Talk:IBM PC DOS/Archive 1

Archive 1

Merge

I propose merging List of IBM PC DOS versions with IBM PC-DOS on the latter's title. "List of IBM PC DOS versions" does not add anything new that IBM PC-DOS doesn't already discuss. SchuminWeb (Talk) 15:53, 4 August 2007 (UTC)

three major operating systems? CP/M, DR-DOS, Commodore, Macintosh, Pro-DOS?

PC-DOS was one of three major operating systems that dominated the personal computer market from about 1981 to 1995.

This introductory sentence is rather ambiguous. If by "personal computer,"- specifically x86 PCs and IBM "clones" is meant, than the third operating system would proably be CP/M by Digital Research. It could also mean Digital Research's DR-DOS which was aquired by Novel and renamed DR-DOS. If one includes Macintosh, it would be the Mac OS, I assume. If one includes Commodore and/or the Apple II as "personal computers," than PC-DOS might not even make the top three. I'm going to remove the sentence, although it could be re-written with a more specific reference and citation. Cuvtixo 17:20, 21 August 2007 (UTC)

IBM offered a choice of PC-DOS (16-bit), AIX (a unix), and CP/M (8-bit) with their computers. AIX and CP/M were priced high to discourage their purchace. In practice, most computers in the 1980s were either eight-bit stuff or ran a clone of PC-DOS as marketed by Microsoft. The dominate eight-bit OS was ROM-BASIC. DR-DOS found some OEM market, but it was not until the release of 3.31 (a success), and 4.0 (a failure because of alpha code), that customers had choice on their DOS. Microsoft really did not start to dominate the market until something like Windows 3.00 and DOS 5.0. --Wendy.krieger (talk) 12:05, 29 March 2010 (UTC)

Its strange, but I actually disagree. CP/M was not even a contender for anything after 1982, and Osborne and Kaypro all but disappeared entirely. AIX which shipped in 1985 was never big enough to count in retail sales. it was never even mentioned as such, neither was Xenix. DOS was king, Pro-DOS all but died out after the Macintosh System Software shipped a few hundred thousand units. DR-DOS sold more units than CP/M+Commodore+ProDOS combined. The three OSs were DOS, Macintosh and DR-DOS.

DOS shipped with almost every IBM PC used by business, and sales of spreadsheets and word-processors. Within its first year, Microsoft was #1, and Lotus was #2. They literally dominated the market until the Macintosh shipped. No one else even came close. As I saw, and worked on, used, and repaired and upgraded, I only saw 4 Xenix machines, 3 of them dedicated to a SABRE training program.

I stand corrected. [1] The Macintosh was insignificant until the early 90s. Commodore sold the C64 in better than 20% of the market for more than 4 years. Apple never has done that. The two OSs that dominated the Market were PC-DOS, and The C-64. —Preceding unsigned comment added by 69.232.193.99 (talk) 10:08, 3 March 2011 (UTC)

IBM-DOS 6.10

There really was a short-time release of an what the manual describes as "IBM-DOS 6.00", what the "dosver" describes as IBM-DOS 6.00, but command.com's ver says IBM-DOS 6.10. This is in a similar packaging to the beta, but was released as a retail upgrade. It features the black icon that was used in PC-DOS 5.0.

There was also a PC-DOS 6.1, which had a similar packaging to PC-DOS 6.3, and features a five-disk install.

If one googles for 'PCDOS613', one will discover that there was also a short-time free stepup from pcdos 6.10 to pcdos 6.30. A similar step-up exists for 5.00 to 5.00.1, and for 7.00 to 2000. —The preceding unsigned comment was added by Wendy.krieger (talk • contribs) 09:51, August 23, 2007 (UTC). Wendy.krieger 09:53, 23 August 2007 (UTC)

IBM DOS history

The split between MS and IBM came while came while DOS 5 was still in the market. IBM decided to go from an OEM only to a compteing DOS, comparable to Microsoft DOS, beginning with DOS 5.00.01. The first retail version is a modified version of IBM-DOS 5, largely to run on generic hardware.

Such a move would also be necessary, because IBM no longer needed to put ROM-BASIC into their machines, and the bios of later machines would be more varied BIOS.

None the same, DOS 5 was released before the split, and the bulk of the system features were set.

Because of the split, the OEM version of MS-DOS was labeled as "MS-DOS and Optional Tools". Together these form the same package as the upgrade version since Optional Tools is different to Supplemental Tools. It is not clear exactly what is optional, but IBM licenced the base code and recompiled it as IBMDOS 6.0. The original package has a manual that has the same cover as the DOS 5 manuals, and it "Starting IBMDOS....". The manual and true-version is 6.00, but command.com says 6.10.

The whole package was overhauled with new manuals, icon and other things, and released as PC-DOS 6.1, Internally, it is still PC-DOS 6.00 (by true-version). Disk compression was by superstor, while many of the optional tools were old-version PC-TOOLS utilities.

PC-DOS 6.3 was a release of PC-DOS 6.1 with many of the fixes that make 6.20 different from 6.00. The DOS in OS/2 PPC is PC-DOS 6.3, with some features (rexx), that would make it into PC-DOS 7. IBM had offered also a free upgrade from DOS 6.1x to 6,30, although this is not widely advertised.

PC-DOS 7.0 replaced many of the features of MS-DOS, such as using OS/2 style PACK2 files, rather than the Compress/expand files. It includes also REXX, and VIEW, with Stacker 4 for DOS and Windows for compression. The manual warns not to install if you use Stacker 4 for DOS and OS/2.

After this DOS, IBM shut down both Boca Raton, and the PC Company. DOS and OS/2 became moribound after this.

PC-DOS 2000 is basically PC-DOS 7.0 with fixes slipstreamed. The manual was a poorer quality printing and cover. In stead of a standard 1.44 disk, and 4 * 1.68 XDF disks, this release was on six standard diskettes.

The PC-DOS download at IBM is actually a chinese version, as can be seen by -ZN- and -PRC-, the latter = peoples republic of china.

IBM continued with a subset of PC-DOS 7.0, with support for FAT32, and little else. The thing has been recompiled several times, some for external use [eg Norton Ghost].

See eg "Upgrading and Repairing PCs - Scott Mueller - Edition 6 - ch 22. Earlier versions had a detailed account of the early versions of DOS.

--Wendy.krieger (talk) 04:08, 3 February 2008 (UTC)

Split_section_out_proposal

Common discussion held at http://en.wikipedia.org/wiki/Talk:MS-DOS#Split_section_out_proposal

See the talk at Talk:DOS 130.101.20.153 16:25, 4 December 2007 (UTC) / discussion moved to busiest page when connecting the links properly. // FrankB 05:41, 10 December 2007 (UTC)

Merges of IBMBIO.COM and IBMDOS.COM

These terms are not specific to PC-DOS; DR-DOS used the same names, as did MS-DOS in its early versions. Merging these here would be like merging steering wheel into Ford Motor Company or leaf into ash tree. JeremyMcCracken (talk) (contribs) 23:27, 10 January 2009 (UTC)

I don't like merging these into DOS either, because now we're moving the other way. Not all DOSes use them. MS didn't in later versions, FreeDOS has its own names, as did early versions of DR-DOS. These should be kept separate; there would be plenty of room for someone with a better knowledge than me of DOS' inner workings to add information. The fact that it hasn't happened yet doesn't necessitate merging them to something. I'm finding book results already; this one looks promising if I can find a print copy. JeremyMcCracken (talk) (contribs) 20:26, 11 January 2009 (UTC)

Probably use this book as well. JeremyMcCracken (talk) (contribs) 20:33, 11 January 2009 (UTC)

$10,000 DOS

Is this a joke or a typo? 2fort5r (talk) 23:26, 28 September 2009 (UTC)

Neither; take a look at the references for that section. I think it was intended as a commercial license. JeremyMcCracken (talk) (contribs) 20:59, 30 September 2009 (UTC)

Worth noting that if you buy an Lenovo desktop and do not order an OS on it, it comes with a PCDOS 2000 licence. 81.23.50.232 (talk) 18:13, 20 March 2010 (UTC)

Because of Anti-Trust?!!!

As a result of the antitrust agreement with IBM and the Department of Justice, IBM had to outsource the bulk of the computer components to outside companies

That is a funny claim to make, because in 1982, the year AFTER the IBM PC began shipping, the antitrust suit was dismissed "without merit." IBM had no reason to try in the year before that to comply with a ruling that would never happen! --Eliyahu S ^Talk 20:53, 24 May 2010 (UTC)

Remove Release History section

I'm removing the Release History section because it is redundant to the Versions section (which includes a link to Timeline of x86 DOS operating systems) which duplicates the list of releases.

I don't see the notability of "official product names". What is official anyway, the legal name given in documents such as license agreements, the name printed on manuals, the name printed on product packaging or the name displayed by COMMAND.COM when the OS boots up? Explanation has not been given, nor any references for "official product name" cited. In my view the official name is The IBM Personal Computer Disk Operating System, which by convention may be shortened to The IBM Personal Computer DOS, or further shortened to IBM PC DOS or simply IBM DOS or PC DOS. Again, I don't feel that the distinction between names is significant or notable.

