On Sat, 3 Feb 2007 13:28:20 +0100 Jeroen van Wolffelaar <[EMAIL PROTECTED]> wrote:
> And in CS it's common practice then to give an error or warning... I > don't see what your problem is with adding a warning to stderr in such > cases. Just silently truncating at the point where you cannot find a > good way to parse something anymore is a pretty lame resolution, not > unprecedented in the world of programming (think PHP at least), but > still. Thanks so much for the agreement JVW! Though I hope nobody fights too hard, as Thijs was being civil, by keeping the bug open though he disagreed with it. 'iprint' itself seems too trivial to battle over, any programmer could fix this bug! The true issue is more social than technical -- like a dress code: should a teacher be permitted to dress like a slob, say. Advocates for a natty dress code would say a slovenly teacher makes his school look bad. OTOH, if this slob were also a great teacher, far better than other more handsome ones, he might argue that his poor grooming might be a defect, yet his talents compensate for it. Which should not be taken as suggesting that any programmer or maintainer is too handsome or savage, rather the comparison is with the program itself. A great program with some minor ugliness is still a great program, but a trivial program with ugliness, that looks cheesey. Here we come to a question about Debian, which has many (16,000?) packages. Should all of these be required to be good, to live up to some agreed upon standards of utility and merit? Or perhaps if no consensus exists on what constitutes merit, should we just let 'em pile up, the good and the bad together, and have critics and librarians separate the gold from the dross after the fact? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
