Vocabulary terms in MANGO

[Laurent Michel]

Hello Pat,


> On 10 Nov 2024, at 23:55, Patrick Dowler via semantics <semantics at ivoa.net> wrote:
> 
> 
> Yes, I have been thinking about it and have similar questions with regards to CAOM and
> vocabularies it uses/refers to.

We are maybe in a good time slot for introducing vocabs in DMs.
For the record, I’ll start to exercise this with the calibration levels.

> 
> For DALI xtypes in particular, I think it is in scope for some DM discussions this week whether 
> and how to define those data types in the base ivoa data model (which has some things already
> that are not purely primitive types). It is technically doable in VO-DML and then would be
> easily imported and used consistently across various data models. I think this makes sense and
> fits my view that xtypes are not really a vocabulary

Yes, technically there no problem with using vocabularies to specify allowed values for particular model attributes.
However, this requires the vocabulary to have a hierarchy setup in order to prevent mixing words related to different domains such as region vs sky coordinates, which is not the case today with Xtypes.
In my understanding, xtypes are not really a vocabulary in a sense that DALI defines them as a flat list of words.
If we use them in MANGO, we must first transform this list into a real vocabulary with a concept hierarchy.

> The question for MANGO would be: if all those data types were in the ivoa base model and DALI 
> was specifying the serialization of a single value, how would that change/work in MANGO?

If we can issue a vocabulary from the xtype definitions, there will be no major change for MANGO.
The question is ot know whether if worth it to build that vocabulary or not.

Laurent
> 
> --
> Patrick Dowler
> Canadian Astronomy Data Centre
> Victoria, BC, Canada
> 
> 
> On Fri, 8 Nov 2024 at 09:29, Markus Demleitner via semantics <semantics at ivoa.net <mailto:semantics at ivoa.net>> wrote:
>> Laurent,
>> 
>> On Fri, Nov 08, 2024 at 04:43:43PM +0100, Laurent Michel via semantics wrote:
>> > I totally agree with you.
>> > Xtypes provide some words that must be reused in MANGO.
>> > The problem is to find a convenient way to connect both worlds.
>> > Xtype definitions are sparse in the VO documentation and there is no hierarchy ordering them:
>> >     -  "xtype=hms" is at the same level as "xtype=moc"
>> >
>> > The connection with MANGO could be done just through the attribute definition:
>> >     - the value of this attribute must match one of (A,B,C) which
>> >     are Xtype allowed values as defined in XYZ
>> >
>> > Or dare I suggest another option which is to put Xtype values in a
>> > specific vocabulary (xtype.rdf)?
>> 
>> I think the idea of having a machine-readable list of xtypes as a
>> vocabulary has come up before, although I couldn't say where.  I
>> could see some applications beyond MANGO, in particular when we
>> consider re-opening xtypes beyond DALI.  Hm.
>> 
>> Pat, are you reading this?
>> 
>>            -- Markus
>> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/semantics/attachments/20241111/00b0b6f1/attachment.htm>