As recently back as 19:07, 13 December 2007 there was an extensive list of versions in the Versions section. http://en.wikipedia.org/w/index.php?title=IBM_PC_DOS&direction=next&oldid=176934900

That list was removed and merged into Timeline of x86 DOS operating systems, but some of the old information was lost. Perhaps a version of the list should be reestablished here, containing only information not duplicated on the Timeline page, such as the hardware the release was distributed with. —Preceding unsigned comment added by 64.128.111.166 (talk) 15:52, 29 December 2010 (UTC)

I note that the name seems to have changed to IBM DOS around the time of release of the PS/2 and reverted back to IBM PC DOS around the time PS/2 was dropped. That would be notable. 75.63.204.178 (talk) 23:37, 29 December 2010 (UTC)

PC DOS 7.1

Sadly I can't confirm the year 2003 for "7.1", but "something" is certainly included in some obscure archives for deployment tools. This "something" is a PC DOS 7.x, it has a FDISK allowing to create FAT32, I can't vouch that it supports LBA or FAT32 outside of its FDISK, and its internal structures do not work with PC DOS 7 share.com (in other words, I can't merge the "7.1" parts in the wild with my complete PC DOS 7 +/- Y2K Euro additions). –89.204.153.226 (talk) 20:38, 5 November 2011 (UTC)

There are several revisions of PC DOS 7.1 from 1999 to 2003 based upon the copyright and build dates found in the core files (IBMBIO.COM, IBMDOS.COM and COMMAND.COM). In all cases the major new feature is FAT32 and LBA support. PC DOS 7.1 IBMBIO.COM clearly makes EDD Int 13h calls and has references to FAT32.

PC DOS 7.1 was never released as a retail product but came with several versions of Norton Ghost and what appears to be the last revision (build 1.32 dated December 1, 2003) was part of IBM's ServerGuide Scripting toolkit. Asmpgmr (talk) 04:52, 18 June 2012 (UTC)

Proposed merge of IBMBIO.COM and IBMDOS.COM articles into PC DOS article

It has been proposed to merge these two articles into this IBM PC DOS article. While both articles are still short and need expansion, I don't think it would be a good idea to merge them, since IBMBIO.COM and IBMDOS.COM are not specific to PC DOS. These filenames were also used by DR DOS 5.0 - 8.0 (with the exception of DR-DOS 7.06), PalmDOS, Novell DOS and OpenDOS, and possibly other DOS-compatible operating systems as well. There's a lot more that could be said in regard to these two files, their format and their functions in various operating systems. This would be way beyond the scope of the PC DOS article. --Matthiaspaul (talk) 01:21, 4 January 2012 (UTC)

I would suggest that the IBMBIO.COM article be merged into the IO.SYS article with IBMBIO.COM redirecting to IO.SYS since these are really the same system file under different names. IBM uses IBMBIO.COM while Microsoft uses IO.SYS. In both cases this is the system file containing the system initialization code and the builtin DOS device drivers.

Likewise I would suggest that the IBMDOS.COM article be merged into the MSDOS.SYS article with IBMDOS.COM redirecting to MSDOS.SYS since these are really the same system file under different names. IBM uses IBMDOS.COM while Microsoft uses MSDOS.SYS. In both cases this is the system file containing the DOS kernel. Asmpgmr (talk) 22:00, 12 June 2012 (UTC)

While this is more reasonable than the original proposal to merge these articles with the IBM PC DOS article (which I opposed above), please keep in mind, that the IBMBIO.COM and IBMDOS.COM filenames were not only used by PC DOS, but also by DR-DOS (and some other operating systems). While the functionality is similar from a bird's eye view, there are lots of interesting differences if you look closer, and DR-DOS provides several extensions in order to support multiple parallel installations of DR-DOS or PC DOS on the same disk. These features apply only to the IBM*.* filenames, not to the IO.SYS/MSDOS.SYS names. Even MS-DOS IO.SYS and PC DOS IBMBIO.COM of the "same" version are not really identical function-wise. And things get more complicated if you count in MS-DOS 7.x/8.0. So, assuming that such information will be added to the articles over time, I'm not sure, if merging them would be really helpful in the long run. Therefore, I would suggest to leave the names as they are right now and check again in a couple of years. --Matthiaspaul (talk) 23:07, 12 June 2012 (UTC)

