched (and found the sufficient success that eluded it until 4.0) DOS-based Windows was positioned as both the client operating system for business and the computing interface for home, a trend which continued until Windows 98 Second Edition. Multimedia Presentation Manager enabled OS/2 to support Multimedia Windows programs which makes this the point that IBM had its opening to put OS/2 on the mass market map.
OS/2 3.0
And they really tried, just ineffectively. Everyone knew Windows 95 coming and would likely reshape the computing landscape so 1994 was the time to get a jumpstart before it was too late. On usability OS/2 Warp was the best yet, it's DOS and Windows support even better than Microsoft's while OS/2 native application support had finally peaked. Windows NT offered equivalent stability but not accessibility and its virtual DOS machine was notoriously poor. OS/2's DOS gaming support left something to be desired but dual-booting with a contemporary DOS (which many OS/2-equipped computers came with anyway) solved that issue in a manner remarkably reminiscent of how Windows 9x would.
OS/2 4.0
IBM had its chance in 1994 but it was over come 1996. Microsoft had already begun their transition into the ruthless monster we know them as today but that doesn't change the fact that Windows 95's success was fair because it delivered what the growing computing masses needed at the time they needed it. It was long past the time computer operation was the domain of the well-trained geek who could do his own system administration and compile his own programs. Though Warp 4 was available on the general market and a small number of enthusiasts snatched it up, as a power user operating system it no longer had enough unique capacity to offer in a world getting dominated by Windows 95 on the personal side and Windows NT on the business side (though OS/2 continued for a long time in the embedded system market, one Windows NT was slow to exploit). Closing Thoughts
I was always a DOS and Windows user, the quirks of both something I remember growing frustrated with back in the day (yet working with that is part of the appeal of nostalgia today). I never heard of OS/2 [or Windows NT for that matter] back when it could've made a difference, though if I'm being honest I can't come up with a reason it would've appealed other than as a more stable platform for Windows games (I didn't start using my PC for productivity purposes until after Windows 95). With the luxury of hindsight I can appreciate this more, but even though I keep a 20th century retro PC I since learned OS/2 has quirks of its own that would make my own adoption of it today too much of a hassle. Nonetheless I respect the impact it made hence why I pay tribute to it through today's post.