[mdx] syntax for entity-attributes

[Tom Scavo]

On Wed, Sep 3, 2014 at 2:56 PM, Cantor, Scott <cantor.2 at osu.edu> wrote:
> On 9/3/14, 8:46 PM, "Tom Scavo" <trscavo at gmail.com> wrote:
>>
>>Hmm. What if a server chooses to support the "{sha1}" mechanism only,
>>leaving the client to do the mapping (say, with JavaScript)?
>
> What do you mean by "mapping"?

The short answer is: mapping from URI to SHA-1 hash.

The longer answer: The Federation Info pages

https://incommon.org/federation/info/

depend on a precursor of mdq-server that serves per-entity JSON
metadata. The files are stored on the server as follows:

0a7f1876b670153088592dac6ad2cf47.json
0a09e4a3283527098a04c6588641e4e8.json
etc.

The browser client converts an entityID URI to a SHA-1 hash (in
JavaScript) and passes the latter to the server. No mapping on the
server.

> Just in general, I think it's clear that with this spec, there can be no
> pushing of processing to the client in the context of anything the spec
> itself says the server is meant to do. The fact that you could implement
> such a split doesn't mean the server can rely on that in general.
>
> I may just not be understanding your point here.

I'm questioning the removal of {extname} from the base spec. I'm
wondering out loud if {sha1} should be put back into the base spec?

Tom