More xrefs with malformed or nonlocalized book abbreviations

[DavidHaslam]

The following book abbreviations do not match those specified by the 66 \toc3 markers:

1 Cor
1 Kho
1 King.
1 Pet
1 Tes.
1 Tim
1. Pet.
1. Sam.
2Chr
2 Chr.
2 Joha
2 Kho
2 Kin
2 Pet
2 Tes.
2 Tim
2. Tim.
Deut
Dute
Eks
Ex
Exo.
Ezek
Ezek.
Ezr.
Fil .
Filim
Gal
Gen
Hagg.
Heb
Hez
Hos
I sai.
Is.
Isai
Isai.
Isai;
Jak
Jas.
Jer
J obe
Job
Joel
Joha
L evi
Lev
Lev.
Matt
Mic
Mic.
Micah
Mik.
Nahum
Num
Num.
Obad
Pro.
Prov
Ps
Ps.
Rom
Rute
Seb
Song.
Tit
Titus
Tshu .

Notes:

1. I have retained Hos and Hagg. even though some of those locations were reported before.
2. Some localized abbreviations are missing the full-stop.
3. Some abbreviations are still non-localized.
4. Some are malformed in other ways as well.
5. Several of the items occur in multiple locations. There are about 218 locations in total.

[DavidHaslam]

The attached tab delimited text file is a tentative replacement list.

replace_bad_abbreviations.tab.txt

Notes:

1. Replacements should be restricted to within xref elements.
2. There is a trailing space in each item to tighten the requirements.
3. Some trivial pre-processing fixes are also required.

I have designed and tested a bespoke TextPipe filter to perform the fixes.

[teusbenschop]

Whew, quite a list of abbreviations that don't match with the \toc3 markers. I have found a few of them occasionally but had not through there were such a big list of them. It is also great that you already have a filter to perform the fix. If you won't mind to clone it and fix it, I can then apply the pull request. Thanks!

[DavidHaslam]

Please commit your recent changes first, before I refork and clone the repo. Thanks.

[DavidHaslam]

I see that you already did so 2 days ago. I hadn't checked since earlier on Friday.

[teusbenschop]

Yup, that correct. There's no uncommitted changes at present.

[DavidHaslam]

It's become more complicated than I outlined yesterday. As well as making more corrections, I'm currently developing the filter further to normalise as much of the xref punctuation as might be possible.

However, I've already come across many instances where reference text is impossible to parse readily.

Here's an example to illustrate the problem: Exodus 28:38 reads: \v 38 Izakuba sebunzini likaAroni, ukuze uAroni athwale ububi bezinto ezingcwele\x + Levi 10.17. Nani 18.1. Levi 22.9. Isa. 53.11. Hez. 4.4,5,6. Joha. 1.29. Heb. 9.28. 1 Pet. 2.24.28.43.\x*, abantwana bakoIsrayeli abazazingcwelisa, kuzo zonke izipho zabo ezingcwele; njalo izahlala isebunzini lakhe, ukuze bemukeleke phambi kweN\nd kosi\nd*\x + Levi 1.4; 22.27; 23.11; Isa. 56.7.\x*.

The problem reference text is 1 Pet. 2.24.28.43.

Now 1 Peter 2 contains only 25 verses, so what to make of the .28.43. ? Clearly this cannot be part of 1 Peter, as the book has only 5 chapters. This means that it would not make sense to change it to 1 Pet. 2.24; 28.43. The only possible solution is to assume that the 28.43. is a reference to Exodus 28:43. Yet that prompts the question, "How to make that clear?".

The reference system currently employed lacks these generalised abbreviations:

• ch. for a chapter in the current book.
• v. for a single verse in the current chapter.
• vv. for multiple verses in the current chapter.

Here I'm using English words for chapter and verse, but these would then need to be localized. The reference 1 Pet. 2.24; ch. 28.43. would then be feasible to parse.

There are so many references that are just as difficult, that I fear the fixes cannot be automated.

[teusbenschop]

The example given is proof that automatic fixing wont' be feasable. As far as I remember, the way to indicated that it's Exodus 28:43 is this: 1 Pet. 2.24. 28.43. The full stop indicates the end of the 1 Pet. series. So any next xref entered will fall back to the current book.

[DavidHaslam]

I will refrain attempting to fix anything more than some elementary punctuation discrepancies.

Those that require assessment of multiples of dot separated numbers are too difficult to address systematically, even though I think some of the dots must be typographical mistakes.

Refer to issue #35 for further details.

[DavidHaslam]

I see that your recent commit was for one non-localized abbreviation that I'd noted. Psalm 56:8 had an xref with2 Kin 20.5.

[teusbenschop]

Yup, I had done a bit of work, then thought that it's better to focus on Shona to complete that first.

[DavidHaslam]

The problem with your rule for the interpretation of the full-stop is that it may not work everywhere.

Consider this example: \x + 2 Lan. 29.34. Hlab. 11.2. 32.11. 36.10. 64.10. 94.15. 97.11. Hlab. 125.4. Hlab. 37.14.\x*

On your theory, the full-stop that terminates Hlab. 11.2. ought to indicate the end of the Hlab. series. It's fairly obvious that it doesn't do so necessarily.

The context for this example being Psalm 7:10, it's simply fortuitous that following your rule here makes no actual difference. i.e. that 32.11. 36.10. 64.10. 94.15. 97.11. falls back to the current book which happens also to be Hlab..

We need a counter example to illustrate where the rule doesn't work at all.

[teusbenschop]

Yes, I agree. Sharp observation.

[DavidHaslam]

All the aberrant abbreviations were replaced yesterday in the Editing branch of my fork. My bespoke TextPipe filter also corrected a number of other miscellaneous issues in various xrefs.