<div dir="ltr"><div><div><div><div><div><div><div>Thanks Mark... that is just the kind of input I was hoping for (even though you agree with Markus on most points)<br><br></div><div>From the application point of view, you draw from how you use VOTable in applications..<br></div><div>Which is to look for the parts you want and ignore the rest.. don't necessarily care if it validates.<br></div><div>No problem there.. <br><br></div><div>However, one is still able validate a VOTable instance. It may not be compete, but <br></div><div>it is possible, and if one wants to make 'valid' VOTable instances, (like the example<br></div><div>serializations I'm making for Cube.. I want to make sure the VOTable structure is correct )<br></div><div>they can do so to a reasonable level.<br><br></div>My view is basically this same approach. The modeling approach we have now allows<br></div>one to validate the instance (SparseCube) if one so desires. If you don't care if it is valid,<br></div>you can annotate the instance any way you want. If Cube is defined to use model A-v1 and<br></div>you annotate to A-v2, and that works for you, all is well. It just would not validate.<br><br></div>The one distinction I see in what Markus is suggesting (please correct me if I misrepresent)<br>is this co-reference idea. Cube uses model A-v1 for element X.<br></div></div> + At element X in the annotation, setup a reference to the annotation for that element<br><div><div><div><div><div><div><div><div><div> + have <b>multiple</b> targets for that reference<br></div><div> 1 = annotation according to A-v1<br></div><div> 2 = annotation according to A-v2 (or B-v1 that contains an alternate modeling )<br><br></div><div>I was thinking about this on the way to work this morning<br></div><div> + I like the flexibility of re-using a lot of annotation, if Cube-v1 and Cube-v2 differ only in <br></div><div> the use of Model A-v1 vs v2, then this allows the annotation to easily comply to both.<br></div><div> + But this is <b>purely an annotation issue</b>.<br></div><div><br></div>Mark<br><br><div><div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Mar 21, 2017 at 6:28 PM, Mark Taylor <span dir="ltr"><<a href="mailto:M.B.Taylor@bristol.ac.uk" target="_blank">M.B.Taylor@bristol.ac.uk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi DM.<br>
<br>
I am reluctant to wade in here, because I've really only skim-read<br>
the thread up till now, and I'm generally not well-informed about<br>
data models. So, if you want to dismiss my opinions as half-baked,<br>
I won't be offended. But since apps and ops are being mentioned,<br>
I'll give my reactions.<br>
<span class=""><br>
</span><br>
</blockquote></div><br></div></div></div></div></div></div></div></div></div></div></div></div>