Mirror, mirror on the wall Who's the fairest text of all? |
The SimHebrew Bible is well underway. It has been a fascinating project to be a part of.
I, Bob MacDonald, as technical support, have learned an immense amount about the differences between the pointed text of the Tiberian tradition, and the unpointed text of the modern Hebrew newspaper. I have generated all the Bible in SimHebrew, book by book. This pandemic project is described in a long series of posts such as this one on Unpointing the Hebrew Text. One of its byproducts is the concordance here, a very easy way to see every word and all its uses in the Bible in one place.
Jonathan Orr-Stav, my colleague, and the inventor of SimHebrew, has completed writing the SimHebrew one chapter at a time for the Pentateuch and the Prophets and is well into the Writings, so we are nearing project completion. To check his process, we have developed a validation sequence which we have completed for the first 8 books.
Our validation sequence looks something like this:
- Jonathan produces the SimHebrew text chapter by chapter, using his own algorithm to convert from the unpointed Masoretic text of the Hebrew Bible.
- I generate the SimHebrew text from the Leningrad WLC text, which is fully pointed, book by book, using my own DBATE (databased algorithmic translation engine) system.
- Jonathan then compares his version with mine, for discrepancies and errors in either one, to produce a single reconciled version, and to highlight where the DBATE algorithm is incorrect.
- I adjust the algorithm, and review any other decisions about the text.
- I regenerate any files that are affected by changes in the algorithm to our shared work area.
- I generate the candidate html input to the epub software.
No comments:
Post a Comment