Hoi, I have started anew on an ERD for the UW. It still has few files, they will grow but it is already quite complicated. I have looked at the ERD from Ilse, and found it to be next to useless for a multilanguage dictionary. This is MY first shot at it, I am still in design mode and my next exercise will be to include pronunciations. That is more complicated that it sounds I am afraid because you also have trancriptions to consider and the fact that phonetic script is NOT standard; Americans do transcipe to IPA and SAMPA differently from the French or the Germans.

One lesson I have learned, since the last time I looked at this in earnest I have learned a lot. This has complicated my perception and it should make for a data design that is better at solving the issues as I know them. I have looked at several designs and rejected a few, there are some designs that I want to revisit.

I did send this mail to uw-creations, a google group that we set up for the implementation of a reference implementation for a Computer Aided Translation (CAT) tool. This design is as relevant to them as it is for the Wiktionary-l. Everyone can read the content of this list, only people who have been invited can post there. (drop me a line)

As always, I am happy for your comments. In the spirit of Open Source I will publish often.

Thanks, GerardM