search

MedCyprus / MedCyprusEFES

The Digital Corpus of Painted Greek Inscriptions from Medieval Cyprus (10th – 13th centuries AD), MedCyprus for short, employs EFES (EpiDoc Front End Services), for the publication and wide dissemination of this largely unknown body of epigraphic material.
Apache License 2.0
3 stars 1 forks source link

Index of Ligatures #1

Open emylonas opened 8 months ago

emylonas commented 8 months ago
  1. pulling from wrong attribute value (clarify)
  2. Future: ligatures across words --Do this first
  3. Future: accented letters (need explanation, example)
emylonas commented 8 months ago

043.17 - Μήτηρ shouldn't be the ligature. It's ΜΗΡ. check

TheocharisPetrou commented 8 months ago

I agree. I checked the mark up of all the records and I think I marked up them correct. Perhaps the problem has to do with the fact that ΜΗΡ is ligature and at the same time it is also an abbreviation: Μή(τη)ρ / Μῆ(τε)ρ.

gabrielbodard commented 1 month ago

In other words, each entry in the ligature index is triggered by one of the following occurances:

The other issue is, should we leave out any non-diplomatic elements (e.g. <ex>, <supplied>, <corr>) from the lemma in this index?

emylonas commented 1 month ago

non-diplomatic elements like <ex> and <corr> won't appear, but <supplied reason="lost"> should show up in the index.

gabrielbodard commented 1 month ago

Correct. However…

For the particular of the encoding of ligatures, our practice needs to be that if two parts of a ligature are divided in the edition, by any markup, spaces, restored/expanded/corrected characters, or anything else, it should be tagged as two <hi> elements, with a <join>. (As in my second example above.)

gabrielbodard commented 1 month ago

See e.g. 044.001: markup is <hi rend="ligature" xml:id="lig1"><choice><orig>Η</orig><reg>Ἡ</reg></choice></hi> <addName type="epithet" key="Pantanassa"><hi rend="ligature" xml:id="lig2">Π</hi><join target="#lig1 #lig2"/> (CORRECT), but H and Π appear as two separate ligatures in the index (and in the EpiDoc display, NB).

However in 060.001 the markup is incorrect, so not a good diagnostic of broken index. To check with Theo and Dora.

gabrielbodard commented 1 month ago

I have fixed the incorrect markup in 043,017 to read: <expan><abbr rend="supraline"><hi rend="ligature">Μ<choice><orig>η</orig><reg>ή</reg></choice></hi></abbr><ex>τη</ex><abbr rend="supraline"><hi rend="ligature">ρ</hi></abbr></expan> which should now be correct. We can test the index against this now.

TheocharisPetrou commented 1 month ago

Gabby, with the above updated markup the whole ligature ΜΗΡ of the abbreviation Μή(τη)ρ was separated in two upper ligature lines instead of one as in other examples we' ve done, e.g. 043.031

gabrielbodard commented 1 week ago

@emylonas Can you see if you can fix in (a) the XSLT and CSS and (b) the index, the issue of ligatures broken across words, such as two examples in line 5 of 044.053 (<hi rend="ligature" xml:id="lig1">ν</hi> <hi rend="ligature" xml:id="lig2">μ</hi><join target="#lig1 #lig2"/> and hi rend="ligature" xml:id="lig3">ὰ</hi> <hi rend="ligature" xml:id="lig4">π</hi><join target="#lig3 #lig4"/>), which appear in diplomatic as separate loops above N, M, A and Π. They are also indexed as four separate, single-character ligatures.

This is likely to be hard, so if not fixable, (1) we might talk to Irene; (2) we might have to tolerate this bug (in the latter case we can still fix the index display for examples of <ex> inside <hi>, as above, which is a rarer instance of this problem).