Marking up unitalicized titles?

[c-forster]

In "Telemachus" there are works mentioned that are not currently marked up, because they are not italicized in the print edition, and so were not marked up with <emph> tags. For instance:

Can you recall, brother, is mother Grogan's tea and water pot spoken of
<lb n="010371"/>in the Mabinogion or is it in the Upanishads?</said>

I marked this up as follows in #17196501040 (I caught it incidentally and only remembered after the PR; apologies for confounding two issues in one PR):

Can you recall, brother, is mother Grogan's tea and water pot spoken of
<lb n="010371"/>in the <title rend="none">Mabinogion</title> or is it in the <title rend="none">Upanishads</title>?</said>

Is that a reasonable way to handle this? There does not yet seem to be a title type applicable to these works. There are also references to Hamlet that might be applicable here; though it is conceivable that the reference is to the character, I think the following should also be marked up if we're marking up typographically unmarked titles:

<lb n="010487"/><said who="Buck Mulligan">―Wait till you hear him on Hamlet, Haines.</said>
[...]
<lb n="010545"/><said who="Haines">―What is your idea of Hamlet?</said> Haines asked Stephen.

[yellwork]

Good call. That’s consistent with what we’ve been proposing for unmarked instances of <foreign> and <name>. There will be others, I’m sure.

It might be wishful thinking if we imagine that we’re going to catch everything, but we should certainly (eventually) aim to run all marked instances of a given <title> or <name> or whatever against the rest of the corpus to turn up unmarked instances. These should be encoded with the @rend="none" attribute. So (per #10) we can run <name type="horse">Throwaway</name>, for example, against the rest of the corpus to find all unmarked instances of the horse’s name appearing.

[JonathanReeve]

Since this issue looks settled, I've just added this convention to CONTRIBUTING.md (see 953a5e1#diff-6a3371457528722a734f3c51d9238c13R55), so that we can close this issue. Feel free to reopen later, though, if edge cases come up.