On Feb 3, 2008 8:06 AM, Daniel Aleksandersen <
[EMAIL PROTECTED]> wrote:

>
> …I should probably use <archive
> xmlns="http://purl.org/syndication/history/1.0"/> as well? What is the
> point of xmlns:prefixes="URI" attributes? Really, why make a standard if
> no one follows it?
>

There are basically three types of tools: XML+XMLNS-aware ones, and ones
cobbled together by looking at examples and writing software that works with
the examples are the big two.  Given the reach of Atom and its dialects,
there will be a *lot* of implementations out there that do things the Wrong
Way.  You can either simplify the standard and limit your implementations,
or give the standard some flexibility and maximize your adoption rate.

But I said three types of tools, didn't I?  The third will be feed
aggregators that want to maximize the number of feeds they can handle.
Given the number of bad feeds there are, pumping these through a typical XML
parser will reject many feeds that they still want to process.  These
aggregators, consequently, may benefit from doing their own parsing,
treating Atom like they treat HTML: tag soup.  These applications may fail
to process some strictly standards-compliant feeds, because they do uncommon
things.  You might argue that this is just them doing things the Wrong Way,
but it's an explicit, conscious decision, not due to their own incompetence.

David

Reply via email to