RFC initiated for Simple Cone Search protocol

[Thomas Boch]

Dear Doug, dear DALers,

Here are a few comments about the Simple Cone Search document :

- section 2, part 1 : I find there is an ambiguity about the VERB
parameter. The document states it is an "optional parameter" without
telling if it applies to the client or to the server side. I think it
should be made clear that this parameter is optional in the query, and
that it is not compulsory for compliant services to support it. 

- section 2, part 1 : typo in sentence "If a query includes an optional
parameter ... the service should must ignore that parameter."

- In the References section :
    - [VOTable 1.0] link points towards version 1.1 of the document, not
version 1.0
    - [RM] link is incorrect (final 'l' is missing in the URL)

- Although I perfectly understand the need to be backward-compatible
with existing Cone Search implementations, there are 2 minor points I
would find annoying if I was about to develop a new CS service, and
willing to conform to the standard :
    - the document compels to use the old UCD1, at least for the 3
compulsory fields (unless I repeat these 3 fields with corresponding
UCD1+, which is rather ugly)
    - I feel uneasy about the ambiguity on the location of the ERROR
parameter. If we plan to choose one location over another in a future
document, it will mean changes at the service level anyway. Otherwise,
the document should state explicitely (not only in the editor's
note) that both locations are allowed.


Best regards,

Thomas