Re: GSOC 2011



ons 2011-03-23 klockan 14:19 -0700 skrev Peter Kaminski:
> Simon,
> 
> I think it's a worthwhile project, certainly.  I've dabbled with F-Spot 
> as a user, but I don't really use it.  However, if it were 
> cloud-enabled, I might be able to make good use of it.

Cloud-enabling it is the natural extension of this project, I don't
think there will be time to setup such a service during the summer, but
that and some other ideas are floating around in my head :)

> 
> Another option, 3: Use a server-side SQLite 
> <http://www.sqlite.org/cvstrac/wiki?p=SqliteNetwork> instead of a local 
> SQLite.  Then like #1, synchronize the picture folder on a file-by-file 
> basis.  Rationale: I think you want a single master to store the 
> metadata.  Once you have multiple copies of it, things might get really 
> messed up.

Yes, data in multiple places can easily become a headache. The case for
networked SQLite seems to be that the existing queries can be reused;
just change the connection to a remote one and off you go. 

> 
> I guess you could get a similar result, and maybe make it even more 
> generally useful, by using a remote MySQL instead of SQLite.

MySQL or PostgreSQL likely outperform SQLite too, but I guess the load
on a personal installation won't be that high.

> 
> And a crazy (but cool) option, 4: use Flickr as the primary image 
> store.  Probably with a server-side SQLite to handle metadata, and of 
> course, local thumbs and optional synchronizing of local/remote full images.

Amazon S3 might be usable for this too, though they don't have fixed
pricing as Flickr do. With flickr there's the problem of multiple
versions though.

One thing is clear: there are a lot of possibilities. One thing that I'm
thinking about is if it would be possible to make F-Spot instances
notify each other when changes are made so that synchronization can
happen immediately. Another is that maybe background jobs can be run
directly on the remote copy. That would likely require a server
component though.
Anyway, those things are not something that there will be time for
anyways, but it might influence the design decisions.

-- 
Simon Lindgren



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]