Draft draft 0.04

[Norman Gray]

Folks,

I've implemented some of the changes that Alasdair and Rob discussed,  
partly because I'm going to be heads-down in something else for the  
next couple of weeks and so will have to stay away from vocab stuff as  
much as possible.

On 2008 Feb 8, at 10:19, Alasdair J G Gray wrote:
>> Rob Seaman wrote:
>> stuff
> I have just read the document and will add some detailed comments  
> below. I am happy to go ahead and perform both Rob's suggestions and  
> my own

>
>> version: bump it up to something like 0.94 or nobody will take it  
>> seriously

Version numbers are tuples of non-negative integers: they have an  
ordering defined but no metric.

>> editors: we need to identify these ASAP, the document doesn't need  
>> much work IMHO

At Andrea's suggestion I've expanded the author and editor lists.

>> Status: let's promote this to a public working draft ASAP

I'll keep changing until Andrea (in loco WG chair) calls time.

>> TOC: didn't verify

It's generated.

>> Replace "<what/> element" with "<Why/> and <What/> elements".

Done.

>> Replace "Gamma Ray Burst" with "gamma-ray burst".  Need to identify  
>> and follow a general policy for terminology, especially in a  
>> document about vocabularies :-)

Thanks -- I'll find a lot of these when I next read through the text  
for style rather than just sense.

>> Delete the word "modishly".  No need to be snarky about folksonomies.

Fair enough.

>> Delete ", with its systematising instincts, and aware of the  
>> benefits of standardisation,"

Interesting.  I've changed this to "the astronomical community, with a  
culture sympathetic to international agreement, can do better."  Plain  
'you can do better' has a disappointed-parent sound to it; I think  
it's interesting to think briefly about _why_ astronomy can do better  
without trying.

>> Replace "supernova1a" with " type 1a supernova".

Yes, better.

>> BTW, how are plurals handled in SKOS?  Meaning - I understand we're  
>> supposed to pick a coherent naming scheme, either plural or  
>> singular, but is there some way to recognize SNe as referring to  
>> multiple SN?  Or is my question ill-posed?

If you need to programmatically distinguish one from several  
supernovae, vocabularies won't help.  A user's search for 'supernova'  
and for 'supernovae' should both translate into the same concept  
internally.  The way that would happen is via labels: the prefLabel is  
what would typically be presented in a UI, and the altLabels and  
hiddenLabels contain strings that will help lead you there.

Thus the IAU-93 concept #Supernovas will have prefLabels  
"Supernovas"@en, and would ideally have both "Supernova", "Supernovae"  
and "SNe" as altLabels (the multiple plurals this this one rather an  
awkward example, "Supernova"@fr, "Supernovae"@it -- I've  
experimentally adjusted the build of the IAU-93 SKOS vocabulary to add  
a few extra altLabels such as these).

>> SIMBAD is sometimes SIMBAD and sometimes Simbad.  Which is it?

The CDS front page has it as 'Simbad', so that's what I have, now.

>> 2: good
> Need to rewrite the introductory paragraph as it claims this section  
> introduces something and then points to 3.2.

Well spotted -- fixed.

>> 2.1:
>>
>> Delete "NOTE: The purpose..." or perhaps expand the following  
>> description.

The note in question is indeed a bit too tightly in the context of  
library science to be illuminating.  I've added some description later.

>> I don't think "a vocabulary (SKOS or otherwise)..." needs to be  
>> bolded.

If I thought I could get away with putting it in bold capitals inside  
<blink>, I would.

>> Shouldn't be shy about announcing a preferred format (XML versus  
>> Turtle).

The problem is that the two serialisations are fully isomorphic, and  
so fully interchangeable that I fear that preferring one over the  
other would look odd.  More practically, almost anyone producing a  
SKOS vocabulary would have the tools to translate between the two  
formats trivially; the only reason I don't suggest mandating _both_ is  
because there will I'm sure be some cases where for technical reasons  
someone will find this parallelism difficult, in which case we  
shouldn't inconvenience them by requiring one or the other.

