This article is within the scope of WikiProject Software, a collaborative effort to improve the coverage of software on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Dos as a general term cannot be an "invention"
I could understand it if CP/M was indicated as having been invented by X or Y , or any actual implementation of a "disk operating system" . The notion DOS however cannot be an invention. — Preceding unsigned comment added by 188.8.131.52 (talk) 23:37, 16 June 2012 (UTC)
Ask what DOS means and 'Disc Operating System' trips lightly off the tongue of the expert, and apparently nothing more needs to be said, the meaning being self evident. But that phrase is ambiguous. Everyone assumes it means 'an operating system for discs' but it could equally well mean 'an operating system ON disc'. I have been told by someone who was there that whilst everyone assumes the first actually the second interpretation is the origin - early computers loaded their operating systems from magnetic or paper tape, a real hassle, but then the wonderful convenience of loading the operating system from disc appeared, and swept the scene. The word 'DOS' was to contrast its superiority to 'MTOS' or 'PTOS'. DOS presumably did MUCH more than just operate the discs; why would it's name just concentrate on that aspect? 184.108.40.206 (talk) 20:16, 2 September 2012 (UTC)
In spite of what an over-excited reviewer wrote in 1982, no-one else has called the original PC DOS (MS DOS, whatever) a "scaled down version of XENIX". Hard enough to find sources, and sometimes the sources are wrong. Besides, XENIX didn't really take off till after the introductino of the PC. --Wtshymanski (talk) 19:21, 19 October 2013 (UTC)
Good lord, you're calling a lengthy, detailed article in BYTE—BYTE—in the magazine's prime years "padded" and "over-excited"? The author said that it somewhat resembled Xenix/Unix, which DOS clearly does; using \ instead of / for directories, and copy instead of cp (or pip), are examples. And why does the date of Xenix "taking off" meaningful for the comparison, anyway? Xenix never commercially took off anywhere, except perhaps on the Tandy 6000 series. Bottom line: If a reliable source makes a statement you disagree with, you'd better find other reliable sources that agree with you. WP:IDONTLIKEIT isn't good enough. Ylee (talk) 19:27, 19 October 2013 (UTC)
Anything not chiselled on stone tablets carried down the mountain by the prophet is susceptible to error. The long detailed articles are about the whiz-bang new IBM PC, not about its DOS. The one (1) line in the Williams article says something to the effect that MS DOS is a scaled down version of XENIX with no further discussion or explanation. The much longer "Genealogy of MS-DOS" in Ray Duncan's "Advance MS-DOS" (1986, Microsoft Press, ISBM 0-914845-77-2) pages 3-8, doesn't mention this Wikipedia version of the origin of MS-DOS. Duncan does mention some UNIX_like features of MS DOS in version 2.0, but XENIX does not appear in the history section of this book nor its index. Surely a Microsoft employee would be in a good place to at least *mention* this alleged influence on the development of MS-DOS? --Wtshymanski (talk) 19:48, 19 October 2013 (UTC)
So far, you've:
* been wrong about Duncan's book; Xenix appears several times in reference to DOS. For example, "As discussed in previous chapters, MS-DOS is largely compatible with both UNIX/XENIX and Digital Research's CP/M, and was designed this way to ease the porting of applications into MS-DOS from those two environments" (p. 110).
* backpedaled on BYTE 's quality. Is it reliable, or "isn't very competent"/"over-excited"/"padded"? You've called it all of the above.
* been wrong about whether Xenix predated DOS. When I pointed this out you promptly pretend you never said it, instead mumbling something about the date it "took off".
