Weekly Report #2

[Asgeir Frimannsson]

Just a quick and late report from the 'Implement support for XLIFF 1.1
in KBabel' SoC project:

[Planned tasks for previous week]
 - Branch work/kde4/kdesdk/kbabel to SoC project
 - Work on designing XLIFF based data-model
 - Remove (or abstract) PO centric features to ease the porting to XLIFF

This week has been mainly spent on designing a new data model for a
XLIFF based Kbabel. At first I tried refactoring the current data
model, but after a while I realized that there were too much PO
specific features that depend on it, so I started from scratch, using
Kbabel::Catalog only as a reference. So the new data model will start
from XliffDocument, and support a subset of the XLIFF 1.1 Spec (but
extensible enough to support all). When I've got this libkbabelxliff
working, I'll create a QT4 item/view model around it, specifically a
listview and a hierarhcical view, and implement a QTextDocument based
editor. Finally I'll merge this with the KBabel 'skeleton', where all
PO features are removed.

Yes, this means that KBabel will initially loose a lot of its features
(which must be re-implemented to support the new data model), to quote
aseigo [1]: "sometimes to go forward you first have to go back". So
the plan for the rest of this project is as follows:

 1) Create a XLIFF based data model (libkbabelxliff), using QXml as a
backend, mapped after the XLIFF Specification. This can be used both
for the import/export filters and for the editor data model.
 2) Create a Qt4 model/view (interview) application using the new
model and editors.
 3) Integrating this with the KBabel 'skeleton' application.

As for further progress, I'll be away on holidays [in Norway/Malaysia]
from 23rd July - 12th Aug. I don't know how much internet access I'll
have in that period or how much work I get done, as I really need to
recharge my batteries for starting a PhD in September :) But looking
forward to a intense period of coding 13th-31st Aug.

cheers,
asgeir

[1] http://aseigo.blogspot.com/2005/07/sometimes-to-go-forward-you-first-have.html