This is feedback from Christian-Emil Smith Ore,
c.e.s.ore@iln.uio.no, associate professor at the University of Oslo. Christian-Emil didnât want to go through the motions of signing up for this mailing list etc. but he has given me permission to submit his feedback on his behalf and he has agreed to the OASIS Feedback License.
---
I quickly read
https://www.lexiconista.com/dmlex/It looks ok. Some Âis too concrete. We made an editorial system for Norsk Ordbok (Norwegian dictionary) in 2002 used it tto edit 7 volumes of the 12 volume dictionary (all published in 2015). First we though a xml-system would be best, then we ended up with a relational model.
You are quite right that recursion is problematic in a relational database. We solved the problem by using a root pointer in addition to a pointer to a parent node. which in fact is what you do, The nodes in the tree are what we call sense groups consisting of definition, quotations, literary references (to a separate bibliography), place references (pointers to a separate geo database) Âand pointers to sub entries (better safe than ..). The sub entries are proper entries in the database. All references are stored as separate bidirectional inks as you do. The content of the relational database is also extracted into an xml-document used to store the state of an entry at a given time and to produce printout, web presentations etc.
Each entry in the dictionary is about a single lexical item (Atkins Rundell). Since the Norwegian spelling opens for a lot of variations a lexical item is a set of variants, and we find it useful to open for more than one headword. Each headword is linked to another database with grouped documented usage examples. So the entire system consists of the dictionary proper, a place database, a bibliography and the database with back ground material.
After my quick view on the dmlex, it is one of the best attempt tomake a good model I have seen so far.
Best,
Christian-Emil