Have I missed anything? Ylee (talk) 19:57, 19 October 2013 (UTC)
┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ It's a pity that this all important XENIX/UNIX compatibility (or rather similarity in the ways file handles are used) didn't make it to the index under either UNIX or XENIX, and that Mr. Duncan didn't think this was an important enough influence to include in the "Genealogy" chapter. Obviously the one line in BYTE outweighs the whole book by Duncan. I don't think 1 line in a product preview in BYTE is enough to claim that XENIX was a significant influence on MS DOS. If XENIX was never commercially important, why would it have been an influence on MS DOS? --Wtshymanski (talk) 20:54, 19 October 2013 (UTC)
Obviously when I said "incompetent source" in the edit comment, the only Wikipedia compatible way of interpreting that is that I meant that every word ever published in BYTE through its entire history was a baldfaced lie. Correct? --Wtshymanski (talk) 21:07, 19 October 2013 (UTC)
Mr. Duncan says on pg. 64 that the "handle" based I/O functions work "in a manner very similar to to that used under UNIX/XENIX" but on pg. 68 and following says the handle-based function s were in MS DOS 2.0 and later. So there couldn't have been that much influence of XENIX on MS DOS 1.0. Unlike the influence of CP/M 80 which already had enough commercially important applications available in 1981 to make "transferability" a concern in laying out MS DOS. Did XENIX have drive letters? --Wtshymanski (talk) 21:19, 19 October 2013 (UTC)
Oh, good grief. So Duncan, who you cited as your great trump card, now also must be ignored despite his statements about the resemblances of Xenix and DOS given that you were yet again shown to be wrong. By your logic, since DOS didn't have subdirectories until version 2, that's even more proof that DOS 1 doesn't resemble Xenix in any way, right? (In any case, the Wikiedit in question merely states, as does the Williams cite, that DOS resembles Xenix and CP/M. The cite is clear that the resemblances in both cases are superficial; Williams wrote that the resemblance to Xenix was a little closer, but my edit doesn't get that granular, only mentioning the resemblance without any qualifiers.)
No, Xenix was never hugely successful for Microsoft compared to DOS, but for years Xenix was what it and outsiders believed was Microsoft's future, as part of a more general "Unix is the future of the industry" belief. This didn't change until the company and IBM began the OS/2 project.
Again, make up your mind. Is Gregg Williams, one of BYTE 's most prolific writers (his name is everywhere in the magazine in this period), a reliable source or not for a brief, very innocuous edit with supporting evidence I can cite from elsewhere? (I haven't even bothered to cite Duncan in the article but might just do that if you keep blathering here, perhaps with a strengthened discussion of intentional compatibility/resemblances.) Ylee (talk) 21:28, 19 October 2013 (UTC)
Hey, guys, you know that other editors are rolling their eyes on reading all this. None of these comments seem helpful.
Myself, I’d assume BYTE is a reliable source by default. But of course, they can make mistakes (just like all of us). If so, we just need another source or two that contradict it. Without that, even if BYTE is ‘wrong’, it can still be used as a source. Wikipedia is not here to find the ‘truth’, and certainly not to try to argue about it uselessly; Wikipedia just reflects what’s published. --A D Monroe III (talk) 21:32, 19 October 2013 (UTC)
You're absolutely right; I'm not being helpful, and I apologize for that. It's something I need to work on.
I completely agree that BYTE, like any RS, can be wrong. As you say, the way to show that a RS is mistaken is to find other RSs. As I wrote in my first comment in this thread: "Bottom line: If a reliable source makes a statement you disagree with, you'd better find other reliable sources that agree with you. WP:IDONTLIKEIT isn't good enough." Ylee (talk) 21:59, 19 October 2013 (UTC)
There is an altogether simpler explanation for this discussion. Our resident tendentious user, Wtshymanski is at it yet again. This is a now all too familiar story. Wtshymanski believes some obscure point that he alone believes, and insists on hammering it into an article regardless of what other users try to tell him and what the reliable sources say. How do we know that it is an obscure point? Simple: he produces absolutely zero sources to back his claims up largely because there are no sources. There then develops a long long thread on the talk page. For other examples of his tendentious behaviour see Talk:Power factor and Talk:Electric stove. There is no shortage of other examples.
Wtshymanski, just drop it and walk away. I B Wright (talk) 08:45, 20 October 2013 (UTC)
Sigh. How many times do we have to put up with this tendentious behavior? Once again, Wtshymanski is wrong on some technical point. The fact that DOS resembled both CP/M and Xenix is well known, but more importantly, the claim in the article is supported by an impeccable source.
Once again, multiple editors disagree with him on the technical issue and nobody agrees with him.
Once again Wtshymanski only cares what the sources say as long as he thinks they support his theories.
