<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Hi DMers,<br>
<br>
I have a question concerning the VO-DML mapping RFC document
recently published. At the page 24, we read this constraint (in
blue) :<br>
<br>
<img src="cid:part1.216B1FE0.540ED5B9@astro.unistra.fr" alt=""><br>
<br>
<br>
May be I am wrong, but the obligation of removing DM elements of
the MODEL header if there are no relevant data in the VOTable<b>
will imply a very big constraint on the data provider, notably
in the framework of TAP</b>. Concretely, it implies that the
provider system should be able to associated to each column of its
DB the relevant element of the DM or DMs associated to and
possibly to manage the dependencies with other columns to know if
the DM elements should or should not be inserted in the response
(the list of resulting columns depends of the user query .<br>
For instance, how Simbad database TAP system will be able to
generate the appropriate MODEL elements for this kind of request "<i>SELECT
s_ra from basic</i>").<br>
If I am right, I am afraid that the MODEL mapping will be
extremely difficult to implement in the framework of TAP.<br>
<br>
Cheers<br>
Pierre<br>
<br>
<br>
Le 11/10/2017 à 09:30, Laurent Michel a écrit :<br>
</div>
<blockquote
cite="mid:59c3228d-8af2-c51a-7a9d-40d71b2bf468@astro.unistra.fr"
type="cite">Dear all,
<br>
<br>
The RFC page for the VO-DML mapping is open now:
<br>
<br>
<a class="moz-txt-link-freetext" href="http://wiki.ivoa.net/twiki/bin/view/IVOA/VODML-Mapping">http://wiki.ivoa.net/twiki/bin/view/IVOA/VODML-Mapping</a>
<br>
<br>
Please read, test and comment.
<br>
<br>
Regards
<br>
DM chair & vice-chair
<br>
<br>
</blockquote>
<p><br>
</p>
</body>
</html>