[MOBY-l] Presentation on the WWW site?

[Jason E. Stewart]

"Mark Wilkinson" <mwilkinson at gene.pbi.nrc.ca> writes:

> perhaps a bit of both worlds is the solution, similar to how GO
> started?  A flat file for day to day updating, and a database which
> is rebuilt on a ~regular basis.  Presuming that one day the flat
> files will disappear it allows me to start coding the "end product"
> right away.

That is the same thought I arrived at. I'm not sure though.

Chris is helping me get midori's ontology into a Pg DB, and I'll play
with both the file -> DB updating and the straight DB querying. 

The thing is, the GO DB table structure is really rather heavy-weight
compared to MOBY itself. I think we may want to export the ontologies
to something like XML-RDF and then import them into the MOBY registry
and only keep the information we need, and not all the history
tracking, etc. In order to export to XML, though, we have to have the
ontology in a DB.

> > Why not just use the top-level <MOBY> object for this purpose? The
> > <ID> object would have no new attributes, so it seems redundant.
> 
> That wont work under the current structure.  <MOBY...> is more of an envelope
> than an object.  

Sorry, Mark, my nomenclature was confusing. I didn't mean the
<MOBY>...</MOBY> envelope I meant the MOBY *object* that is the root
of the Type Hierarchy. That object has all we need.

> I can't see any way of getting around having a 'generic' ID object,
> which then has a CV attribute of namespace to describe what kind of
> ID it is.

Right, in my mind that generic object is the class that is the root of
the type hierarchy, maybe we shouldn't call it 'MOBY', (especially
since in XML it would come out as <MOBY> which is already reserved for
the MOBY envelope), maybe it should just be 'Object'?

> In a sense, this suggests that all other objects are overloaded in their base
> form...  i.e.
> 
>     <Sequence    namespace="GenbankAcc"   id="A234742.2">
>         <Seq> ACTAGCTAGCGTATCGATCGTA</Seq>
>     </Sequence>
> 
> is actually read as "The Sequence object with the ID object of namespace
> 'GenbankAcc' and value 'A234742.2' has the sequence ACATGCATG..."   Would it
> perhaps be more correct to write it this way:
> 
>     <Sequence>
>         <ID   namespace="GenbankAcc"   id="A234743.3"/>
>         <Seq> ACTAGCTGATCGTACTG</Seq>
>     </Sequence>
> 
> Which representation do you think is more consistent?

Definately the first! I want to avoid the second at all costs. So
looking back at the hierarchy that Lukas sent out, we would have:

Object
  |
  |-Sequence
  |
  |-MAGE
  |
  \-etc..

Which in the XML would give us:

<MOBY>
  <Sequence namespace="GenbankAcc" id="A234742.2"/>
  <Object namespace="GenbankAcc" id="A234742.2"/>
</MOBY>

I'm still having a lot of trouble with parts of the type system,
though, but that for another email.

jas.