Once again, nothing anyone says on the article talk page causes Wtshymanski to change his position (See Power Factor for an especially egregious example of this)
Once again we have the sarcastic edit comments. Once again we have an edit war (Side note to Ylee; it takes two to edit war, and you have been around long enough to know better.)
And of course we have the standard snide remark on his user page. Did I miss anything?
I, for one, am getting tired of playing Whac-A-Mole in multiple engineering articles. This needs to dealt with for what it is; another WP:FRINGE theory from the engineering equivalent of a conspracy theorist. --Guy Macon (talk) 15:16, 20 October 2013 (UTC)
More tendentious-editor food: BTW, I timed it; it took 7 minutes to find these.
From a Microsoft print ad:
"MS-DOS 2.0 and XENIX share hierarchical file structure and I/O redirection, including piping. MS-DOS 2.0 also provides XENIX-compatible system calls".
"Initially, Microsoft had big plans for XENIX and tried to insure DOS compatibility.
That has tremendously positive influence on MS DOS which in versions 2 and 3 acquired several Unix based innovations like hierarchical file system, pipes, redirections, etc, moving father and father from original CP/M...
Its clear that work on XENIX and DOS proceed largely in parallel and some cross-pollination: XENIX greatly influenced DOS design.
For example Microsoft Press' "MS-DOS Encyclopedia" shows a reproduction of a late DOS 1.25 OEM brochure which mention such future enhancements to DOS 1.25 as XENIX-compatible pipes, process forks, and multitasking, as well as "graphics and cursor positioning".
That shows that Microsoft certainly tried to bring those two OSes closer, but the forks, multitasking, and multi-user support never materialized.
Oddly, the flyer claims: 'MS-DOS has no practical limit on disk size. MS-DOS uses 4-byte XENIX OS compatible pointers for file and disk capacity up to 4 gigabytes."
From the book Inside XENIX:
"The first version of MS-DOS and PC-DOS was very much like CP/M, but the second version introduced some of the fundamental features of UNIX.
These features, including 1/0 redirection and tree-structured directory systems, are quite independent of whether the system supports a single user or many and show the strong influence of UNIX.
An example of a UNIX-like feature found in MS-DOS is redirection through the use of less-than (<) and greater-than (>) symbols.
These symbols allow a programmer and ordinary users to specify any destination, for example, the screen, printer, communications line, or even a disk file for the output of programs.
The symbols also allow input to programs that come from any source, including the keyboard, communications line, or an ordinary file.
In addition, we can use the vertical bar symbol (:) to set up "pipelines" in which the output of one program is fed as the input to another.
These pipelines conveniently combine small stand-alone programs to form larger programs that accomplish complex tasks, such as report generators; word processing tools, such as spelling and grammar checkers; and program generators.
Tree-structured directories also are familiar to MS-DOS and PC-DOS programmers.
These directories allow users to organize information in terms of categories within categories.
At each point in the tree, subdirectories can be given meaningful names according to the information they contain."
From the book Kirshna's Computers and Languages by Dipak Jha:
"In 1983, version 2.0 of MS DOS appeared...
To simplify management of [the PC/XT 10MB hard disk] version 2.0 introduced a hierarchical file directory structure based upon the UNIX model.
Microsoft has recently developed XENIX, its own variant of UNIX; it was thought sensible to move the two systems closer together..."
Guy, unfortunately, the cites don't support my edit in question because they all discuss DOS 2.
The BYTE review is of the original IBM PC and DOS 1.0. It says that DOS 1 resembles (in terms of its prompts, command names, etc.) both CP/M and Xenix, calling DOS 1.0 a "scaled down version of Xenix". The edit in question merely says that DOS 1 resembled both CP/M and Xenix, which I didn't expect would be controversial. In any case the BYTE review is still sufficient, I believe, to serve standalone as citation. Ylee (talk) 05:04, 21 October 2013 (UTC)
I do not support the edit war between Wtshymanski and Ylee at all, but Ylee's sentence in the article, which now effectively states that 86-DOS as well as released versions of MS-DOS 1.xx and PC DOS 1.x would have been influenced by XENIX in any way, is factually wrong and therefore must be corrected for historical correctness. We should make clear, that XENIX-like features first appeared in what we typically call "DOS 2".
