Flutterby™! : time doesn't wait

time doesn't wait

2009-06-11 18:20:38.396085+02 by Dan Lyke 3 comments

Making <time> safe for historians, on the new HTML <time> attribute and the difficulties of specifying dates.

[ related topics: Web development Archival ]

comments in descending chronological order (reverse):

#Comment Re: made: 2009-06-12 16:08:16.697888+02 by: Dan Lyke

Eric, I think the point is that HTML wasn't initially a page layout language, it was a semantic markup language, and the people who care about HTML would like to see it go back to those roots. People have glommed on to "no tables" like it was a religion, but the core of that movement was "tables for tabular data, not layout".

Sean, I sympathize. One of the problem with human readable stuff is that people tend to implement it as though that's all that has to be done with it.

#Comment Re: made: 2009-06-12 10:20:30.369274+02 by: spc476

Gah, dates! I've been working on a program to index my email and dates have been one of the more problematic areas, oddly enough. 400 lines of code just to parse the Date: header (and I think it's mostly correct, although the thought of personally checking 72,385 messages is a bit daunting) and that's for a supposedly well defined header (and in an ideal world, parsing the date header would only be a few lines long).

Nice idea in theory …

#Comment Re: made: 2009-06-12 07:39:06.67005+02 by: ebwolf

Sounds like the basic "why lat/long in WGS84 is really very good for location" problem - but in history rather than geography.

But I don't really get the argument. HTML5 is a page markup language, based on XML. If there is a <time> tag in HTML5, then it's for layout, not semantics. The real problem with the web has been people confusion markup with semantics (and that dates back to Sir Berners-Lee). A separate History-ML should establish semantics for time - just as GML/KML establishes semantic location.