James M Snell wrote:
What is the type of the resource pointed to by the in-reply-to href? It

It's whatever type the server says it is when you GET it (Content-Type header).

Well ideally you wouldn't want to GET or HEAD every href in order to determine its type. I'm not positive this is a problem, but it may be depending on what a client may want to do with this extension.

The related link allows clients that are not familiar with the Feed Thread extension to at least be able to indicate that there is some kind of relationship between the two resources.

Imagine that a client doesn't know anything about this extension. What do you think such a client would typically do with related links? I don't know about anyone else, but what I do is display them in a "see also" list that the user can launch in a browser window. This works very well when the links are to web pages or other resources that can be handled usefully by a browser, but it's a complete waste of time if the link is pointing to an Atom feed. Most of the time the user will just get a page of meaningless XML - at best they'll get the option to subscribe to a feed that they're quite likely already subscribed to.

But that's just how I handle links. Maybe there's a better way. Can you imagine some other way in which clients might handle related links (without knowing anything about this extension) that would be able to do something useful with a link that pointed to an Atom feed - specifically the Atom feed that the current message is replying to?

Now if the in-reply-to href was pointing to a web page I could understand recommending a related link that pointed there too. So maybe you could limit the recommendation like that. Or otherwise just leave out the recommendation altogether and let the feed producer decide whether a related link is a good idea or not. Related links are already part of the Atom spec - there's nothing stopping a feed producer using them - but I'm not sure they should be encouraged to do so with links which may not be at all appropriate.

Regards
James

Reply via email to