For several years now I’ve been trying to continue the DOS history series and write (or rather finish) a DOS 5 page. While tracing the history of DOS 1.0 or 2.0 is quite difficult and the amount of source material is very limited, with DOS 5.0 there’s the opposite problem, too much information. Way too much.

Thanks to various lawsuits, thousands of internal Microsoft documents were made public. Dozens if not hundreds are relevant to DOS 5. There are documents which outline the development plans in detail, and there’s even a fairly comprehensive post-mortem report which is a great source of information about what actually happened (as opposed to the usual Microsoft pie in the sky unrealized plans, like a DOS 5 release in late 1989).

Finding the relevant documents is not easy. Some are long e-mail dumps where only a few bits are pertinent. Others are awful scans which defeat any OCR, but are still readable by someone who has a bit of context information. There is a lot to go through.

The next problem is how to condense the huge amount of raw information into something informative, readable, and accurate, without ending up like the infamous Wikipedia FAT entry which bears no resemblance to an encyclopedia article and makes a solid argument that too much information its just as bad as too little.

But I’m trying.

This entry was posted in DOS, PC history, Random Thoughts. Bookmark the permalink.

14 Responses to DOS 5: TMI

  1. Yuhong Bao says:

    DOS 5.0 and Windows 3.0 is probably worth mentioning. In retrospect both Windows 3.0 and MS C 6.0 probably should have been delayed.

  2. Chris says:

    If you need some help with converting the awful scans to searchable PDF’s, I’m willing to convert some. I think it’s important for this and future generations to have a full searchable computer history.

  3. dosfan says:

    Would it help to split the DOS 5 page into two pages: one for technical info and another for behind-the-scenes business stuff ?

  4. Michal Necasek says:

    That would be cool, but an awful lot of tedious manual work would be needed. It is really tens of thousands of pages. There are many files that are not OCRed (or not completely) but basically legible, like this one: There are others which are marginal, like this one: That is where a human can do a much better job with knowledge of the subject matter. But some are pretty hopeless, like It is actually possible to make out something, but not enough.

  5. Michal Necasek says:

    Yes. Or at least clearly separate it into business/history part and a technical part. The technical bits are comparatively much easier 🙂

  6. Actually, I think I could transcribe that last PDF for you if you want; it gets considerably more legible zoomed in.

  7. Michal Necasek says:

    Can you get it 100% though? It’s possible to guess most of it, but probably not everything.

    It also occurred to me that maybe higher quality scans exist somewhere. Probably not public.

  8. Almost 100%; I can make out the entire subject and body except for a total of five words/abbreviations in the last three body paragraphs, and even the date headers are (barely) legible.

  9. Michal Necasek says:

    How well can you do with the org chart on page 5 of ?

    What I noticed is that the PDF viewer used matters. I don’t think any of them does it “wrong”, but some end up with less illegible text.

  10. Can’t make out more than a few words on the chart, alas; the tiny font used is really a killer, and not in a good way. (The rest of the document is actually more legible than the previous barely-legible one, though; you know what they say about clouds and silver linings.)

    >What I noticed is that the PDF viewer used matters. I don’t think any of them does it “wrong”, but some end up with less illegible text.

    I’m using Chrome’s PDF viewer; I also find that zooming way in helps quite a lot.

  11. If you want, I could try to get together a diagram of that chart with the few words that I was able to make out…

  12. Michal Necasek says:

    Nah, not necessary. I just wanted to make sure I wasn’t missing anything obvious — I’m not, it’s really illegible 🙂

    There are some other versions of that org chart in other documents I think, some more legible. The names and positions of the people involved are mentioned elsewhere in the documents, so most of it could be reconstructed. But it’s not that important.

  13. Jack says:

    MS-DOS 5.00 Project starting in 1989. MS-DOS 5.00 Codename was lifeboat. MS-DOS 5.00 first MS-DOS Beta Builds numbers. MS-DOS 5.0 was post be completely finish in June/July 1990 and it relay for 1991. MS-DOS 5.00 Beta Builds made in 1989 these Alpha Builds. MS-DOS 5.00.224 Beta Dosshell have name Program Manager, File Manager and Color Manager like Windows 3.00. Windows 3.0 and MS-DOS 5.00 was in development at time. It will be nice to have MS-DOS 5.00 selections.

  14. Michal Necasek says:

    Lifeboat was actually the codename of PC DOS 4.1 (OS/2 1.x used boat codenames like Sloop or Cutter, OS/2 2.0 was Cruiser). When Microsoft took over, they didn’t follow IBM’s plans, although there were some things in MS-DOS 5.0 (like high memory support) that were started as part of Lifeboat.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.