New Sindarin Dictionary application

[Lúthien]

Awaiting response from either Christopher Gilson or the Tolkien Estate - both seem to consider our questions re. copyright not worth answering it seems

[Eryniel]

Indeed, I have sent a mail to both, but have not even received any indication they are reviewing the matter. And currently I am unsure if it would be wise to send a "reminder"....

[Eryniel]

UPDATE: After finally sending a reminder of sorts via the official Tolkien Estate website I got a short reply from Adam Tolkien (really could use a jumping up and down smiley here)

He said it might take a considerable amount of time until we got an answer, but he would like more detailed information. A good start after a long wait I would say...

So I replied giving all the info as well as a link to download the application... I am hoping they will find that we are giving considerable thought to the quality of the data as well as having a solely non-profit approach and will "sanction" our project.
Let's wait and see, fingers crossed...

[Lúthien]

(really could use a jumping up and down smiley here)
...

?

[Eryniel]

Hahahaha, I guess that one would do

[Lúthien]

Hi all,

I moved the exchange with Mr. Willis to the Staff forum because it was offtopic.

If you have any questions about it (and cannot access the staff forum) let me know and I'll give you access, or send you a copy of the discussion.

[Lúthien]

Update - we have been talking with Aran about using his word list as possible source for the dictionary. I'm looking into that possibility now, and it seems very promising. It could probably also hugely improve performance, because the structure of his dataset is a lot less complex than the TEI subset that I used until now (which came from the Hisweloke list). On top of that, his dataset includes also the PE17, 18 and 19 entries.

I'll first be looking at the data itself, because the English (and possibly French) entries will need to be re-added. I don't think that that will pose any major problems, though.

[Lúthien]

Suilad Aran,

I had a better look at the data the past two days. There's a couple of questions that came from that, as well as a first draft for a database model that would accomodate other languages besides German.

I don't understand the meaning of the columns srek, saltrek and s2rek in table wortliste.
These columns only take the values 0, 1, 2, 3 or 9 - it seems as if the names suggest they refer to the columns sind, sind_alt and Sind2 - although I can't figure out what -rek refers to ?
It doesn't seem to be an index to separate otherwise identical rows for the above columns, but maybe I should check that again (though it'd be great if you could just tell me )

Also, I don't know what the second table wordliste_komm is for. By its contents I'd think that these are the somewhat uncertain entries that you talk about (like _lorn_) but if so, how does this table fit into the data of the first one?

I puzzled a bit with a possible database design. As to what you mention about (for instance) homophonous words; I want to keep the internal cohesion exactly as it is. The changes that I'm thinking of would therefore leave all that intact; it would just distribute the components in another way across tables so that it would become possible to add other (modern) languages. It would also eliminate the need to enter redundant data, like when there's two rows in wortliste for two translations of one Sindarin word. Here's an ER diagram:



It's of course entirely possible that I missed out something, so if you can look this over and comment, that'd be great!
I also left out the srek, saltrek and s2rek culumns because I don't understand what they mean; and the second table is not yet considered here.

I'll also post this message on the topic about Aran's new German <-> Sindarin word lists on Mellyn Lammath (SD -> GER, GER -> SD)

Thanks again!

NB - it just occurred to me that this model does not only allow for other modern languages to be added (like English, French ... ) but it can also accomodate another elvish language like Quenya (in that case it'd be better to rename that one Foreign Key column to ELVISH_ENTRY_ID)

Or, even better: if Quenya words would be added, you'd also get a Sindarin-Quenya dictionary for free .. at least for the words that exist in both languages, of course.

[Lúthien]

Just a quick heads-up in here because the discussion is mostly taking place at the German forum at the moment:

We're now trying to cast Aran's data in a form that a dictionary application could use. That's the relational database model I pictured above. Unfortunately, things aren't as straightforward as we'd like, even though this is not very difficult in essence. There are all sorts of complications that come with handling characters like the diacritics used in Sindarin (and also German): how to sort them, how to compare them, etcetera. This can be a real nuisance if you compare two sets of data, ie. is _dol_ the same word as _dôl_, or not?
We've compared the Hisweloke list to Aran's list, and that yields some interesting problems like described above. Anyhow, we're progressing still!

[Lúthien]

A little update:

work is still continuing, even not as fast as I would want it to. That's mainly because of the same old "being very busy at work yadda yadda issue" - sad, but true.
But we'll get there.

What we've done: I have completed the data model, and Eryniel and Roman have completed the data set, which was a very time-consuming task.
I'm almost done migrating those new data into the database, and then I can create a first version of the application.