>> A SKOS entry may contain a "definition for the concept, where one  
>> exists in the original vocabulary".  This notion permits the many- 
>> to-many mapping we need.  A particular vocabulary may include the  
>> limited "I am a spiral galaxy" type of definition.  A separate  
>> vocabulary may include a concept with a more fleshed out  
>> definition, perhaps including (or simply plagiarizing) links  
>> covering all of what astronomical science has to say about the  
>> birth, life and death of spiral galaxies.

It's also technically easy for one person to define the concept <http://me/myvocab#Term 
 >, and for a separate person, at a separate URL, to assert <http://me/myvocab#Term 
 > skos:definition "means stuff".  This of course raises issues about  
provenance and trust which are big, interesting, complicated and which  
we had probably best not talk about just now.


>> 2.3:
> This section needs a complete rewrite to reflect the changes to the  
> skos standard.

I'll leave that to Alasdair.

>> I remain a bit unclear about where the equivalences are going to  
>> live.  What document will contain the example "iau93:#SPIRALGALAXY  
>> map:exactMatch ivoat:#spiralGalaxy"?  I suggest the need for some  
>> document external to both iau93 and ivoat to make this  
>> equivalence.  How are multiway equivalences conveyed?
> These would be in a separate document from the skos vocabulary  
> meaning that users can pick and choose which sets of mappings to  
> use. When I rewrite this section I will expand the discussion. We  
> also now have a first mapping file relating the terms in the A&A  
> keywords to the AOIM. I feel that a description of this should be  
> added to section 4 and a copy included in the appendices.

Sounds good.  I think that the detailed curation and distribution of  
the mappings files is one of the small set of technical issues  
remaining to be worked out, alongside the issue of versioning.

> What does the normative mean in the title? What purpose does it serve?

None, now.  It's a relic of what I'd _thought_ was a decision about  
which of the distributed vocabularies was to be regarded as a formal  
product of this standard, and which were just examples. I've removed  
them pro tem.

>> #5 - typo "defintions" (saw a similar typo somewhere else in the  
>> document)

Multiple places!

>> #8 - Again, does the publisher of iau93 map to ivoat, or does ivoat  
>> map to iau93, or both, or one or more third parties?
> This needs more explanation in the document. Both parties could  
> publish mappings along with anyone else who decides to create a set  
> of mappings. Of course, the mappings themselves all have inverses,  
> so if you have a mapping from vocabulary A to vocabulary B then you  
> automatically have one from vocabulary B to vocabulary A.

Provenance, curation, trust: all fashionable concerns, no easy answers.

>> 4: ok (meaning looks like you guys did a vast amount of work)
> We claim that all of the vocabularies are normative versions except  
> for the IAU. This is not correct by my understanding. The normative  
> version of each vocabulary is surely the version published by the  
> vocabularies creator, e.g. the publisher of the A&A journal or the  
> IVOA document that defines the AOIM. We are merely providing skos  
> versions of these existing vocabularies.

True, but until A&A start publishing SKOS, I suppose we can assert  
that ours is the IVOA-normative version of the SKOSified one.

> 4.2 There are some alternative labels and scope notes.

I've just now added some support for the generated IAU-93 SKOS, which  
allows very mild garnishing of terms with further altLabels (in  
particular <#Supernovas> skos:altLabel "supernovae")

> 4.6: How is the IVOAT going to be published? Is it time that we  
> started preparing a document for its publication so that we can  
> reference it here as an early version?

A good question.  I favour the IVOAT being published initially as one  
of the normative vocabularies in this document, and being thereafter  
maintained via a separate process.

>> Bibliography:
>>
>> If the UCD authors are listed, so should the VOEvent authors.

BibTeX glitch.  I've updated the .bib file and the relevant BibTeX  
style file to cite editors over authors for documents like these.

Thanks for the detailed comments.

Norman


-- 
Norman Gray  :  http://nxg.me.uk
eurovotech.org  :  University of Leicester