zeitgeist team mailing list archive

[Mikkel Kamstrup Erlandsen <mikkel.kamstrup@xxxxxxxxx>]

2009/10/9 Mikkel Kamstrup Erlandsen <mikkel.kamstrup@xxxxxxxxx>:
> 2009/10/9 Seif Lotfy <seif@xxxxxxxxx>:
>> Hey guys here my quick take on why we are under time pressure:
>> 1) We need to show GNOME that we have been coming forward and not standing
>> still after the 0.2.1 release
>> 2) We need to get 0.3 rolling quickly since alot of projects are starting to
>> depend on the API of 0.2 and incase we want to break it should happen now!
>> 3) Keep up the momentum
>> 4) We can take a big break afterwards :P
>> 5) Teamgeist and Docky want to fully depend on Zeitgeist thus we should meet
>> them halfway (You might think why care if they want us they should stay with
>> out momentum, but this is a wrong approach, especially if teamgeist is
>> being developed by Collabora which will guarantee us alot of success if we
>> make it (if we don't they have the manpower to rework zeitgeist from scratch
>> making us redundant. Plus Docky is one of the most used apps on GNOME which
>> will give us a prestige, they don't need us but we have the honor of them
>> wanting to depend on us)
>> 6) tackle open issues and narrow down problems for the hackfest
>
> Agreed on all points. Breaking API does not worry me as much as long
> as the transition is easy. I don't expect Docky to have a lot of
> problems, Teamgeist seems more closely related to Zg so we should be
> on the lookout here.
>
> Our big problem is that the real use cases are only starting to
> trickle in, up until recently we been designing the API for the
> Journal and some vague extra ideas. This is a very thin foundation for
> building a general purpose API, I hope people can appreciate that. But
> as I already said I still think we should make the ride as smooth as
> possible.
>
> One thought I did have was to write a general purpose migration tool
> for the DB. This is of course huge over engineering if we only need to
> do this once, but if we have to do it more then it'll save us some
> time. Plus - it could function as a backup tool too. Here's the simple
> idea, much like Tracker does: Import/export RDF in N3 or Turtle (we
> should use the same as Tracker). It would require that we somehow
> version the database so we can automagically detect schema upgrades.
> I'll try and write a blueprint for this.

Here you go; follow the link to LGO for the full spec:
https://blueprints.launchpad.net/zeitgeist/+spec/db-migration-and-backup


-- 
Cheers,
Mikkel