home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!wupost!cs.utexas.edu!sun-barr!ames!news.hawaii.edu!galileo!tholen
- From: tholen@galileo.ifa.hawaii.edu (Dave Tholen)
- Newsgroups: comp.os.os2.advocacy
- Subject: Re: Portable?
- Message-ID: <1992Jul26.151507.14399@news.Hawaii.Edu>
- Date: 26 Jul 92 15:15:07 GMT
- Sender: root@news.Hawaii.Edu (News Service)
- Organization: University of Hawaii
- Lines: 67
- Nntp-Posting-Host: galileo.ifa.hawaii.edu
-
- Pete Skelly writes:
-
- > Could it be that you inferred that the discussion had anything to do with
- > NT only due to the fact that some of us involved are MS employees. If that
- > is the case, then you're inference is incorrect.
-
- No, I inferred it because what other operating system is being written from
- scratch in C with portability to non-Intel platforms in mind?
-
- > Let me restate what I once said.
- > SO WHAT. Would you put your bank account on the fact that there is a small
- > chance that IBM may not introduce any bugs during translation. What is your
- > cut off for acceptable bug introduction? For an acceptable risk. As I see
- > it, there is a very high probability that any program translated from
- > assembly to C will have some bugs introduced.
-
- It makes no sense to bet the bank account unless the probability of
- introducing bugs is less than 50 percent, and I never said it might be
- that low. All I've said is that it is not 100 percent, whereas the
- others have been using terms like "will" and "are", which imply 100 percent.
-
- >>> Actually, the probability statistics were posted by a non-MS
- >>> employee. Have you been reading this thread at all?
-
- >> Yes I have been reading it, which is why I know that Phil (you), Pete,
- >> and Raymond (all Microsoft employees) have all insisted that translation
- > ^^^^^^^^^^^^^^^^^^^^^^^^ SO WHAT.
-
- The so what is because in Phil's earlier posting (see above), he said that
- the probability statistics were posted by a non-MS employee. I have no idea
- who he was referring to. I was referring to the postings from the above-
- mentioned people who said that bugs ARE (read 100 percent probability)
- introduced. It's those probabilities that I take exception with.
-
- > I suspect most other SE's will agree with the conclusion, that Bugs have a
- > high probability of being introduced during translation.
-
- An increased probability, yes. How high it is depends on the talents of the
- programmer.
-
- >> Actually, in certain circumstances, translation could be easier than writing
- >> from scratch, because the algorithm already exists. Coding a new concept,
- >> depending on the complexity of the concept, could be much more time
- >> consuming. I know from my own experience that when I need to do a
- >> particular task, it is sometimes faster to use published examples of code and
- >> translate into my language of choice rather than attempt to fully understand
- >> somebody's word description of an algorithm and then code it from scratch
- >> myself.
-
- > Let me be clear on this. Are you implying that OS/2 is one of those
- > "certain circumstances"? If not, SO. If so, well, the project is big.
-
- Parts of OS/2 probably do fall into this category. Just as an example,
- tell programmer A to write a new file system from scratch in C with the
- following specifications (insert those for HPFS). Tell programmer B to
- translate the assembly code for HPFS (assuming it was written in assembly
- to begin with and has already been debugged) to C. Which of the two will
- likely have fewer bugs? Depends on the talents of the programmer, but I
- wouldn't be surprised if programmer B introduced fewer bugs. Other parts
- probably also fall into this category, but I don't know just how modularized
- OS/2 is, though it's hard to imagine an operating system as complex as this
- being monolithic rather than modularized.
-
- > Taking up someone elses code can be a pain in the *** sometimes, especially
- > if their coding style is different from your own.
-
- Agreed.
-