@disciple I’m no help for the programming part of this, but for the BH part: you’re welcome to ask text critical questions here. We have a dearth of such questions about the Hebrew bible, and I’ve recently been interested in changing that, so I’d be curious to know more about what you’re wondering about.

For the Hebrew part: Can you tell us where this mysterious shin is? For pe and samekh, etc., you may want to check out this teaching grammar. See, for that topic, page 191. Lesson 29 (starting on p. 236) goes over other “meta” Masoretic information.

@disciple BH and this chat are a great place to discuss the text, linguistics, hermeneutics, etc. But if you are interested in more technical discussion such as the nitty gritty of iterating over the text in Python (or any other language) there is a discussion group along those lines just getting started using Slack that might appeal to you.

@JackDouglas, @PaulVargas, @Davïd, @swasheck and anyone else interested in the intersection between coding and Biblical data sets, this might interest you too.

@Susan Thanks for your interest. I have put a sample (mismatched lines from Genesis plus a few lines) in hastebin.com/iqolurawow.lisp . Gen 50:26, line 1532 of my texts, I think line 198 in my sample, has {ש} (I think that's shin, but I'm not always sure of my Hebrew letters). I know the vast majority of my differences are artifacts but need more info to programatically or manually eliminate them.

@Caleb line 34 is probably a case where the manuscript has a smaller letter, so Costas fonted it accordingly. There are some lines differing only by a vav letter which are probably "defective" vs "complete" spelling, but it would take me a long time to verify because I would have to look up every word. I would love to produce an article about variants in Hebrew texts, but preferably it would be written by a Hebrew expert.

@disciple That’s a sin/shin alright, but I have no idea why it’s there. It’s between two lines of repeated text that differ only by two words being stuck together as one in the first line and separate on the second. I’m used to those two words (son-of-hundred-->hundred years old) being stuck together with a maqqef, but you’ve got none of those there. It looks like there are a bunch of them where lines get repeated for that reason. ?

(AFAIK, sin is not used like pe and samekh are for punctuation.)

@Susan Thanks for the quick reply! the format is: <verse number, 1..23000> <book>|<chap>|<verse> <text from wlc> <text from aleppo> These are broken into fields with commas and divided into lines with a pretty-printer function. The rtl / ltr doesn't behave well in some contexts, although hastebin seems to do a nice job, at least according to my browser. I've gotta go, will be back. I have eliminated the vowels and punctuation in wlc. Maybe the shin represents the maqqef in aleppo?

@disciple It's not in the place of the maqqef, though; it's between a repeated line that 'normally' (i.e. in the pointed texts I'm accustomed to) contains a maqqef, her re-written with and without a space. I don't think I exactly understand what you're doing. Are there supposed to be a bunch of repeated (but slightly variant) lines?

I haven't looked closely, but it seems like most of them are because a couple words are rewritten together or apart (where a maqqef would be in a pointed text). At least one I see is a spelling variant (with or without a mater lectionis).

Oh, I get it (sort of) - that's why you're saying they're 'mismatched lines'. And the mismatches are mostly defective spelling (as you pointed out) or word spacing (which I guess is artifact)? Still no idea about the sin, though... Anyway, now I've gotta run, will be back and look closer later. Interesting stuff!

@Susan My program accepts lines as the 'same' when they only differ by spaces (and /xa0, which is a breaking space), but I wanted to retain the spaces when printing. I know that any line still having any ascii characters must have meta-information differences, but I must understand and parse these before I can get clean consonant-only text. I've seen qure/ketiv differences marked with brackets for example; I must figure out how to interpret these programatically.

I decided to paste my code, even though it's very crude as I am still learning the basics of using python efficiently. I will become unavailable in a few days. I would be happy if someone takes my idea and runs with it as long as they eventually share with me. I would be even happier if someone shows me where something similar has already been done. hastebin.com/lulozanipa.py Ideally, we need a good library to deal with thml and/or the wlc's internal format, and more ocr'ed texts.

Some manual interpretation: 34: The "small" ה is an insertion. Possibly a doubtful character in aleppo.

200: wlc has choice of כ or ק insertion just before אתך 203: each has one vav not in the other 276: aleppo inserts a shin שם | ש | ם line 34 is just weird, line 200 has brackets to deal with; others can be found with character diff alone.