Re: Simple alternation

Thank you all. It *is* a nice tidy example, isn't it?

I take Lou's point that the 'unclarity' might be said to consist of
a pair of alternative readings, and should therefore contain them;
though I'm not sure that it is entirely wrong to reason that there
are a pair of alternative readings, each of which is to some degree
unclear.

As for the decision (I almost said 'choice) to pair off entire
words rather than letters (as Peter and Lou suggested), that 
was something of a policy decision, made long ago, in full knowledge
that it was less than precise: viz., the decision not to allow strings 
smaller than words within <unclear> -- for reasons very like 
those suggested by Piotr (ease of search, ease of lemmatization, 
ease of tokenizing.)

I appreciate also Lou's point that <choice> implies exclusivity and
obviates the need for @exclude. Thanks for that too.

pfs

On Tue, Oct 18, 2016, at 12:10, Lou Burnard wrote:
> By definition all the children of <choice> are mutually exclusive, so 
> you don't need anything like the second proposal.
> 
> The pedant in me says that your first solution is slightly misleading in 
> that it suggests there are two possible ways of encoding, both of them 
> unclear. But that is not the case: there are two ways of encoding, and 
> they are both equally clear -- it's either a W or an M: no possibility 
> of it being a Z or a hippopotamus. So I would say there's only one thing 
> that's unclear and tag it like this
> 
> <unclear>
> <choice>
> <seg>m</seg>
> <seg>w</seg>
> </choice>
> euyn
> </unclear>
> 
> Or, if this sort of thing happens more than once, I might box clever and 
> treat that "w or m" thing as a special glyph :
> 
> <unclear>
> <g ref="#wOrm"/>
> euyn
> </unclear>
> 
> Of course that means I have to define the pesky w or m thing somewhere, 
> and make up my mind how to represent it by default in my output.
> 
> p.s. I don't buy Magdalena's suggestion of using <app> for this: there's 
> only one witness here.
> 
> On 18/10/16 16:02,
> Paul Schaffner wrote:
> > Given this presentation in an edition:
> >
> > http://www.umich.edu/~pfs/mec/queries/meuynweuyn.jpg
> >
> > (which nicely and succinctly says, 'this word is either "meuyn"
> > or "weuyn" but I can't tell which), what is the simplest TEI
> > way to capture this meaning? (not the appearance, just the meaning.)
> >
> > Does this suffice?
> >
> > <choice>
> >     <unclear>meuyn</unclear>
> >     <unclear>weuyn</unclear>
> > </choice>
> >
> > Or is something like this required?
> >
> > <choice>
> >    <unclear exclude="#weuyn" xml:id="meuyn">meuyn</unclear>
> >    <unclear exclude="#meuyn" xml:id="weuyn">weuyn</unclear>
> > </choice>
> >
> > Or is there another, better, way altogether?
> >
> > pfs
> >
> 
-- 
Paul Schaffner  Digital Library Production Service
[email protected] | http://www.umich.edu/~pfs/