that article you link doesn't make any sense at all to me

could you be a bit more specific with your criticism?

could you be a bit more specific with your criticism?

Sure. If you have something like a "<" character in XML text, you have 3 possibilities:

1. It is part of a XML element
2. You encode it as &lt;
3. You embed it in an unparsed CDATA section

The first solution is not a solution at all unless we want to start messing around with XML namespaces like the article suggests. How many RSS reader will convert namespaced XML elements to HTML tags? Not many, I think. On the other hand, &lt;, &gt;, &amp;, &quot; and &apos; are predefined XML entities that are always valid and should always be decoded to the characters they represent. I'm totally perplexed by the claim that most XML APIs won't resolve those entitites, because both DOM and SAX sure do (I think I should know). Also, I find it curious that the author claims that most RSS readers don't use a proper XML parser - well, if that's the case than it surely is not our job to conceal the holes in that software's pseudo-parsers.

Again: First we have to decide if and where to allow HTML tags (for example I'd say stripping HTML tags from titles but leaving them in the descriptions would be a sane decision). If we do have HTML tag characters, the way to encode them is definitely as predefined character entities in PCDATA.

thanks for the clarifying words, hannes. i'm +1 for encoding html and +1 for completely removing html from the title.

i also would like to see the descriptions being without any markup – at least in rss readers. but maybe then i have to make the right choice of software (ie. a reader that strips the tags before displaying the feed). what do you think?

to complete this rss task, i also would like to discuss how long the descriptions should be. i am really against putting the whole story in there, instead i'd prefer a short teaser (maybe a little bit longer than in the history e.g. on this site). however, we could leave it up to the user's choice.

last but not least: in the current version of antville the new content model can lead to rss feeds that are completely useless, ie. when the site's stories aren't using the title and/or text properties anymore (which the rss skin relies on).

so the question is: fully enable the rss skin in the skin manager or create a meta editor to assign a story content part to the title and description fields of the rss output (a suggestion by robert, btw)...?