SoylentNews is people
SoylentNews
A silly new app has been doing the rounds this week: Windows 95 as a standalone application. Running on Windows, macOS, and Linux, the Windows 95 "app" combines Electron (a framework for building desktop applications using JavaScript and other Web technology) with an existing x86 emulator written in JavaScript. The emulator can run a bunch of operating systems: for the app, it's preloaded with Windows 95.
This is, of course, software piracy. The developer of the app has no rights to distribute Windows 95 like this, and I'm a little surprised that the app hasn't been yanked from GitHub yet. And for now, the app is just a toy; there's no real reason to run Windows 95 like this, other than the novelty factor of it actually working.
But Windows 95 (and software that runs on or requires Windows 95) was an important piece of computing history. I think a case could be made that it's Microsoft's most important Windows release of all time, and its influence continues to be felt today. Not only was it technically important as an essential stepping stone from the world of 16-bit DOS and Windows 3.x to 32-bit Windows NT, and not only did it introduce a user interface that's largely stayed with us for more than 20 years—Windows 95 was also a major consumer event as people lined up to buy the thing as soon as it was available. A full understanding of the computing landscape today can't really be had without running, using, and understanding Windows 95.
(Score: 3, Informative) by requerdanos on Sunday August 26 2018, @05:39PM (12 children)
Perhaps it was important, but probably not as important as Windows NT, which was an actual operating system as opposed to a nice, feature-filled GUI for DOS.
No, it wasn't. Windows NT 3.1 was released July 1993 [wikia.com]. Windows 95 didn't come out until August 1995 [wired.com]. The stone had been stepped years previously.
More people used Windows 95 than Windows NT 3.1, and perhaps more people remember it. But after 2000 or so, Windows NT was the only Microsoft offering for PCs, servers, and laptops (under the names "NT 4" and "2000" and "XP" and "Vista" and "7" and "8" and "10"), while Windows 95 had a few brief updates (98, ME) and then faded from view entirely. Because it was a nice, feature-filled GUI for DOS, while NT was an operating system.
(Score: 2, Interesting) by Anonymous Coward on Sunday August 26 2018, @06:10PM
Win95 pretty much just used DOS to load itself up into memory. Then scooped the brains out and put itself in place on all of the interrupts and put itself into protected mode. I think you overlook the importance of it and what writing a protected mode program on x86 was like back in 94/95 (hint you pretty much had to ignore all of DOS and write your own 'os'). Going from DOS to NT was a challenge. But going from win95->NT programmatically was a good move for MS. The API was similar enough that many times it 'just worked'. The EXE format is the same one they defined in 95 and ports easily. Win3.x was a DOS shell and they reused as much of DOS as it could get away with. The API was dissimilar enough you spent usually many months hunting down compatibility bugs.
(Score: 4, Informative) by SomeGuy on Sunday August 26 2018, @06:18PM (1 child)
Simply having NT available was not quite a "stepping stone".
Go back in time to when NT 3.x was released and try to run it. None of your commercially available applications would run natively on it. Those that did run would run like pigs in the NTVDM. Virtually none of your common hardware would work. If you called for tech support you would get laughed at. If you were to try to deploy this across a business, things would grind to a halt as critical applications and drivers would not work.
It took years for applications to migrate to 32-bit NT and vendors to write drivers. And even longer yet for people to buy all new compatible hardware and software. Small special purpose items took much longer and even more money. Hell, there are still some that never got migrated that are still in use today.
The key selling point of Windows 95 was that it was almost 100% compatible with exiting DOS and Windows 3.1 software while providing an integrated and more complete Win32 API. You could load that proprietary MS-DOS CD-ROM driver in your config.sys and it would STILL WORK! Meanwhile you could run that bleeding edge 32-bit application and be one step more ready for the NT world.
The improved Mac-like user interface was a huge bonus, as was peer to peer networking.
(Score: 1, Interesting) by Anonymous Coward on Sunday August 26 2018, @09:35PM
It was interesting how hyped NT was-and had been in the pop tech journals for a few years prior to its release about how compatible, safe and faster it was than anything else, namely UNIX/AIX and OS/2. Even after the release it was universally pushed as more compatible than the original systems it was emulating.
(Score: 1, Insightful) by Anonymous Coward on Sunday August 26 2018, @06:33PM (1 child)
Nobody said it was first, or technically superior compared to Windows NT. They said more important.
