Re: Re: Accepting invalid input [was RSS 1.1]

On Thu, 20 Jan 2005 20:38:35 -0000, ecomputerd <ecomputerd-/[email protected]> wrote:
> 
> 
> It is very unlikely that a popular aggregator would reject invalid
> RSS altogether. If any aggregator is able to read the feed, one of
> the first things the developer will hear is "I can read the feed
> with YYY reader, why not yours?" This has been my experience.

Yep, this was (IMHO) the strongest argument against error-only
reporting of ill-formed XML in Atom.

I still think had there being a will within the developer community to
insist on rejecting bad XML as a general rule, then the feedback
effect would have meant that the proportion of non-XML feeds would
have declined sufficiently for "I can read YYY" having been so
marginal as to avoid proliferation of liberal parsers.

This would only have been viable with Atom because it was effectively
starting from day zero. Even then many developers will be extending
existing (liberal) parsers rather than building anew. The approach
wouldn't stand a chance with "simple" RSS, or probably for that matter
with RSS 1.01. Whether it's appropriate for 1.1 I don't know, but
given the attitudes of people in the Atom WG (a more spec-respecting
bunch than average) I suspect the display-but-shout-loud approach to
ill-formed XML is probably most appropriate for RSS 1.1.

One thing that is sure, having a well-known online validator and tests
early on encouraged spec conformity among (0.3 beta) early-adopters. I
would imagine this will be doubly important with RSS 1.1, where the
validity of RDF/XML is also an issue.

Just a thought, if the normative spec is say OWL Lite, then what
happens with modules which e.g. use >1 cardinality?

Cheers,
Danny.
-- 

http://dannyayers.com


 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/rss-dev/

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/