home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.advocacy
- Path: sparky!uunet!decwrl!netcomsv!mork!xtifr
- From: xtifr@netcom.com (Chris Waters)
- Subject: Re: Portable?
- Message-ID: <k+kmwwl.xtifr@netcom.com>
- Date: Sun, 26 Jul 92 01:36:41 GMT
- Organization: Netcom - Online Communication Services (408 241-9760 guest)
- References: <1992Jul25.224915.15078@news.Hawaii.Edu>
- Lines: 60
-
- In <1992Jul25.224915.15078@news.Hawaii.Edu> tholen@galileo.ifa.hawaii.edu (Dave Tholen) writes:
-
- >Phil Lafornara writes:
-
- >> This whole thread was started in response to the periodic postings
- >> of "when Portable OS/2 comes out." I questioned that, and asserted
- >> that IBM would be introducing new bugs into the code when they did
- >> the translation. End of story.
-
- >And I asserted that the probability of introducing bugs was increased, but
- >not 100 percent. Once again, you're using "would be" as though it was
- >absolutely guaranteed that portable OS/2 will be buggier than "something
- >not translated from assembly" (insert Windows NT here if you feel that it
- >has been implied).
-
- Um, I think I'll jump in at this point. First of all, my credentials:
- over 12 years professional experience with a variety of languages. I've
- specialized in maintenance and porting--it's not a glamorous field, but
- it pays well. And I've rooted around in some of the ugliest code you'll
- ever hope not to see. I've translated code from C to assembler, and
- vice versa, as well as to/from Pascal, Forth, Basic, Fortran, and Algol.
-
- I have no axes to grind--I'm using OS/2 because it's available, but I've
- always been a fan of Microsoft, and I'd probably feel more comfortable
- with an MS produced OS. But XENIX is not what I want on my home
- machine, and NT isn't out yet. And OS/2 is working--better than I
- expected, actually.
-
- Anyway, I have to disagree with the statement that "translating code
- from assembler to C introduces bugs." Believe me, I've done a *lot* of
- it, and this is far from *always* true. Often true, perhaps. But I
- *don't* blindly try to do a one-to-one mapping between the C and
- assembly code. I find out what the assembly code is doing, and
- *design*, then write C code to do the same thing. But *not* necessarily
- the same way. With this approach, I find that I cure *almost* as many
- bugs as I introduce, on the average.
-
- This is essentially what MS has been doing with NT--working with the
- design specs of DOS and Windows, and writing C code to do the same
- thing. I'd bet money on this. And, I strongly suspect that IBM is
- doing the same thing in their efforts to create a portable OS/2. They
- don't *need* to read the assembly code to port it to C--they just have
- to know what it's doing (in detail), and write C code to do the same
- thing.
-
- Now, I admit that I don't work for IBM or MS, so I don't know what
- either the OS/2 or NT code is really like. But from where I sit, I
- cannot see one good reason to think that more bugs will be introduced in
- the creation of portable OS/2 than in the creation of portable
- DOS/Windows (NT). If NT has to read FAT filesystems, if it has to
- handle interrupts generated by DOS programs, then it has code, DOS code,
- ported from assembly. Or so it seems to me.
-
- If Phil has a rational explanation of why MS's porting efforts are less
- likely to introduce bugs than IBM's, I admit I'd be interested to hear
- it. But all I've heard so far are comments that should apply equally to
- OS/2 or NT.
- --
- Chris Waters | the insane don't | NOBODY for President!
- xtifr@netcom.COM| need disclaimers | Because Nobody's perfect!!
-