Integrating Lotte Kéry's volume

Now that Lotte's volume is online and we have more and more articles with page-specific links to the digital version of her Canonical Collections, we can think about (semi-automatically) integrate information on manuscripts and collections into the Wiki. As Linda used Lotte's book for her own handbok, for most collections this is not really necessary; but the case is different with manuscripts.

1. Identify mss that have an entry in Canonical Collections but not this Wiki.
2. Extract basic information: Place/Archive/Shelf mark, rough date, canonical collection(s) contained in this manuscript
3. Auto-create new pages based on this information.

These new articles would be stubs containing ...

• Page name is "Place, Archive, Shelfmark.
• Categories are Category:Manuscript, Category:Manuscript of ... (see [[1]]), Category:Manuscript saec ...,
• DEFAULTSORT, see here and here.
• an Infobox.

I have tried whether ChatGPT 3.5 can do some of the mechanical work involved in turning Lotte's book into articles. My prompt was

compile a table that contains in every line the following information: place name in colum 1 , library name in colum 1, shelf mark in colum 3, date in colum 4 using the following information on three mansucripts: "Bamberg, Staatsbibliothek, Msc.Can.5 is complete copy of the Collectio Anselmo dedicata; base manuscript for the analysis of this collection in Clavis database. It was produced in Italy in the tenth century but later given to Bamberg by Henry II. Barcelona, Archivo General de la Corona de Aragón, San Cugat 63 is a twelfth century manuscript. Orléans, BM, 221 is a copy if the Hibernensis from 1213."

and later I inserted OCRed passages from Lotte's book. This all went very well; country, city name, library name, and shelf mark were extracted with 100% accuracy, even if the proper names contained several OCR errors and if some of the information was somewhat hidden (for example, in one case the shelf mark was only given in form of the sort order command DEFAULTSORT). The only thing that was more difficult (but still doable) was to extract terminus post quem and terminus ante quem from the texts I fed the bot with. So I am confident that we can compile tables containing basic information on the mss, hopefully also the "is copy of X" information without too much manual work. --Christof Rolker (talk) 12:51, 27 February 2024 (CET)

also, I have talked to WikiData people (Lucy Patterson and Jonathan Groß) about "our" (and other) mss in Wikidata; the notability issues I was worried about are less of an issue, having entries for all our collections and perhaps even all of Lotte's mss seems feasible. --Christof Rolker (talk) 14:35, 27 February 2024 (CET)

If we integrate Lotte's book, it would be good if this could be done in a similar way to the Clavis handbook, i.e. via her account (or a separate account clearly linked to her). --Christof Rolker (talk) 10:30, 3 May 2024 (CEST)

Yes, good idea. The safest way to achieve this, might be to create a special account for this purpose and do all the steps for this integration from this account. Afterwards, I can take all contributions of this special account and re-attribute it to Lotte Kéry's real account. Another, equally viable approach: We work under our own names, but prepend all comments that deal with this integration with a fixed prefix (e.g., "HMCL" or whatever works for us). Then I can identify all those contributions and re-attribute them to Kéry's account. --Clemens Radl (talk) 11:05, 3 May 2024 (CEST)

Lotte Kéry kindly agreed to the former arrangement, and I have set up the new account User:Lotte Kéry which can be used to integrate her book (in addition to her personal account User:LKery). --Christof Rolker (talk) 12:25, 14 May 2024 (CEST)