On Tue, 2007-06-03 at 13:47 -0500, Joshua Hutchinson wrote:
> On 3/6/07, John W. Kennedy <[log in to unmask]> wrote:
> > If they're going to use TEI, then they should do it properly. Otherwise,
> > let 'em stick with HTML, and save both the temporary labor of converting
> > to TEI and the permanent hardware overhead of converting back again on
> > every use.
> The problem, as I've said in other messages, is that our markup is
> done on a volunteer basis by people who tend to think in layout terms
> vs semantic terms (I admit that the shift in my own thinking still
> comes back to bite me at times ... I don't always think in semantical
But if you see what you're doing as structural transcription rather than
editorial markup, the two are quite close in this case. <i> can be seen
in structural terms as sugar for <hi rend="italic"> and <b> as <hi
rend="bold">. Your example about abusing <p> is a different and deeper
issue, since <p> can't be sugar for <lb> (though I suppose lb could be a
milestone of sorts for p). As you no doubt know it is (?was) very hard
to get from tei:p to html:p when there are intervening elements like
tei:quote/tei:p in the structural paragraph. But since the tei:p maps
better on to print practice--"use tei:p whenever you see an indented
paragraph"--it shouldn't be too hard to train.
Given the issue and feelings involved, I think myself that the most
consistent decision to take on a large project like this might be to
define what you are doing on the transcription~editorial continuum first
and then produce guidelines based on that... and I'd define myself very
close to transcription given the issues involved. You can then explain
the reason for the encoding practice in the way Peter has been
suggesting. In this case, I would say something like "This project's
primary goal is the transcription of print sources in structural markup.
Since our goal is to produce texts that are accurate and useful for
subsequent higher level work by others, our markup will be as neutral as
to semantic meaning as possible: in other words we will concentrate on
describing structural features of the text before us as much as possible
rather than interpreting their semantic value..."
I'd then stick to hi. I don't think I'd go as far as using ab for
undistinguished chunks, though.