On Fri, 6 Jan 2012, Lou Burnard wrote:
> The one thing I hope we all agree would be bad news is
> <p><hi rend="italic">blah blah</hi> bling <hi
> since this encoding does not make significant the real thing of interest in
> the source: the fact that the bling is visually distinct from its
Bad news or not (and I agree that it is), we perpetrate a lot of this
sort of thing -- or rather, our keyers do, and we are apparently helpless
to stop it: partly because they find it easiest simply to <hi>light
everything in italic (a simple rule to apply); but partly because it is
often genuinely difficult to tell what exactly is being highlighted with
respect to what. Certainly there *are* paragraphs like the one above in
which arguably the single bling word should be un-<hi>'d, but the keyers
are apt to go by majority rule, i.e. the predominant typeface is regarded as the
unmarked one, even when the semantics argue otherwise. Layers of nested
<hi>s are even more problematic (an emphasized word in italic within a
differently emphasized phrase in roman within a differently emphasized
paragraph in italic within a predominantly roman div). Interpretive
markup at the word and phrase level is probably our single greatest
struggle, and I often think that we would have done better to jettison
<hi> altogether in favour of a purely descriptive set of <seg>s with
But leaving aside the question of what constitutes markedness, the
general rule surely should be that @rends are to be regarded as inherited
if not overruled by a nested @rend on the same semantic 'axis.'
But that raises the question of how exactly to define those axes, and what
to do with mixed @rend values (<p rend="boldItalic"><hi rend="roman">
or <p rend="italic rightJustify"><hi rend="bold"> etc.) This is a problem
as old as time, or nearly so. I'm not sure that anything has really
changed since the days of <i> <b> etc.
> On 06/01/12 10:34, Franz Fischer wrote:
>>> Consider<p rend="italic">blah blah<emph rend="roman">bling</emph>
>>> blah</p> for example. The intent here surely is that the bling 's
>>> roman-ness replaces its italic-ness; it doesn't complement it.
>> But accurate encoding should then be as follows:
>> <p><hi rend="italic">blah blah<hi> <emph rend="roman">bling</emph> <hi
>> shouldn't it?
>>> On 06/01/12 09:55, Barry Cornelius wrote:
>>>> On Fri, 6 Jan 2012, Sebastian Rahtz wrote:
>>>>> I suspect I know what people will think, but let me try anyway.
>>>>> <tei:hi rend="subscript">b,<tei:hi
>>>>> mean that the inner "i" is subscripted or not?
>>>> Do you need to distnguish between the TEI document and its processing?
>>>> When looking at the TEI, I'm not sure if inheritance is relevant. I mean
>>>> as far as the TEI is concerned the letter i is inside a hi-rend-italic
>>>> that is inside a hi-rend-subscript. That's the end of the story.
>>>> When processing this TEI, e.g., by a XSL script, it's up to your XSL as
>>>> what you want to do with characters that are inside a hi-rend-italic that
>>>> is itself inside a hi-rend-subscript. Most people would want the
>>>> If the opposite of italic is normal then how about:
>>>> <tei:hi rend="normal">b,<tei:hi rend="italic">i</tei:hi></tei:hi>;
>>>> Presumably this is valid TEI but your XSL might produce an error.
Paul Schaffner | [log in to unmask] | http://www.umich.edu/~pfs/
316-C Hatcher Library N, Univ. of Michigan, Ann Arbor MI 48109-1190