Hi, On Mon, Feb 27, 2023 at 5:59 PM Aitor Santamaría <aitor...@gmail.com> wrote: > > As a matter of curiosity, given that we were unable to find a precise answer > to why Microsoft dropped IFS after MS-DOS 5.0 (although we had speculations), > I thought, why not ask ChatGPT (Bing) about that? > > I had hoped that there was some page that we weren't unable to find. Or even > it could have been better if ChatGPT had any access to information from > Microsoft that was not easily publicly accessible. > Neither seems to be true.
Search engines are notoriously bad for anything niche. Heck, even popular stuff gives lots of wrong answers. There's still info out there, but you have to ask the right person. Most stuff is tucked away in storage or hidden in some retired guru's brain. The tech world overall is sadly not very interested in archaeology, neither preserving nor emulating nor documenting "obsolete" stuff. "Today and tomorrow" are all they care about. Everything else isn't important to them (mostly). > As a lover of computing archaeology, I fear that when Mr. Gates passes away, > we will have to do with ChatGPT's answer. Oh well. I'm not understanding exactly what you want to know. (But someone like Michal Necasek of OS/2 Museum probably could answer.) PC-DOS was the version for IBM hardware, and MS-DOS was licensed to others. While Microsoft was mostly responsible for DOS (and OS/2), IBM still played a heavy role. At least Wikipedia says "MS-DOS 4" was mostly IBM's work. When IBM "fired" MS, (AFAIK) IBM still had access to sources for Windows 3.0 and MS-DOS 5.0. So they kept updating PC-DOS until, what, 2003? MS-DOS 5 came out in 1991, right? And OS/2 1.0 came out in 1987. OS/2 was supposed to be "the future", and MS was primarily responsible for OS/2 1.x (the 16-bit version). 32-bit OS/2 didn't come until 1992, and 32-bit Windows NT (aka "MS's portable OS/2 rewrite") was 1993. NT did support HPFS, for example, for a few years. So why did the "buggy" DOS 4's IFS interface not survive? They were focused on other OSes. Wikipedia says that OS/2 later got drivers for UDF and JFS. Obviously NT preferred NTFS (journaled, unlike HPFS, with larger max file size) and eventually dropped booting atop FAT (e.g. Vista) for security and performance reasons. It could also just be a RAM issue. IIRC, the standard answer why Windows tools won't create FAT more than 32 GB is because it's too slow (esp. for a real-mode driver). Maybe they never could happily get a working HPFS driver for DOS. (The 16-bit HPFS driver for OS/2 allegedly only supports a 2 MB cache maximum.) So if nobody uses it or it's too buggy or it only helps a competitor OS, then why bother? (Sigh.) For example, a few years ago Linux removed the ext3 [sic] driver because they only needed ext2 and ext4 (backwards compatible). The main reason for keeping ext2 as a separate driver was because it was much cleaner and simpler code. (Besides, I guess some storage media won't require journaling.) Sorry if I am confused, I'm just trying to understand it myself. _______________________________________________ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
