« WoT 2011 Deadline Extended | Main | From AI to BI »

Thursday, April 28, 2011


Feed You can follow this conversation by subscribing to the comment feed for this post.

Jan Algermissen

Hi Erik,

(1) do not media type negotiation is a runtime issue and should not be baked into representations because then you'll need to update the implementations of resources when more or less representations are available for other resources. Even worse, other's might point to your application/foo version which makes it hard to stop providing that media type without breaking those pointers.

(2) Hints on the subjects of feeds can already be expressed using Atom categories in service documents.

My 2 cents :-)



I think marking feed variants is a great idea!


@jan, thanks for your comment, but i cannot quite follow your argument. i agree that content type negotiation ideally should be a runtime issue, but how do you do it, then? it seems that the basic feed model of wrapping content in feed formats makes standard HTTP content type negotiation a bit useless, because you're negotiating the feed type, not the actual feed/entry type. so how can we expose the interesting information, the feed/entry content type, to runtime content type negotiation? i don't see your comment addressing the "how", only that it should be done. my original post was exactly about the "how", so what do you think, how should it be done?


This would be really useful. I think it could greatly save on the number of requests clients need to make to get data in the desired representations.

Although, I should say that it would be really nice if good Web citizenship was more widespread so that more people published data in multiple representations in the first place.

The comments to this entry are closed.