Thursday 24 December 2015

Project status

Merry Christmas to all and to all a good day. (There is but one and that one ever - Herbert knew this even before Einstein.)

I really will slow down for 10 days I am sure. I have just synchronized my backup database offsite - proving I could recover if I lost my computer. Nice to know.

Completed is a relative term when it comes to projects. But here's the relative progress.

Now for interpretation: chapters completed so far 262 of 929. These are still subject to change. I had tightened my algorithm on the relationship between stem, gloss, and semantic domain too soon. There are so many misinterpretations in my data that I was masking my own errors. I will figure out a way to document the acceptable deviations from my concordance patterns with a new table of exceptions - soon. In the meanwhile I removed the domain from the test and exposed all my unacceptable errors, and there were many (over 400 at first, now - two days later - reduced to less than 100 - only .1% - I will try and keep it as low as is reasonable, given the nature of language).

That should slow me down a bit. Many of these are internal data inconsistencies rather than direct gloss errors - i.e. where I have transcribed the same name with two differing stems. This is acceptable if the English is a homonym, or there is a special reason like for an acrostic, but sometimes, I have just got an error in the stem definition. And of course at this stage the domains are very fluid.

In October I had over 75000 complete unknowns. Today this has reduced to 61724 largely as a byproduct of minor changes.

The count of words done is an estimate based on the sum of consecutive words anticipated by the software. Actual words read for completed chapters and verses is currently 74,077 of 304,730 (25.4%). So on all counts we, I and my computer programs, are over 1/4 through.

The work surface evolves also as I need new ways of searching, updating, and correcting. It is a very easy surface to change, ideal for an old database programmer such as I am to manipulate.

In the first quarter of 2016, I will be back at UVIC for a second fellowship. I hope to start a book on the music.

No comments:

Post a Comment