Difference: UtypeTigerTeamMinTel3 ( vs. 1)

Revision 12012-09-19 - MarkusDemleitner

 
META TOPICPARENT name="UtypesTigerTeam"

Minutes of the Telecon of the UTypes Tiger Team, Sep 19, 2012

Participants: Matthew, Mireille, Pierre, Gerard, Markus, Pat, Jesus.

Action item from last week: rank use cases and requirements. Only 50% have actually done the ranking -- so, everyone else, please contribute.

Longer discussion about why 0 of 2 polled wanted R#3. Markus will eludicate why he thinks that's important in a Use Case.

Gerard says that VO-URP provides a mechanism that lets you specify that at position X within a data model there is an instance of Y, which would take care of some of the motivation for having two utypes on one thing. Markus says that this doesn't really help when, e.g., Obscore only has RA and Dec, but full STC allows derivatives, errors, etc on them.

On the use cases ranking: Again serialization and deserialization are favourites.

On the possible use case of representing enumerations using utypes, Mireille points out that utypes are no ids, and thus no such use case should be adopted; on this, there's broad agreement.

Markus notices that UC#1 has a hidden requirement that there is a metamodelling language as part of the utypes spec. He's not sure there couldn't be a way to get a sane utypes spec without having to talk about metamodelling, which he fears could delay the spec (so, he'd like to drop his support for UCs 1 and 7).

Gerard points out that having a metamodel helps you figure out how to serialize not only in VOTable and avoids ad-hoc hacks; whether there's many use cases that make metamodels very useful at run time he doubts.

Gerard promises to have a look at the list of requirements mailed around by Omar and will comment how much of it is solved by VO-URP.

There's going to be a plenary on utypes at the next interop. We must be able to present what our discussions have yielded there.

Matthew will insert VOUnits and VOEvent material into the current usages document; Gerard will touch up the SimDB section. The Photometry DM discussion we had on the mailing list should also be reflected, but it seems the matter is still a bit in flux. Omar and Jesus will try to write something.

Matthew feels we need at least two telecons before the interop. Omar and he will put out a doodle poll next week.


<--  
-->
 
This site is powered by the TWiki collaboration platform Powered by Perl This site is powered by the TWiki collaboration platformCopyright © 2008-2024 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback