Forward: Page web inacessible - Unreachable web page

[Carsten Svaneborg]

On Thursday 09 February 2006 15:04, Jonas Bosson wrote:
> I have some trouble finding dependencies.

Ask if this is not sufficient map of dependencies:

All the plugin scripts depend on the database and the wiki,
but not on anything else.

New data is entered into the DB using import.pl

import.pl depends on the HTML patent texts that are downloaded
by pleech.pl

updatebot.pl depends on the DB to schedule patents due for updating
and pushes data into the Add,Del,Update tables.

adddelupdate.pl processes the Add,Del,Update tables, kills patents
in the DB, uses pleech to download new HTML patent texts, and
import.pl to import these.

download_ebd.pl runs weekly and downloads EBD files from EPO.
It depends on process_ebd_sgml.pl to do the processing.

process_ebd_sgml.pl processes these files, and schedules applications for
update when an EBD file lists them as being granted.

optimize.pl is run periodically and depends only on the DB. It generates
precalculated indices like the ECLA browse tree, and some of the slow
threeway merges, that can't be done online.

I think that is basically it.

Currently with the DB being at the centre of things, the problem is to keep
the HTML patent files and the DB concent in sync, that is done by 
adddelupdate.pl  (calling pleech.pl to update the files, and import.pl to
synch the DB).

Currently all the patent related  info in the wiki is dynamically generated
from the DB, so no synchronisation is necessary.

With static wiki pages, these needs to be tied into the adddelupdate
process as well. E.g. delete needs to delete all instances of an idx in
the DB, delete the HTML patent file (actually move to backup), and
now also delete the static wiki pages.

Add will have to download the new HTML file, import it into the DB, and
create a new wiki page(s) and so on.

> I am not sure if moin moins single directory-layout will work well.
AFAIK it has started to put files into directory tree structure, or am I
wrong on this?

> A well structured compact format and best a good directory layout is
> probably the most important issues since 75% time is spent waiting on
> IO.

Is this IO on viewing patents, or IO in general?

-- 
  Mvh. Carsten