Re: whenToUseGet, test licensing, and the RDF validator

On Wed, 2005-02-23 at 09:28 +0900, Martin Duerst wrote:
> Hello Dan,
> 
> At 02:55 05/02/23, Dan Connolly wrote:
>  >
>  >I'm composing an RFE
> 
> I have no clue what an RFE is.

Request For Enhancement

(it's a pretty common term, isn't it? yes, it's
in FOLDOC...
http://wombat.doc.ic.ac.uk/foldoc/foldoc.cgi?query=RFE
)

>  Google didn't bring up anything
> sufficiently suspicious for me to guess.
> 
>  >that involves running the RDF validator
>  >on my homepage... I can't bookmark the results because
>  >it seems to use POST. ew.
> 
> The RDF validator understands GET, so if this "RFE" is something
> automatic, you can easily change it to GET.
> 
> Btw, the main reason I made sure GET works is that it's
> much easier that way to test the RDF validator, see e.g.
> http://www.w3.org/2002/08/rdf-i18n-tests/.
> 
>  >Perhaps the reason is
>  > "- I grant the W3C permission to save this RDF and to use it for
>  >developing test cases"
>  >
>  >which can't be done with GET.
> 
> Well, strictly speaking, you are right, because giving W3C permission
> to save the RDF changes the state of the server.

Right; in particular, it's an action of the client that's beyond
"please send me some info".

>  On the other hand,
> this operation is actually idempotent,

"idempotent" is a red herring. It has very little to do with
safety, which is the main distinction between operations that
should use GET and those that should use POST.

(this was explained in early drafts of the TAG
finding on the issue
  http://www.w3.org/2001/tag/doc/get7.html
but seems to have been lost in later drafts. hmm.)


>  executing it twice doesn't
> provide more content or give more permissions (although the data
> will appear on the server in two different files with different
> numbers). Also, one may see the data collection as just a side
> effect, some sort of 'extended logging', so it's not a clearcut
> case at all.
> 
> I think the main reason for using POST was that for any kind of
> serious RDF, the GET URI gets extremely long, to the extent of
> being unmanageable.

I wasn't pasting the RDF into the form; I was using
the by-reference option. The URI needn't be much
longer than...

http://www.w3.org/RDF/Validator/ARPServlet?q=http://www.w3.org/People/Connolly/


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Wednesday, 23 February 2005 03:54:34 UTC