The next file format

[Anša Vernerová]

Hello everyone,

here are my "advanced user" thoughts on the new format. They are all
concerned with the idea of editing aspects of the collection outside
the apps provided within kde-edu.

With the current kvtml, I have been using simple scripts for tasks
such as moving all images to a new location, swapping the order of two
languages, and copying sound entries from one language to another. I
think it is not just me: if the collection would not be stored in some
kind of self/explanatory plain text format, I would never dream of
being able to accomplish those tasks. Even just hiding the
self-explanatory plaintext into a compressed file with obscure ending
might discourage me from trying. Really, calling it collection.zip
instead of collection.xyz would make a difference!

I perfectly understand the need to have a single compressed file for
uploading collections with media files. However, what I also enjoyed
about the current way of handling things was the possibility to
actually edit the media files independently of the kvtml file. (Like
changing the image table.jpg for a different table.jpg without even
opening parley; replacing a sound file with a more intelligible one;
reducing the size of all images in the directory or exporting
thumbnails for html preview; running a script to equalize the sound
volume across sound files acquired from different resources.) I also
reuse images a great lot, both within one collection and across
collections. How will this be affected by the new format?

I also want to support the idea that IDs used for crossreferencing
within the new format should be meaningful. Sometime long ago on on
this discussion, someone has suggested something along the lines that
the text content of the entry should act as the id unless there is a
collission. I think this would be great, not just for people who want
to edit the file(s) by hand, but also for people who are looking into
the format because they want to write a simple script to accomplish
something. Numerical IDs are hard to make sense of.

Anša