Link typing
Atom feeds for Concordia will use atom:link elements to relate web resources to each other.
Relevant link rels built into Atom are: "alternate", "related", "self" RFC4287. The rel type 'alternate' is used in links to alternate representations of an entry in general, and to corresponding web pages in particular. Commonly, these alternate representations will be from the same domain as the entry. The rel type 'related' is a used in links to related web resources. Commonly, these related resources will be in a different domain. An entry about an IAph entity can thereby indicate a generic catch-all relationship to a Pleiades entity and vice versa.
Extension link rels for Concordia come from the vocabulary defined at ConcordiaThesaurus.
example
<feed>
<entry>
<category term="person" href="http://atlantides.org/concordia/terms"/>
<link rel="self" href="http://clas-lgpn2.classics.ox.ac.uk/id/123456/atom"/>
<link
rel="http://concordia.atlantides.org/concordia/terms/belongsTo"
href="http://atlantides.org/batlas/cyrene-38-c1"
/>
...
</entry>
</feed>
example 2
<feed>
<entry>
<category term="inscription" href="http://atlantides.org/concordia/terms"/>
<link rel="self" href="http://ircyr.kcl.ac.uk/inscriptions/C61300.atom"/>
<link rel="alternate" href="http://ircyr.kcl.ac.uk/inscriptions/C61300.html"/>
<link
rel="http://concordia.atlantides.org/concordia/terms/foundAt"
href="http://atlantides.org/batlas/cyrene-38-c1"
/>
<link
rel="http://concordia.atlantides.org/concordia/terms/attestsTo"
href="http://clas-lgpn2.classics.ox.ac.uk/id/123456/atom"
/>
</entry>
</feed>
Feed paging
We anticipate very large numbers of entities, prohibiting single-document feeds. We will implement syndication and archive feeds as described in RFC5005. A syndication feed document containing the N (or fewer) most recent entries will be served as the primary entry point. This document will link to a previous archive document containing the previous N entries. That document will likewise link to its preceding archive document. The concordia reader/crawler will traverse these links to access all entries.
Page size
The value of N shall be up to the feed provider and need not be the same for all pages. Keeping individual page documents under several MB is preferable.
