[KPhotoAlbum] (yet another) html front end

[jedd]

On Sun Jan 4 2009, Jan Kundrát wrote:
> The ultimate plan (tm) is to drop the XML backend in some point in
> future. Nobody knows when is this going to happen and it's possible
> that we'll change our mind. What is very probable is that the next
> release will include both of them with XML being the default. We'd
> love to make the SQL stable and ready for use as well.

 Hi Jan,

 Yeah - that was pretty much what I'd deduced from what
 I'd seen, and given it's been about two summers (I'm from
 Australia, so tracking Google<tm>'s idea of summer is
 ever a tricky thing) since serious development of a SQL
 backend had commenced.

> When I wrote my web frontend, XML parsing in Python using expat was
> too slow to be usable and I had to convert the DB to a custom SQL DB.
> having to wait for five seconds or more with a database of 20k items
> was not acceptable.

 {nod}

 What I do is keep a very stripped down subset - just the images
 that are thus tagged for publishing - in an array dump within
 PHP.   It's 116k for the 200 images that I've got published at
 the moment, for example.  If the index.xml is younger, then I
 pick it up and create a new cached array dump.  Otherwise I
 just load this smaller file.  Performance appears pretty snappy
 and I think this is a convenient trade-off.  My primary database
 has just breached 20,000 images, btw.  The initial load is pretty
 fast, but don't have numbers to hand (a few seconds I think,
 on an Atom 1600mhz).

 cheers.
 Jedd.