One thing that needs to be kept in mind... and that feed reader
developers need to make sure they're handling properly... is that Atom
intentionally provides two locations for human readable content:
atom:content and atom:summary. If the atom:content element does not
contain readily displayable content (like HTML), then the atom:summary
element should be used to provide some kind of displayable approximation.
- James
Martin Atkins wrote:
Richard Salz wrote:
Happy New Year.
I want to add content to an atom:entry, but it's intended for use by
computers not humans. I know atom:content can hold arbitrary XML, but
I'm not sure if I should use that or if I should put it as a direct
child of the atom:entry element.
I don't like usurping the "human-readable" intent of atom:content.
Also, since it's Language-Sensitive, I have a slight concern about
someone translating machine field names and the like. There's a
(slight) chance that the content could from different components,
which is why I prefer inserting elements into atom:entry instead of
"rewriting" atom:content. I don't have any justficiation for that
other than "gut feel" however.
I want to be able to associate a schema with the data (see
http://www.ibm.com/developerworks/blogs/page/soma), and using
atom:entry seems cleaner. But it's a small matter to say the schema
link relation is scoped to the atom:entry atomInlineOtherContent type.
In practice, you don't want to put anything other than HTML in
atom:content if you intend this feed to be consumed by general Atom
processors and aggregators. Many aggregators (and feed re-publishers)
will misbehave with anything other than <content type="html"> and
escaped content.
You can add arbitrary XML elements as children of atom:entry, and use
xml:lang to define what human language is within.
Of course, if you intend this feed to only ever be consumed by your
own software it doesn't really matter what you do as long as your
producer and your consumer agree.
If you are able to give more details about specifically what you are
trying to do folks might be able to suggest a way to do it that does
not involve extending Atom at all, at least at the XML level.