86-DOS was modelled after CP/M-80 (and, according to Tim Paterson, also resembles some ideas found in Northstar DOS). 86-DOS 1.10 became MS-DOS 1.10, the first version of MS-DOS (there never was a MS-DOS 1.00 version, although many books state so when they incorrectly lump together early MS-DOS and PC DOS version numbers). Either MS-DOS 1.11 or 1.14 became PC DOS 1.0, and somewhat later MS-DOS 1.24 became PC DOS 1.1. In 1982-03, MS-DOS 1.25 was the first version of MS-DOS available to OEMs different from IBM, Microsoft and SCP. None of these operation systems was in any way influenced by XENIX (or any other Unix alike). Like CP/M, they all used FCBs (file control blocks) for file I/O - the concept of file handles was completely absent. These versions also did not support sub-directories, piping and redirection, background processes, networking or a switchar different from '/'. Non-XENIX features also not found in these versions of DOS were a CONFIG.SYS configuration file, dynamically loadable device drivers or TSRs. The internal data structures and interfaces to support any of this simply did not exist at this time.
The exact moment, when Microsoft started to plan to introduce XENIX-like features into DOS is difficult to determine, but given that Gregg Williams already speculated about it the January 1982 release of BYTE ("However, the IBM DOS is a scaled-down version of Microsoft's 16-bit Unix look-alike, the Xenix operation system.") seems to indicate that the "wishlist for DOS 2" was already open at this time and either Microsoft or IBM had told him about it. Perhaps, his statement is the result of some over-simplification for casual readers or he simply mixed up what he was told. In either case, "as is", his statement is factually wrong in 1982-01. There was zero XENIX-influence in these early versions of DOS, and none of the XENIX-like features found in later issues of DOS were implemented or even worked on at this time.
Work on the future "DOS 2" features began in 1982-03. Support for dynamically loadable drivers was added in 1982-03 for MS-DOS 1.27, support for safe context switching and background processes was started in 1982-03 as well, redirection and piping were added in 1982-04 for MS-DOS 1.28, sub-directories were added later in 1982-06 for MS-DOS 1.40 and the support for handle-based file I/O and the corresponding new set of API functions were finished with MS-DOS 1.54 in 1982-07. A configurable switchar (either '/' or '-') was also supported at this time. To the general public, all these features first showed up in PC DOS 2.0 (and MS-DOS 2.0), and that's why they are typically named "DOS 2" features.
As I understand it, DOS 1 and DOS 2 were not even the same product or even produced by the same organisation. Although DOS 2 was a Microsoft product through and through, DOS 1 was bought in from a third party supplier. Unfortunately, neither of our recollections or beliefs are citeable material in Wikipedia (and that applies to all editors including Wtshymanski). It has been suggested that the BYTE reference may be an unreliable reference, but in Wtshymanski's own words, "A poor reference is better than no reference". If the article is to be altered from its currently (when I last looked) referenced position, then good and reliable references are required to support the change. Until that happens, this discussion is going nowhere. I B Wright (talk) 09:57, 21 October 2013 (UTC)
(Goes away and does some research) Ah. I see what you (Matthiaspaul) are talking about. In my mind I was lumping DOS 1.28, DOS 1.40 and DOS 1.54 as "Dos 1". I am not calling them MS-DOS because, as I remember it (and it was the first version of DOS I learned), 2.0 was the first one to by called MS-DOS. I am fine with anything (including Wtshymanski's preferred version) as long as it is sourced. My memories might be a useful guide as to what to search for, but they can not be the basis for anything in the article. In fact, making the article match your personal knowledge while saying "the sources are wrong" is a hallmark of tendentious editing. You have to follow the sources whether you like them or not, and not make changes when you have no sources backing you up.
Getting back to similarities and reusing ideas, file control blocks, file handles, volumes, subdirectories, etc. could be thought of as non-core features on a computer that could and did run just fine with a cassette tape and no hard or floppy disks. The core feature that is found on XENIX, CP/M, and all versions of DOS is the basic concept of programs not directly accessing hardware peripherals, as was common on systems such as the Commodore 64 and even some early mainframes. The programs talk to the operating system (today we would call it using the API -- does anyone remember when that terminology became popular?) I think that isolating the programmer from the underlying hardware by providing a set of uniform interfaces to the various hardware is pretty much the core feature of XENIX, CP/M, and DOS, and they all got the idea from UNIX. --Guy Macon (talk) 10:04, 21 October 2013 (UTC)
Let's try to identify the actual core of the problem and how we can solve it within the limits of our policies and guidelines. Actually, it isn't that difficult IMO.
