Stewart Stremler wrote: > begin quoting Bob La Quey as of Fri, Jul 06, 2007 at 12:06:32PM -0700: >> Standard off the shelf validating parser please. > > Well, there's java.util.Properties, aside from the block-of-text syntax. > But there are standard ways for that, too. See below. > >> I do not want my programmers wasting their time >> writing parsers. Even simple ones are hard, if >> one expects them to be reliable and handle flaky >> inputs. > > I've never found XML to handle flaky inputs at all. Mostly, it falls > down and dies, giving cryptic and/or non-useful errors. > > I've also had XML that Will Not Parse on a machine that isn't connected > to a network. With validation turned off. Whatever good a "standard > validating parser" gives is insignificant to that one little problem. > >> I also want off the shell tools like XSLT to >> produce html for reports. > > What sort of report do you get from a printer configuration? > > (And presumably, the report is only needed because XML is so > gosh-darned annoying to read...) > >> BTW, what do you call that? It is not an s-expression >> is it? > > I call it "structured keys". > > [snip] > > Third format (java.util.Properties style): > > consumable.comment.en=Toner Black (4000 prints) - $80 \ > Toner Yellow (4000 prints) - $110 \ > Toner Magenta (4000 prints) - $110 \ > Toner Cyan (4000 prints) - $110 \ > Transfer Belt (35000 prints) - $47 > > Most scripting languages just execute configuration files, so they've > basically got their own validating parser built-in. >
Are you not mixing syntax validation (well-formedness) with the more specific (ie, DTD) document validation (that blq _may_ have meant)? Not that DTD validation is always needed/desired, but when it is, the choice of xml would seem to be a no-brainer. Regards, ..jim -- KPLUG-List@kernel-panic.org http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
