How are Sifrei Torah typically checked today?

Question

Our shul has about 10 Sifrei Torah (no exaggeration) that are all passul. I am curious how complicated it would be to discover where the problems are.

Someone told me that today, many sofrim (scribes) use computers to discover where errors are. Is this true? If so, how does the process work? Do they put the Torah scroll on a flat-bed scanner or do they wave a wand scanner and then compare each column with some standard? How would that work, since everyone's writing style is a bit different and also, the column starts are endings in each Torah may be different. (Note: for this answer, I am seeking a somewhat "computer / technical" answer, if possible. Is programming built into the system so that it is smart enough to recognize font nuances and compare it accurately with the computer's own font "standard"? Are there settings for, e.g. Sefardi writing styles?) Is it possible that the computer may declare an error when there really is none b/c it cannot recognize font nuances in good detail?

If the process is done manually, how long does it take, typically for a sofer to inspect an entire Torah scroll? I would imagine that this is a very laborious process if one doesn't know where the mistake(s) is / are.

Approx. what percentage of sofrim use a computer to assist them vs. completely manually inspecting?

Lastly, what does it typically cost, in the U.S. to have a sofer inspect an entire Torah and repair it (assuming these are minor repairs such as correcting a letter here and there vs. replacing an entire parchment segment.

Answer 1

Computers are used quite effectively for checking for errors in STa"M. Your understanding of how it works is basically correct. The scanned text is compared with a standard which is known to be kosher. The program would have a variety of "standard" versions for different writing styles and column formats.

To correct for differences in handwriting, a tolerance for difference is built into the program. This can be set to a low value in order to tip the balance of error toward type I errors. This technology is very reliable and has been around for a long time. Torah scrolls are even easier than usual handwriting because the font is so standardized. In general, you don't really have to worry about the tolerance being so high that mistakes would be missed. I imagine that the most common mistakes are entire missing letters/words or letters touching each other. An entire missing letter would easily be spotted and whether two letters touch each other is an objectively true or false question... exactly the kind of question that computers are good at answering. (For Torah scrolls that are already "in the wild," probably the incidence of scratched letters or holes in the parchment are more common than they are for brand new scrolls. Still, the program's tolerance is low enough that it should be able to detect such problems.)

(Source: I am a software developer and generally familiar with the capabilities of computers.)

After running the scroll through the program, the computer will spit out a list of what it thinks are problems with the scroll (even including things that don't make the Torah pasul, but rather simply less mehudar). The sofer will then look over this list and confirm the issues. If the tolerance for error is set to a very low level, there might be some "issues" that aren't really issues at all. The sofer can simply leave those alone and fix the real problems.

As a word of advice, I know a rabbi who is a sofer. He advised me that I should never buy any klaf that has not been computer-checked. Computers don't get tired after a long day of staring at tiny font and miss the fact that whole words are missing. You don't want to be batel the mitzvah of wearing tefillin every single day because the sofer who checked them missed an issue.