Originally, the article was incomplete, but correctly stated:
"IBM PC DOS (and the separately sold MS-DOS) and its predecessor, 86-DOS, were loosely inspired by Digital Research's CP/M, which was the dominant disk operating system for 8-bit Intel 8080 and Zilog Z80 based microcomputers. PC DOS ran on Intel 80x86-family processors."
At present, it incorrectly states:
"IBM PC DOS (and the separately sold MS-DOS) and its predecessor, 86-DOS, resembled both Digital Research's CP/M—the dominant disk operating system for 8-bit Intel 8080 and Zilog Z80 based microcomputers—and Microsoft's Xenix variant of Unix. PC DOS ran on Intel 80x86-family processors."
If we would change that to: (or something along that line)
"IBM PC DOS (and the separately sold MS-DOS) and its predecessor, 86-DOS, resembled Digital Research's CP/M—the dominant disk operating system for 8-bit Intel 8080 and Zilog Z80 based microcomputers. DOS instead ran on Intel 8086 16-bit processors. Starting with (MS-DOS 1.28 and) PC DOS 2.0 the operating system incorporated various features inspired from Xenix, Microsoft's variant of Unix."
we could include the previously missing Xenix information without stating anything untrue. There are uncountable sources supporting the CP/M influence, and there are also lots of sources stating that DOS 2 added various Xenix-like features. The only part for which it might be difficult to find public sources (but which is nonetheless true) is the "(MS-DOS 1.28 and)" bit - in the worst case we could just skip it and still not state anything untrue. (I don't need to have that in the article, it could just serve as some further explanation for where some confusion might have come from. In either case, this is bonus background information and wasn't subject of the edit war between Wtshymanski and Ylee, so we can ignore it for now.) We could also say: "MS-DOS/PC DOS 2.0 (and higher) also incorporated various features inspired from Xenix, Microsoft's variant of Unix." (thereby simply ignoring the fact that Microsoft must have been working on these "DOS 2" features for some while - and was, but not before MS-DOS 1.28 in 1982-04.)
Regarding the BYTE article, WP:V and WP:RS certainly do not make it mandantory to use any source in existance and deliberately seek for incorrect statements in them just to pick and add these inaccuracies into the article. After all, our goal is to present as much accurate information as possible. So we could simply choose to ignore that particular statement in the BYTE article (until someone finds other sources supporting it - but this will never happen, as versions of MS-DOS before 1.28 (and thereby also any version of 86-DOS and any version of PC DOS 1.x) simply were not influenced by Xenix in any way. This is not a matter of interpretation, but a historical fact.) Things would be different, if "86-DOS and PC DOS 1.x were influenced by Xenix!" vs. "DOS 2 was influenced by Xenix!" were two competing views in the industry. In that case, we would have to present both views in the article. But there never was any dispute about it (statement 2 is correct, and so far nobody came up with statement 1, if we don't count that odd statement in the BYTE article), so it is irrelevant for the article (for as long as we don't state anything untrue).
I like Matthiaspaul's solution. --Guy Macon (talk) 11:41, 22 October 2013 (UTC)
I am OK with the proposed text. Ylee (talk) 17:55, 22 October 2013 (UTC)
Fine. :-) While I wished the other participant in the edit war, Wtshymanski, would have commented on my suggestion as well, he has vanished for a week now. Since the proposed text includes and refines both positions, I think it is safe to assume that he would be fine with it as well. Therefore I have updated the article now. --Matthiaspaul (talk) 14:36, 27 October 2013 (UTC)
This is behavior that we have come to expect whenever Wtshymanski rediscovers the first law of holes. The pattern will continue in whatever engineering article he targets next. The good news is that he does eventually give up when it becomes apparent that his latest fringe theory isn't going to fly. --Guy Macon (talk) 16:56, 27 October 2013 (UTC)