Windows NT was a good OS for its day, in the role of file and print server for Windows based networks, or as a Microsoft ecosystem database server. It also had limited compatibility with applications designed for windows 3.1, essentially no ability to run DOS software and no value for running games, beyond minesweeper and solitaire anyway. It was no more useful as a desktop OS to the majority of people than Linux is today.
Windows 7 eventually showed that the world envisioned by NT was good, but it took almost two decades to get there. If not for Windows 95 and its successors, everyone would have been stuck with DOS and probably would have eventually ended up abandoning Microsoft entirely.
Microsoft could have gotten along without NT. It couldn't have without 95.
(Score: 3, Interesting) by requerdanos on Sunday August 26 2018, @06:50PM
If you're talking about Windows NT 3.1, I slightly disagree in that I was not very impressed with it. It required a minimum of 16MB RAM during the days when having 8MB ("filling up all 8 slots") was remarkable. It didn't support long file names on FAT. It was slow and clunky. The 16-bit virtual machine in which it ran existing windows software was especially slow on the hardware of the time. NT 3.1 was advanced in that in was a true 32-bit operating system with pre-emptive multitasking and pretty much the features of a modern operating system, but in very rough form.
If you're talking about Windows NT in general, its "day" is 1993-present, as it's the current windows [microsoft.com] being sold by Microsoft (as "Windows 10").
Microsoft has provided and updated Windows NT from 1993-present, 25 years and counting. Microsoft provided and updated Windows 95 from 1995-2000, 5 years etched into a headstone.
Your position seems odd in part because with no Windows NT, Microsoft would have had no operating system product for the last twenty years. That would be fine with me; I don't use Windows anyway. But it doesn't reflect reality.
(Score: 2) by Nuke on Sunday August 26 2018, @06:56PM (2 children)
Technically Win 3.x was a GUI on DOS (not the only one, or the first - which was VisiOn*), but it has to be admitted that Win9x was an operating system separate from DOS, although re-using parts of it. It was crap anyway and should never have been extended to Win98/ME by which time MS could have produced a lite version of NT that would have run on any entry level PC by then. It was the games on Win9x (and not on NT) that kept it going, on life support.
NT was the first even half-decent OS that MS wrote themsleves (they bought DOS); even so to write it they poached a team of coders from DEC who (it is believed) brought some VMS (DEC's OS) code with them. There was later some financial settlement between MS and DEC. Despite NT being around from 1993 for professional use (at a professional price), WIn 9x was the stepping stone to NT (now eveolved to XP) for most people.
* VisiOn, a GUI for DOS, was shown at at a trade show in 1992, Gates saw it and demanded that MS write their own GUI, which was launched as Windows 1.0 three years later.
(Score: 0) by Anonymous Coward on Monday August 27 2018, @02:45AM
Of course they used VMS in naking WNT. Add 1 to each letter in VMS and you get WNT.
(Score: 2) by Nuke on Monday August 27 2018, @11:52AM
Sorry, I made a typo in my comment above. VisiOn was shown in 1982, not 1992.
(Score: 4, Insightful) by Arik on Sunday August 26 2018, @07:11PM (2 children)
And this was the best thing about Windows 95, in that comparison.
There was mention of the best version of Windows - it's not 95, but it's close. 98SE was without any doubt the best version of Windows ever made. And it was essentially Windows 95 SP3 - the final, most polished release of the series.
It was better than XP precisely because it wasn't an OS. It could still be unloaded, there was still a real OS with a user-accessible interface waiting underneath. There was still a way to get it out of the way and get work done.
You can't do that with NT. As sophisticated as it is in some ways (and it brought real improvements under the hood,) it's always been a toy by design. It's like a sleek, powerful motorcycle - with integral training wheels. Yes, they always wanted to get everyone to move to it eventually - but there was real resistance and in the early years of NT they seemed more than happy to pursue a dual product line and continue to maintain a version of DOS that came with Win32 compatbility etc. so that it still had access to their newer applications - without having to give up the old ones completely.
DOS had made them a ton of money, both directly and indirectly, and DOS was something many of us loved even if we loved to hate it. It let us take off the training wheels and get work done in a way that nothing Win32 ever approached.
But there was one longstanding problem with DOS from their point of view - Gary Killdall, by most reckonings the architect and primary author of DOS, had never licensed it to them or been their employee. MS had bought DOS from a hobbyist who had gotten an early sample of the new CPU and managed to hack up a functional version of CP/M to play with while everyone was waiting on DR to wrap a proper release for it. They had then sold this OS to IBM, who were similarly unwilling to wait for DR to wrap it up proper for them. Now, if this were just a matter of matching the publicly known interfaces, it wouldn't have mattered at all (this was before the invention of modern "intellectual property") but from the very beginning they had every reason to think it involved large scale copying instead.
