home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.advocacy
- Path: sparky!uunet!microsoft!hexnut!gordonl
- From: gordonl@microsoft.com (Gordon Letwin)
- Subject: Re: Can IBM write code?
- Message-ID: <1992Jul21.231022.5094@microsoft.com>
- Date: 21 Jul 92 23:10:22 GMT
- Organization: Microsoft Corporation
- References: <1992Jul13.215408.12815@engage.pko.dec.com> <2148@unet.UUCP> <1992Jul14.213905.2857@desire.wright.edu>
- Lines: 58
-
- In article <1992Jul14.213905.2857@desire.wright.edu> jstewart@desire.wright.edu writes:
- >In article <2148@unet.UUCP>, jrs@zippy.unet.com (John Switzer Frame 3.0) writes:
- >> For a previous job, I disassembled numerous bits and pieces of DOS because we
- >> needed to know the internals. In particular, I spent some time on DRIVER.SYS,
- >> which in DOS 3.3 was about 1.8K in size. In DOS 4.0, though, it was over
- >> 5K, but when I took a look at it I found that the basic core of the program
- >> was virtually unchanged - only one bug about 1.44MB formatting was changed.
- >
- > Correct me if I'm wrong, but isn't PC-DOS 99.9% MS-DOS. This brain
- >deadedness was almost certainly fromn MS.
-
- You're wrong. He states that he's comparing DOS 4 to DOS 3.3. DOS 4.00
- was entirely an IBM product. (MS produced 4.01, a bug fix release, but
- didn't substantially change anything).
-
- I worked on DOS 5 and in the process studied DOS 4 very closely; that
- ridiculous code plus a lot of other sh*t was indeed put in there by
- the brilliant boys in boca. As you may recall, DOS 4 ballooned in size
- vs. DOS 3.3 yet added little functionality. It didn't sell well for that
- reason, as well as the bugs that IBM installed. MS wrote DOS 5 which is
- much smaller and faster (even without HMA). I remember seeing a project
- planning memo re: DOS 5 which started out, essentially,
-
- "First we rip out all that IBM code"
-
- I can say from personal experience that it's a great pleasure to wield
- one's DELETE key like a mad slasher, in circumstances like this.
-
- In sum, it's fair to say that DOS 3.3 and DOS 5.0, successful and
- popular products, are "Microsoft writing DOS". DOS 4.00, bugs and
- obesity and all, are a good example of "IBM writing DOS". IMHO, the problem
- is based in IBM's inability - or unwillingness - to distinguish between
- their good programmers and their poor ones. I think that they feel that
- just as it would be discriminatory to give the important jobs only to
- men, or whites, it's also discriminatory to give key jobs only
- to the good engineers. Jobs are assigned by seniority and an unblotted
- copybook, rather than skill or ability.
-
- So you get folks who can have enough skill to stick N instructions together,
- but not enough experience or judgement to decide if it makes sense, or if
- it's stupid, to put a 4K hunk of code into the resident kernel.
- For example, DOS 4.00 had "optimizations" added to it whose net effect was
- to slow things down. The decision was made by someone who knew what
- an optimization was, but didn't have the experience to to evaluate such
- things - both on paper before and by stopwatch afterwards - to see if they
- make sense.
-
- Some IBM guys have been offended because they feel that I'm denegrating
- all IBM programmers. By no means. I worked with some sharp guys there.
- But I also worked with at least as many utter bozos. And more often then
- not, the sharp guys (who were considered untrustworthy "mavericks" because
- they said what was true vs. what management wanted to hear) were doing the
- scut work and the utter bozos were running things.
-
- gordon letwin
- not a spokesperson for anyone
-
-
-