medieval-mss icon indicating copy to clipboard operation
medieval-mss copied to clipboard

updates and url stability

Open holfordm opened this issue 7 years ago • 5 comments

@ahankinson , what is the best way to deal with updates to authority files that impact on existing URLS? For example, https://medieval.bodleian.ox.ac.uk/catalog/work_14680 and https://medieval.bodleian.ox.ac.uk/catalog/work_14679 are the same work and there should be only one entry. If work_14679 becomes the canonical URL, should we (can we?) set up a redirect from work_14680 to work_14679, and (if so) how should that be expressed in works.xml?

holfordm avatar Sep 04 '17 08:09 holfordm

We would have to do that manually on a case-by-case basis, so it's not very scalable.

ahankinson avatar Sep 04 '17 09:09 ahankinson

I envisage changes to works.xml being made gradually and manually, as they are noticed (together with associated changes to the manuscript records of course) - similarly with duplicates in persons.xml. I don't think there is an enormous number of duplicates, but inevitably there will be some.

holfordm avatar Sep 04 '17 09:09 holfordm

You mean the redirects would have to be set up manually? I was hoping it might be possible to set up redirects with appropriate markup in works.xml (something like <note type="redirect" target="https://medieval.bodleian.ox.ac.uk/catalog/work_14680">Merged into https://medieval.bodleian.ox.ac.uk/catalog/work_14680 and deprecated</note>)

holfordm avatar Sep 04 '17 10:09 holfordm

Do you mean you want to just provide a link pointing from the deprecated work to the 'real' one? Or you want it to automatically re-route the user?

The first is possible, but it will mean that the deprecated records will appear on the Source and Person records, so you would get duplicate records in those lists.

The second is considerably harder (if not impossible). Re-directing and rewriting URLs is handled by the server, and there is no way to 'tell' the server to forward a user to the new URL from the TEI, since it's just content in a Solr index.

ahankinson avatar Sep 11 '17 09:09 ahankinson

option 1 would be fine were it not that the deprecated records would appear in the index. Is there any way of getting round this? (e.g. putting deprecated records in a separate directory and indexing them separately? would that help with implementing option 2?)

holfordm avatar Sep 13 '17 09:09 holfordm