[mdx] Joe on section 2.2

[Ian Young]

Joe's comment:

> -- 2.2 specifies that "all metadata retrieval requests MUST use
>   the GET method" -- would this preclude the normal use of HEAD
>   as an alternative to test for changed data, etc.?

This one had quite a lot of back and forth from Leif and Scott.  Leif thought that it would be nice to support HEAD as an alternative to conditional GET. Scott was initially quite negative because he didn't see that having more than one way to do it added value. In the end, though, as either way the spec wasn't going to *require* his client to implement HEAD, he thought the server-side people might be the people who had to care about this.

My remarks:

I don't see a strong use case for permitting HEAD as well as GET, unless there are HTTP client libraries that are unable to implement conditional GET properly. Outside such a case, I don't see a use for HEAD except as a way of invalidating a cache without fetching the new data in the way that a conditional GET would do. I don't think that's very likely to be an issue for clients; they are more likely to ask for metadata when they think they need it and that means that a HEAD would always be followed up by an immediate GET, so using conditional GET is a more efficient solution.

I propose, then, that we leave things as they are (only GET supported) unless we find evidence that this would make an MDX client impossible to build in some library. Action item on Leif to look at the Python libraries, perhaps?

Comments please.

	-- Ian



-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4813 bytes
Desc: not available
URL: <http://lists.iay.org.uk/pipermail/mdx-iay.org.uk/attachments/20131001/f0ef31d4/smime-0001.p7s>