[
Lists Home |
Date Index |
Thread Index
]
Rich Salz wrote:
> He goes a bit overboard, but Mark Pilgrim makes some good points:
> http://diveintomark.org/archives/2004/01/14/thought_experiment
> (What, Mark going overboard? I'm shocked I say, shocked.) Don't lose
> his point:
> it's not necessarily invalid markup, but different character
> encodings, too.
The problem is an architectural one: if you have feeds coming in from
sources
out of your control, you need to have a fallback mechanism for when
feeds come in incorrectly, whether incorrect beause of bogus encodings
or crap
markup.
If your application has not been coded for errors, of course it will be
fragile.
People who think documents should allow pieces of different encodings are
thinking of the English case where characters in different encodings are
probably
decorations or prissiness: but if I have half of a generated HTML page
in Japanese
Shift JIS and half in Japanese EUC-J somehow, which one should be rendered
correctly? Either way the page is crap, and the users will complain.
Underneath Mark's issue is the same old one: that there are serious systems
and casual systems, and the disciplines required for one is burdensome for
the other.
Cheers
Rick Jelliffe
|