From what I saw the difference between IBM and Microsoft was just names and copyright texts. Basically if you built DOS without specifying an "IBM" define you would get IO.SYS and MSDOS.SYS and utilities that referenced them (like FORMAT and SYS) would change accordingly. The code was essentially the same though you couldn't do a binary compare because of the differing names and copyright text lengths caused offset differences. Anyway if the IBMBIO.COM and IBMDOS.COM articles are to be kept then they should be updated to have the same essential information as their IO.SYS and MSDOS.SYS counterparts (obviously not including the MS-DOS 7.x merging). Asmpgmr (talk) 23:55, 12 June 2012 (UTC)

IBM DOS 4.0

I hope the DOS 4.0 section is one of the worst on Wikipedia, because it's truly awful. It is completely unsourced and contains bizarre claims. DOS 4.0 was for testing ideas for DOS 5.0 aka OS/2? To begin with, OS/2 1.0 was released more than six months before DOS 4.0. That seems like an odd way to test ideas for it. And what were those ideas, anyway? It would help to list at least one.

"Taking more than twice as much memory as DOS 3.30"? That's highly doubtful, unless it's an apples to oranges comparison with DOS 4.0 configured to use lots of buffers, FASTOPEN, etc. compared to DOS 3.30 in a minimal configuration.

Statements like "heavily rewritten" and "slipshod" had better be supported by some sources. And where did the nonsense about removing FCB support come from? That is utter fabrication. See "Developing Applications Using DOS" (1990) by Feigenbaum, Christopher, Saliga, ISBN 0471522317 for details of FCB implementation in DOS 4.0. IBM only shipped drivers for IBM memory cards with DOS... well, it was supposed to be used with IBM hardware, so why is that a bad thing?

DOS shell was "poorly designed and irksome to use"? This is supposed to be an encyclopedia, not an opinion piece. The DOS 4.0 section appears to illustrate the fact that DOS 4.0 had a bad name by repeating or inventing nonsensical "bugs" instead of documenting the real issues. Unless the section is significantly improved very soon, I'm going to strip it down to the bare verifiable essentials. Codegen86 (talk) 11:05, 15 April 2012 (UTC)

The stuff regarding the removal of FCB support has been deleted from the article as that is totally incorrect. Asmpgmr (talk) 23:58, 12 June 2012 (UTC)

This is the edit that introduced the "stripped out support for the obsolete DOS 1.x file control blocks" statement. Though not all of this IP editors "contributions" are incorrect, too much of this is wrong or opinion and you know it's all sourced off the top of his head. I'm inclined to revert all 3 of his/her edits to this article, as all of their contributions are questionable. Particularly irksome is that they inserted some of their unsourced information between sourced information I previously added, and the reference for that source, to make the article imply his stuff came from the source I cited. –Wbm1058 (talk) 03:24, 18 June 2012 (UTC)

The stuff about the removal of FCB support was absolutely wrong. FCB support was never removed from any version of DOS. What happened in PC DOS 4.0 was that the code which translated FCB accesses to SFT accesses was moved to SHARE.EXE for greater than 32MB partitions and a message was displayed if SHARE wasn't loaded on such systems. This code was reintegrated into the kernel with DOS 5.

As for his other edits. I agree that they seem to be mostly wrong. DOS 2 was indeed a major rewrite to add support for things like subdirectories and device drivers but it wasn't completely rewritten. Also the item about DOS 3.2 being the first retail version sold by Microsoft is wrong; MS-DOS 5.0 was their first retail version. I've just removed that sentence. Finally there is too much stuff specifically about MS-DOS which should be in the MS-DOS article instead of the PC-DOS article. Would you happen to have the old MS-DOS Encyclopedia ? That would be a great source of information. Unfortunately I don't have it. Asmpgmr (talk) 04:37, 18 June 2012 (UTC)

Yup. Got a lightly used one for a really nice price from amazon. Sorry the best price is higher now. Click on the ISBN to link to amazon. Wbm1058 (talk) 04:56, 18 June 2012 (UTC)

By all means use that to verify stuff. I believe it covers DOS up to 3.2. Also I rewrote the PC DOS 7 section since I was the lead on that version so I know what was done for it.

Just a thought: shouldn't the PC DOS and MS-DOS articles have similar structuring ? They currently do not.

Asmpgmr (talk) 05:11, 18 June 2012 (UTC)

The bit about MS-DOS 3.2 being the first retail version is almost completely wrong, but not quite. MS-DOS 3.2 was the first 'packaged product' version, sold by Microsoft to system builders. It was not a retail product, but it was the first generic MS-DOS (available in parallel to the classic OEM versions of MS-DOS). IIRC this is mentioned in the MS-DOS Encyclopedia. Codegen86 (talk) 14:26, 8 July 2012 (UTC)