Re: [Tracker] My disagreement with a recent change in the ontology for GNOME Notes

[Philip Van Hoof <philip codeminded be>]

On Thu, 2013-07-11 at 13:52 +0200, Pierre-Yves Luyten wrote:

Hi Pierre-Yves,
 
> > I should probably have watched the ML more carefully now that it's
> > committed and too late.

> My bad, I hope the fact there was no release yet allows us to revert
> without concern.

No problem, no hurry, take your time to fix Notes first.

> > We can't query it, because its contents are formatted and
> > we don't have query capabilities to query your formatted contents. And
> > even if we would allow storage of formatted content like XML, we should
> > enforce the format using a DTD or other XML schema so that we're sure
> > that the inserted format is valid and known.
>
>
> I think pushing data into Tracker make sense if the format is
> understood by several applications. Which is why i first thought "HTML",
> and pushing there only the *representation* of the note.

Not only should the format be known, common and specified; we also need
to have a possibility to query it.

With XML data we would have to bring in XPath inside of SPARQL to query
on fields that are XML. This would have profound effects on the query
speed as this goes completely outside of the realm of SPARQL (meaning
that on for example a Cartesian product, during a join, and/or subquery,
we'd have to perform the XPath query on each row: bye bye any form of
reasonable performance, even when dealing with a few hundred rows).

And of course XML allows you to store subfields in fields:

<note><person direction="from"><name>John</name></person></note>

How do we query for all notes that came from John? With SPARQL (without
having to introspect the nfo:xmlContent)?

Instead you should of course use the NCO ontology and relate a contact
as the from of your note (like the NMO does for MIME among onther
message types).

> However in the specific case of notes, XML seemed correct since Tomboy
> format [1] is already understood by several applications, such as
> Tomboy, Gnote, Conboy, Tomdroid, ongoing Macboy - and Bijiben. (In the
> XML case however, we can notice if we stick to Tomboy format, both rich
> content and metadata are to be stored.)

The text/html and the iCal format are also already understood by a
multitude of applications, and yet neither NFO, NCO, NMO nor NCAL are
storing HTML (NFO, NMO), from/to as iCal (NCO) or the calendar item as
iCal (NCAL): the ontologies model the data instead.

You need to do the same for notes. I'd be ok with creating a Nepomuk
Notes Ontology, if that suits the needs of Tomboy, GNote, Conboy,
Tomdroid, Macboy and Bijiben. Get those people's heads together and
negotiate a appropriate Notes ontology. Then talk with us and/or KDE's
Nepomuk team and then we can as a group in agreement propose this to
upstream Nepomuk.

I can't imagine nfo:xmlContent being accepted and so this way Tracker
will be the only Nepomuk implementation that has it. That's not good for
anybody.


> > Don't abuse Tracker to solve your 'whatever' storage needs.
>
> Sure, this is not the idea at all.

Then we agree :-)


Kind regards and thanks for picking this up,

Philip

> Regards,
> Pierre-Yves
>  
> [1] https://wiki.gnome.org/Tomboy/NoteXmlFormat

Sure. Nice format. For you to store.


-- 
Philip Van Hoof
Software developer
Codeminded BVBA - http://codeminded.be