Tim Bray wrote:
I got an email last night from a well known syndication implementor
pointing out an obvious bug in my Atom feed. The feed's valid, but the
stuff in <content> was full of relative URIs which were broken because
I'd borked the xml:base. So I went through the code and got the
xml:base right and ruthlessly pruned all the pointers. The feed looks
a little weird and it's giving the (pre-alpha) validator a heartburn,
but I sort of think it's right. I also think it's good practice.
Check it out:
<feed xmlns='http://www.w3.org/2005/Atom'
xml:base='http://www.tbray.org/ongoing/'
xml:lang='en-us'>
<title>ongoing</title>
<link href='' />
<link rel='self' href='ongoing.atom' />
<logo>rsslogo.jpg</logo>
<icon>/favicon.ico</icon>
I agree that relative URIs are a good practice.
I didn't realize that "path-empty" was a valid URI-reference.
Looks like another test case to write. ;-)
- - -
While it clearly shouldn't be the default behavior, longer term (i.e.,
sometime well after basic Atom 1.0 support is more complete), how much
value do you think that there would be value in an option to attempt to
verify all potentially dereferencable URIs resolve to accessible resources?
- Sam Ruby
