Project Update
Christopher Guy Yocum
ORCID: 0000-0002-7241-3264
I have been thinking about various aspects of this project for a while. As you may have noticed, it went on hiatus for the last nine months due to various life events as often happens. This also gave me some time to think about various aspects of the project. I have come to a few “executive decisions”.
First, in line with the Data on the Web Best Practices Recommendation 8.6, I will be versioning the data from now on. The way I will do this is to use git tags to tag each merge into the master branch. The format will be the ISO 8601 date plus a dot then the number of the merge that day starting at 1 and incremented from there. For instance, if I were to release for the first time today, the tag would be 20211016.1 and the second would be 20211016.2 and so on.
This versioning scheme will allow anyone who wishes to reference IrishGen to show where in the history of IrishGen they got their information. This is in contrast to needing to use “Accessed on” or the bare git commit hash.
Second, I will be removing the ‘#’ from many of the entries and
putting that in the @base
for each file. This is to cut down on
repetition and visual “noise” when looking at the files. I may in
the future swap from using .trig#
to using /
(the hash versus
slash debate). By centralising the #
, it will be easier to swap in
the future.
Third, I will be auditing all the files. This for the sake of accuracy and as my understanding of the medieval Irish genealogies grows, the audit will reflect that growth. This may never actually be finished as I am doing the audit on my own in whatever free time I feel like devoting to it.
Fourth, as I am auditing, I will be adding Agrelon predicates to the entries. This is to allow people to use Agrelon rather than the Relationship Ontology. I may in the future remove the Relationship Ontology completely. Some of the reasons for doing this are outlined in this post.
Fifth, I am considering removing Named Graphs from the database. Named Graphs did not, in the end, really serve the namespacing purpose that I wanted to use them for. Additionally, I feel that anyone using this database will want to interrogate particular manuscript sources rather than all the manuscripts mashed together in one search. I thought that Named Graphs would serve the purpose of namespacing the manuscripts and the triples contained within but it seems to have caused more confusion and frustration than clarity. Strict namespacing seems to be better controlled at the repository level within whatever RDF graph database that you are using.
Sixth, I am also considering purchasing a domain name for the database. This will allow other people to reference the database in various ways. The cost is fairly minimal at around £10 per year so it should not be too much of a burden.
The project will continue and will be trundling along as my energy levels allow. If you use IrishGen, please let me know. I am always happy to hear from anyone using it or interested in it.