Re: TECH: SQL Database Schemas

From: "Arthaey Angosii" <arthaey@...>
Sent: Sunday, March 26, 2006 2:55 AM

> It seems as though many of us have used or are planning to
> use SQL
> databases to store our lexicons. I'm planning the same
> sort of thing,
> and I was wondering what schemas others had settled on and
> why they
> didn't use Kura instead.
I run Windows on my machine and this won't change very soon.
Back when I had Linux on my old machine (R.I.P.), one day
Kura decided not to run anymore because I made a stupid
error occur. I couldn't uninstall the program either for
some reason. Anyway, I never really understood how Kura
works ... there are so many dialogs. OK, I could use Shoebox
on Windows, but I decided to work with an SQL database. So
the schema I use is as follows:

  attributes(WordID, attribute, _AttrID_)
  translations(WordID, translation, FromExample, ToExample, comment, _TransID_)
  words(_WordID_, word, pronunciation, wordclass, date)

I think it isn't that good to have the 'date' in this
relation. It should rather be part of the 'translation'
relation because it wouldn't be impractical to see when
additions/changes have been made to an entry. It should
nevertheless be possible to dump the stuff of that column
into an xml file and read it in again in 'translation'. I'm
using a PHP/HTML based interface to make entries etc.
Deleting entries is yet only possible using phpMyAdmin.
Other than that, I'm contented with this schema. The only
real problem is that deleted indices in 'words' are not
filled again, so that the actual number of entries and the
'WordID' differ a little.

Carsten