Closed MisterTJB closed 8 years ago
URLs should probably be of the form host/path/identifier
e.g. colenso.org.nz/letters/PrLBrghtn-0001
Suggested browse filters:
creation/date/@when
)correspAction[@type = "sent"]
and //titleStmt/author/name/text()
)//correspAction[@type = "received"]
and //correspAction/name/text()
)Methodology:
Retrieve {document uri, creation date, sender/author, receiver, title} for every document in the database. When filters are selected, use jQuery to show/hide elements in the results list.
Tested execution time for retrieving every document – BaseX reports 45ms and total page load time is approximately 1500ms. Note, though, that rendering only takes about 400ms.
The methodology of retrieving and displaying every document, then filtering via jQuery hiding is probably fine (as opposed to, e.g, pagination or progressive loading)
Alongside each filter, indicate the number of results in the current result set that are associated with that filter. E.g:
Sender Person A (43) Person B (11)
Recipient Person C (12)
Maybe browsing should be interpreted as "browsing" through search results from the latter view? I.e:
Proposed methodology:
Using jQuery:
$.get( '/ajax/letterID',parameters, function(data) { $('#letter').html(data);
Browse should be ordered by date written to preserve temporal continuity
Actually, no, the single page paradigm doesn't fit well with the requirement that page URLs can be emailed.
All users want to browse the documents, and view the letters in a nice human readable way. They should be able to save the raw TEI of a document that they are looking at. If they email the current URL to a colleague both will see the exact same document.