CP/M, though not 'Free Software' was at least 'Software' - that is, licensees got access to source as well as binaries. So they bought what was openly described as 'a quick and dirty cp/m clone' by an author known to have access to the actual CP/M source code and at the time no one cared but even back then a man with the business smarts of Bill Gates probably realized it would bite him in the butt eventually.
This hadn't caught up with them yet when they first started working on NT though, and notice that the early versions of NT were not aimed to replace the DOS line in 'hobbyist' applications - it was mostly aimed at business users, the one area where there had always been some demand for crippled or toy systems - systems that, in theory, would be harder for poorly trained employees to screw up rather than systems that would offer advantages to those who were already using PCs fluently.
NT had little to offer for the latter market, and little appeal beyond 'be the first one on your block.' The first one to struggle with lack of drivers, the first one to come face to face with the reason why software means source when you hit a showstopper bug that you can't possibly fix and the manufacturer doesn't care about. I definitely played with it, but it was a step back from 95 as a workstation, and still inferior to Novell as a server for years to come. And if you didn't mind tinkering, why not tinker with Linux, where you get the actual source code and can fix the odd misplaced semicolon yourself if you have to? NT found basically no traction in the home, and even Microsoft didn't seem to consider that a problem.
But time went on and by early 2000 Killdall's copyrights finally caught up with them. Their lawyers couldn't stall it anymore, they were forced to settle. The terms were confidential but clearly included large back-payments, and were known to also include a source-code access provision to allow Caldera to more easily work around manufactured incompatibilities in the future. It may have also included per-unit royalties going forward. At any rate, by the end of the year Microsoft had terminated the DOS product line and by the end of the year after they had XP on the market. And after 20 years of working to put DOS on every PC, they were now working hard to keep DOS OFF every PC.
They did make a quick attempt to make XP more palatable outside of the office environment before release, but other than porting DirectX and getting some games working they didn't really do much on that front that I recall until the release of powershell.
If laughter is the best medicine, who are the best doctors?
(Score: 5, Interesting) by Nuke on Sunday August 26 2018, @07:59PM (1 child)
Killdall did not write DOS, nor was it written by a hobbyist. The original PC-DOS operating system was written by Tim Paterson while working for Seattle Computer Products [SCP] for their own 8086 CPU board (a micro computer kit component). Microsoft bought DOS from SCP and hired Paterson to adapt it for the IBM PC. They then sold it at a massive profit to IBM and other PC makers. Later, SCP successfully sued Microsoft for misrepresenting the intended use at the time of the sale.
Killdall wrote the rival CP/M system, which was also available for the early IBM PC, but it failed as being priced too high.
http://www.patersontech.com/dos/byte%E2%80%93history.aspx [patersontech.com]
http://www.billgatesmyths.org.uk/dos.html [billgatesmyths.org.uk]
(Score: 5, Interesting) by Arik on Sunday August 26 2018, @09:02PM
The hobbyist in question. And yes, he was a hobbyist. Anyone that was competent to port CP/M at the time would have been described as a hobbyist, even if they were also a working professional. And he did indeed describe it as a 'quick and dirty' hack of CP/M for the new processor. And he did indeed have a printout of the CP/M source code available for a starting point.
Since they settled all the exhibits from trial is sealed, but the outcome itself seems a pretty strong indication that large parts of it were copied wholesale in the process.
If laughter is the best medicine, who are the best doctors?
(Score: 2) by shortscreen on Sunday August 26 2018, @11:22PM
I've been hearing this story for twenty years. I don't know why people love it so much.
If Win95 is not a "real" OS by your definition of OS, that's fine. But it must not have anything to do with DOS because DOS is hardly more than a file system driver and some basic I/O functions. Even memory management was mostly handled by separate installable drivers (ie. HIMEM). Win95 has its own drivers to replace that stuff. Even Windows for Workgroups had an optional 32-bit file system driver to take over for the 16-bit DOS one.
WinNT came first, but it was too bloated to run on what was affordable hardware of the time. And it was less useful because it couldn't run DOS stuff that people needed to run. Marketing was not yet sufficiently advanced at that time to spin this as a good thing.