Path: utzoo!attcan!uunet!husc6!cmcl2!rna!kc From: kc@rna.UUCP (Kaare Christian) Newsgroups: comp.sys.ibm.pc Subject: Re: Microsoft bugfix non-support (& open letter to uSoft) Summary: Borland Better?? Message-ID: <167@rna.UUCP> Date: 26 May 88 15:05:42 GMT References: <5537@megaron.arizona.edu> <32505@amdahl.uts.amdahl.com> <232@octopus.UUCP> Organization: Rockefeller University - Neurobiology Lines: 56 I'm not an independent observer of Microsoft, so I don't want to comment on their support. (But I'm not in bed with Microsoft, and I believe that I can report fairly on Borland. I am not a Borland basher, I have occasionally used turbo pascal for short projects, and I have taught Pascal using Turbo.) In article <232@octopus.UUCP>, pete@octopus.UUCP (Pete Holzmann) writes: > Borland: First release of the compiler. I note one of the 4 bugs that everybody > found. Call Borland. They immediately ship me, FREE, an update with > all 4 bugs fixed. They also tell me the patches over the phone so I > can get going immediately. I find another, more subtle bug; they send > yet another update for free. As version 1.5 is released, I find another > 1.0 bug. They send me version 1.5 for free! [1.5 has significant new > features, not just bug fixes] I had just the opposite results with TurboC, and I abandoned it because of its many bugs. While working for half a year in Germany, I bought a copy of TC version 1. During my first few weeks I ran into a serious bug every few days. This resulted in four formal bug report letters to Borland, each accompanied by example code that irrefutably demonstrated the bugs. Some of the reports detailed several bugs, so we are talking about 8 or so of the buggers. There were problems with far and huge pointers, problems with floating point, problems with printf, and a couple of problems with library routines. In response to each letter I got a very polite response admitting that there was a problem, and pledging to fix the bug soon. (Each response was also accompanied by a compuserve freeby, which is pretty useless in Germany.) I never received any fixes. One of the bugs was in the lfind procedure. I wrote a replacement in about ten minutes (tho it took about a day to find the bug originally, 'cause who would suspect that lfind couldn't look through an empty list). Their response to this trivial to fix bug (just send me a new lfind) was "our engineers are working on it and we expect ...." Cripes sakes. If more than one person is working on it for more than about ten minutes then you should go get some real programmers. Sometime later I was at a trade show in Munich, and I talked to a person at the Heimsoeth booth (they sell Borland stuff in Germany). The person said that 1.5 was just out and it was much better, and that he would send me a copy if I returned my 1.0 disks. I did, and he did, but there were no docs, no license, no nothing but the disks with the 1.5 software. I don't know if he was authorized to give away 1.5, or if he did it illegally. None of the important bugs was fixed. Approximately two weeks of my time were wasted because of turboc bugs. (I worked with turboc for about six weeks, but during that time I only made about 4 weeks of progress, because the rest of the time was spent finding compiler bugs.) In the end, my source was strewn with #ifdef __TURBOC__ statements to work around the bugs. It was a waste of time; my bug reports obviously didn't help make a better 1.5. Kaare Christian Research Assoc. The Rockefeller Univ.