Lists Home |
Date Index |
Tim Bray wrote:
> Dare Obasanjo wrote:
>> The "Desperate Perl Hacker" argument was a bogus claim for XML 1.0
>> because of the existence of entities and CDATA sections but is quite
>> farcical now with the existence of the Namespaces in XML
>> recommendation (and it's bastard spawn "QNames in content").
> Empirically false, at two levels. First, lots of people process XML
> with perl (or equivalent) all the time. Second, the real requirement
> was to make it tractable to take a large body of document data and make
> quick programmatic changes on it. Which, obviously, XML makes way easier.
The requirement as we understood it years ago was that the person would
be working with regular expressions and _no XML parser_. (otherwise
syntactic choices like minimization would have been irrelevant)
Dare is right that XML-as-specified was never simple enough that you
could reliably process XML with Perl 5 regular expressions. On the other
hand, the XML one tends to receive, tends to be simple enough that it
works most of the time. You don't want to build a system around it but
for one-offs on data sitting on a hard drive, it usually works. CDATA
sections and entities are easy to "grep' for.