intertwingly

Yup, you point out the problems with "View Source" and relying on tools instead of reading specs. I modelled my feed off of yours http://www.intertwingly.net/blog/index.rss which at the time of this posting claims to be RSS 0.91 and has no image tag.

Also when I validated it against the validator at http://feeds.archive.org/validator/ it never complained about that. It does complain about a relative URI reference which I haven't changed because I can't edit my K5 diary and I'd rather not tweak the code that generates the feed for that one issue since it isn't a fatal error.

PS: I do think <image> being a required element is poorly thought out and will probably switch the version number on my feed instead of providing a bogus <image> element simply to be spec compliant.

I think someone else has said this before me, but the greatest thing about XML is that it's human readable. That's also the worst thing about it. Particularly when the instance documents are often more readable than the specs they're supposed to comply with.

My feed claims to be version="3.14159265359", so I don't have these problems.

Also, Dare, your RSS 1.0 feed contains HTML in the description tag, which is not allowed in RSS 1.0. (Put it in content:encoded instead. See http://www.aaronsw.com/weblog/index.xml for an example.)

Also, Dare, your K5 diary declares itself to be HTML 4.0 Transitional, but the W3C validator shows 57 validation errors:

http://validator.w3.org/check?uri=http%3A%2F%2Fwww.kuro5hin.org%2Fuser%2FCarnage4Life%2Fdiary

I'm writing a browser that can only display valid HTML. Could you please fix your site so I can read it? Thanks.

I've seen screen shots that lead me to believe that RSS Bandit will happily consume my "invalid" RSS 0.91 feed.

Clearly, these issues are not cut and dry.

*chuckle*

Mark,
I don't see anywhere in the RSS 1.0 spec at http://web.resource.org/rss/1.0/spec#s5.3.3 where it describes the <description> tag that mentions anything about HTML content. All I see is a suggested maximum length. Please provide links to normative documentation that bans textual content that happens to be HTML from being the content of the <description> element.

Sam,
According to the DTD your RSS feed claims to comply with at http://my.netscape.com/publish/formats/rss-0.91.dtd your feed is valid and the <image> element is optional. Now we seem to have to contradictory statements about the mandatory nature of the <image> element. Fun, fun, fun.

IMHO, if all you people have to say about the points I raised is to attempt to find flaws in my RSS feeds then obviously there isn't much to discuss. Y'all should read http://www.datanation.com/fallacies/subject.htm sometime. :)

http://diveintomark.org/archives/2002/11/22/debate.html

I think it's pretty clear from http://web.resource.org/rss/1.0/spec#s3 that HTML is only allowed in places where explicitly stated. Ask anyone on the RSS-DEV working group.

Everything I have to say about your points I've already said on XML.com. The rest of this is just blowing off steam.

It was fun arguing with you. Let's do it again real soon.

Mark,
I probably agree with your points more than I let on and actually toyed with using an co-worker's SgmlReader instead of our XmlReader to process RSS feeds when working on RSS Bandit.

Joshua and I talked about your article some and he also felt torn and could see the merit of both arguments.

Either way, it's been fun debating. :)

PS: I can't wait for your next column, the first two have been enlightening and I plan to quote your "What is RSS?" article in my MSDN column next month. Good stuff.

Ah... Dare noticed the DOCTYPE in my RSS 0.91 feed. It turns out that my 0.91 feed is valid when compared to this 0.91 spec, but not this 0.91 spec. Dare said he modelled his RSS feed off of mine, but since he didn't copy the DOCTYPE, his actually ...

Pingback from Sam Ruby: What should the RSS validator validate?