Caluser2000 wrote on 2021-04-22, 18:27:
Caluser2000 wrote on 2021-04-22, 16:52:
All MS Dos 4 had was the gui file manager thingy loading up first. As did MS Dos 5. This was optional. So I'm confused with the MS Dos 4/Topview comparison. MS Dos 4 failed because it had a few major problems which are well documented. One reason I went to DR Dos. Topveiw was just a gui tool on top of Dos.
Systems with MS Dos on were still being sold new well into the mid '90s.
IBM produced a retail version of PC Dos until the late '90s
I am quite aware of the multi tasking MS Dos 4. It was never adopted by the major industry playersd so it was really in the game anyway and rearaly discussed on forum s like these. There were other multi tasking Dos variant such as Concurent Dos, MSIs Real/32, Paragons PTS Dos 32, etc etc.
I used DrDos matched with GeoWorks Pro 1.3 back in thge early '90s.
Look - this was out of a discussion of whether the adoption of DOS with multitasking back during the early/mid AT era ('86) would've changed the trajectory of DOS and influenced the timeframe for the adoption of NT or NT-like operating systems - as much as it was not a "major player" it was still released by Microsoft, and it was similar to another piece of software sold by IBM - both of them (along with Digital Research or DR) were the parents of DOS on x86 commodity PC hardware as we know it, so they weren't the players...they were the dealers in the game. While it was a stepping stone and an evolutionary dead-end, it could've been a potential turning point for DOS if the OEM response back in '86 were more positive.
Out of the 3 examples you cited only Concurrent DOS (Concurrent DOS/286 - around '86) really existed around the same timeframe as DOS4 and Topview, and could be considered its tech contemporary - except it had its own CP/M derived compatibility baggage. DR did release a version later called Concurrent DOS/XM later with XMS support, and it could've bank switched its way to 8MB of RAM. XMS support was something that DOS4/Topview didn't do, which was likely why they both went nowhere after the first OS/2 . If both DOS4/Topview seen good uptake it'll would've probably evolved into something like Concurrent DOS/XM, but simpler.
Real/32 grew out of Digital Research's Concurrent DOS/386. Concurrent DOS/386 came out around 1988 and was a multiuser/multitasking DOS in the same family lineage as Concurrent DOS/XM, but it's not a single user multitasker like Win9x - it's more like a hypervisor suite. That allowed someone to buy a multi-socket Compaq SystemPro 386 server, throw Concurrent DOS/386 onto it, run serial breakout boards (or modems on dialup pools) to serve real mode DOS (in multiple v86 instances) to multiple users simultaneously off dumb terminals back in the days. It required DOS applications that did not do something crazy like use TSRs, perform special hardware tricks or require 386 protected mode facilities (DPMI wasn't a thing until 1989) - its uptake were not high mostly because it was expensive, and also partly because commodity PC hardware of the late 80s aren't quite
“enterprise grade” yet. If you had to spend the money to bring up a bunch of dumb terminals in '89, you'll probably connect it to something like an IBM mainframe, a Tandem Nonstop VLX (what NYSE ran at that time) or something like a Sun-3.
Real/32 did support DPMI, but it didn't hit the market until '94 when Novell gave up on DR DOS , and a year after the release of Novell DR DOS 7 (which grew out of the same lineage but were single user/multi-task). Those 2 were preemptive multitasking DOS variants but both came out too late to change DOS's trajectory - which was to eventually evolve into IBM OS/2 Warp (didn't happen) or Microsoft Win9x or NT.
Paragon's PTS-DOS 2000 is not the same as PhysTechSoft's PTS-DOS 32 which came out much, much later on the DOS lifecycle (past '95), and I did not remember if multitasking (preemptive or cooperative) was even a thing with them. PTS-DOS 32 is Russian only, so detailed english language tech writeups are scarce.
GigAHerZ wrote on 2021-04-23, 08:21:
Agreed. And the evolution might have been even uglier than the shoehorning of Win9x. (Which, let's be honest, isn't too beautiful under the bonnet either.)
What i don't agree is that you imply dos would have stayed like we know from our real-world timeline. I believe, in case of MS-DOS 4 success, all the 386 enhanced features and memory management and stuff like that would have arrived to MS-DOS natively. (Remember, prior to windows 3.0 (or special version of 2), it also didn't support any 386 enhanced features. But starting from 3.0, some were implemented.) And for better or worse (probably worse), we would have had longer time on consumer level running on top of MS-DOS and some gui achieving something like Win9x did.
It's a path of history that never happened, but it would have been somewhat interesting. (Like some other historic possibilities - imagine Xenix becoming popular and today we would all run on unix-like op. systems?) Yet, i do agree with you, that world (MS-DOS 4.0 success) probably would have been worse than better...
What I implied since the beginning is that the unlikely success of MS-DOS 4 back in '86 will speed up adoption to IBM's OS/2, rather than leading to something like DOS5 and eventually Win9x. 386 protected mode will still arrive later but it'll likely be a component of OS/2, and its timeline will be dictated by IBM, and not Microsoft. Maybe DOS would've died along with the 286, and it'll be REXX scripts instead of batch files.
