Re: Simple alternation

I think you're exactly right, David, with one caveat:

TEI is a language for modeling texts, and the nature of the model will
respond to what you're trying to do. So if you care about modeling the
state of the document you're encoding, putting the choice at the level of
the character is the right decision. You might not be worried about the
state of the document though, but instead be focused on linguistic
variation or something like that, and then you might care more about whole
words and less about what was happening in the source.

The software you plan on putting this TEI document through may itself
embody such opinions, and you might tailor your markup to make processing
easier. By doing so, you'd be making the same decision I outlined above,
but maybe without having thought about it, which, as you note, is
sub-optimal :-).

All the best,
Hugh

On Fri, Oct 21, 2016 at 10:18 AM, David Farmer <[email protected]> wrote:

>
> I am relatively new to TEI, but in other work I have encountered
> something similar to the issue of which of these is correct:
>
> <choice>
>    <unclear>m</unclear>
>    <unclear>w</unclear>
> </choice>euyn
>
> OR
>
> <choice>
>    <unclear>meuyn</unclear>
>    <unclear>weuyn</unclear>
> </choice>
>
> The argument for using the second possibility is that you can
> easily find the word 'meuyn'.  I don't think there was anything
> else in its favor, because the first option correctly
> (and minimally) captures the source.
>
> Assuming I understood that reason, it is not good reasoning because
> it suggests that you should use sub-optimal markup because of a
> limitation in the tools you have.
>
> The purpose of the XML source is to capture the structure of the
> document.  The XML source is not necessarily what you use to study
> the document directly: you need to transform it in some way.  If you
> need to look for specific words, then you transform it in a way that
> lets you more easily locate the words.  Applying a simple transformation
> to unclear tags is one of several things you may need to do in order
> to make a new (temporary) document that lets you find all words
> using a simple string search.  But it is just one-time work to set that up.
>
> Apologies if my comments are based on a misunderstanding, or if I am
> saying something that already is obvious to all of you.
>
> David
>
>
>
> On Wed, 19 Oct 2016, Raffaele Viglianti wrote:
>
> On Wed, Oct 19, 2016 at 7:56 AM, Magdalena Turska <[email protected]>
>> wrote:
>>
>> If, alas, the aim of the TEI encoding is to capture that ur-source then,
>> yes, <unclear> might have a place there.
>> Such a reconstruction feels odd though and in my opinion calls for some
>> careful policy and documentation at least.
>>
>>
>> I don't think this is about reconstructing a supposed "ur-source", rather
>> it seems to me that Paul is translating
>> editorial markup from the original document he showed us into TEI markup.
>> The question marks in the original document
>> indicate uncertainty, not variance.
>>
>> I would lean towards an econding that is closer to what I see in the
>> original document:
>>
>> <choice>
>>    <unclear>m</unclear>
>>    <unclear>w</unclear>
>> </choice>euyn
>>
>> but as Poitr has pointed out, it may be more useful to indicate the
>> lemmas separately, depending on what the encoding is
>> for.
>>
>> