Path: utzoo!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!munnari.oz.au!goanna!ok From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) Newsgroups: comp.lang.prolog Subject: Re: Arity's so-called upgrade Message-ID: <4989@goanna.cs.rmit.oz.au> Date: 18 Mar 91 08:15:23 GMT References: <11768@uhccux.uhcc.Hawaii.Edu> Organization: Comp Sci, RMIT, Melbourne, Australia Lines: 73 In article , dowding@ai.sri.com (John Dowding) writes: > Because no vendors will respond favorable[sic] to the request to declare > publicly to a potential (or actual) customer all the known bugs and > shortcomings of their product. I wonder about that. When I was a student at Auckland University, we had a Burroughs B6700. A couple of times a year Burroughs provided a FTR/TAR collection (FTR = Field Trouble Report, TAR = Trouble Analysis Report). These were fairly thick lists of all known problems, both the ones that had been fixed and the ones that had not been fixed. From time to time I would try some of the "not yet fixed" ones to see if we had those problems, and we usually did. I don't recall them being subject to any form of non-disclosure, and there was nothing to stop any potential Burroughs customer wandering in and looking at them. DEC used to do the same sort of thing for RT-11 and VMS, and I presume they still do. One AT&T documentation set for UNIX V.3 that I saw had a list of known remaining problems. I have no idea what Quintus's current policy is. I don't recall them publishing a list of bugs when I used to work for them, but if someone asked "is this your bug or mine" they got an answer by return E-mail, and nobody ever said "don't tell". Vendors are in it for the money. Some vendors are big enough to ignore individual customers, but most Prolog companies that I know about are not. If you go to Quintus, ALS, Integrated Solutions, LPA, &c, and say "I am interested in buying X number of copies of your product, but only if you give me a list of problems currently known not to be fixed, and yes, I am willing to sign a non-disclosure agreement", then they are going to think very hard about doing it, because your purchase is big enough to be noticed. > Over the years I have submitted a number of bug reports to a variety > of vendors. In each case, I have never known if they knew about this > bug already, or if mine was the first report. Who knows how many > times I have spent hours tracking down bugs in products that the > vendor already knew was there? (That is a rhetorical question, Richard) How many times did you *ask* whether they already knew about the mistake? From my own experience, SUN *did* tell you whether your problem was a known one (after you submitted it), provided the person submitting the report was the official contact person for the site. By the way, my understanding of a "support contract" is that information like this "feature such and such has mistake so and so, we intend to fix it soon/in the next release/eventually, in the mean time here is how to work around it" is one of the things that such a contract is FOR. Again, I would point out that most Prolog companies are small and hungry. If you demand this kind of thing as part of the "support" you are paying for, they are hungry enough for the support money to listen and maybe do it. (It takes a fair bit of staff and time to track bug reports and distribute this kind of information, so I think it is perfectly fair of vendors to ask for money to do it.) > In any case, it is certainly not the vendors[sci]place to tell us what > should and should not be discussed on the net. I'm not a vendor. And the point of my posting was not to say "thou shalt not" but to say "DO YOURSELF A FAVOUR, go FIRST to the people who really DO have the answers, as opposed to the people who only think they have the answers." The law varies from country to country. My understanding is that in NZ and OZ vendors can get away with pretty well anything in their advertising, but that if you write and ask "does your product do X" and they state in writing _in reply to a specific request_ that their product does X, then they are legally required to be telling the truth. Do you begin to see how something like this could be used to shape the situation so that it is in the vendor's interest to tell the truth about known problems? -- Seen from an MVS perspective, UNIX and MS-DOS are hard to tell apart.