page 6 - syntax for region searches.
is the ' ' quoting the kind of region necessary? it doesn't seem to
fit with the rest of the adql syntax.
going further, is the whole REGION ( ' .... ' ) syntax
redundant? if not, why?
---
page 7 - Top n construct.. a few typos in this para.
INTO and #Upload - would it be possible to widen this spec from a 'vospace location' to 'a URI'. This would allow queries that fetch and write to other storage systems (ftp, http, SRB, something else not yet invented) to be expressed.
I also think the #UPLOAD keyword needs describing more carefully - I'm not clear how it works, or how it's to be implemented, at the moment. Maybe a concrete example of addressing a particular table within a votable would help here.
--
"Table name qualified by a service identifier MAY be supported to
specify a table
that belongs to another SkyNode service."
is this skynode-specificity appropriate in the adql spec?
A short name of the service MAY be
specified, however note that it does not guaranty the uniqueness in
the VO.
[( <service_identifier> | <short_name> ) : ]
<table_name>
e.g. ivo://jvo/sxds:tableName
e.g. sxds:tableName
the full form - which will probably be used in non-skynode
implementations is a little bit misleading. In particular, it looks
like a URI string, but is an invalid one - as a ':" can't occur after
the initial 'scheme:' fragment. This may cause client libraries that
try to handle such a reference to fail with a parser error.
if '?' or '#' were used instead of ':', then this would be valid uri
syntax - and would then be a convenient way of unambiguously
addressing a single table in the VO. ('?' and '#' have different
meanings for URI - although their usage in VO isn't perfect at the
moment)
---
http://wiki.astrogrid.org/bin/view/Main/NoelWinstanley
Senior Java Developer, AstroGrid Project
Jodrell Bank Observatory, University of Manchester
Received on 2006-07-04Z01:05:38