ISO/IEC JTC1/SC22/WG5/N1565 From: Richard Maine RESPONSE TO N1558 (NAGGING DOUBTS) 29 July 2003 I won't try to address the meat of this paper. I haven't studied it well enough yet to do a good job of that. I wish to note only that I do *NOT* consider these to be "minor" technical changes (much less "very minor"). I see these as significant technical questions that we have absolutely no business in changing in a draft that becomes a FCD. There is new syntax here. There are changes in definitions of things that could affect stuff all over the standard (type compatibility). If we try anything so foolish, we will get it wrong. Kurt argues that we have gotten several of these things wrong. Perhaps he is correct, but I believe that in making a last-minute change we would break more things than we would fix. The term "last-minute" understates things. It is far past the last minute. I don't think this is something that should have been distributed "some weeks ago"; I'd say more like a year ago. Although Kurt describes these as decisions made at the recent joint meeting, most of them look like things that have been around a long time to me. In fact, I'm having trouble identifying any related decisions made at the last meeting; perhaps I missed something, but this material does not look like fixes of decisions made at the last meeting to me. We did make lots of decisions at the last meeting, and I think that includes mistakes that need fixing, but I'm having trouble making the connection with the questions of this paper. The reallocating assignment is part of the TR, for example. The only recent changes relating to it have to do with applying it to extra cases; Kurt's concerns seem addressed more at the basic functionality as described in the TR than to those extra cases. The function side effect question is far from new and although C interop is new, it isn't new to the joint meeting. This one can get very subtle; I'd bet that if we try to say anything of subtsance on it now, it will be nonsensical in at least some cases and wil result in interp questions. Perhaps these are all thing that we really need to fix before going out with an FCD. I defer to WG5 on that question. However, I strongly feel that if we make changes of this order, we should slip the schedule by at least 6 months. I feel it irresponsible to make such technical changes directly into the FCD. If we put out an FCD that has technical flaws because of after-the-last-minute changes, then it will slip our schedule by a lot more than 6 months (my understanding is that it kicks us back to the CD stage). Of course, I have to be suspicious that if we slip the schedule by 6 months, we will just get a bunch more technical changes 6 months from now. Are we ready for FCD or not? If we need to make these kind of changes, then we aren't ready for it. To me, that is the most fundamental question here. -- Richard Maine | Good judgment comes from experience; Richard.Maine@nasa.gov | experience comes from bad judgment. | -- Mark Twain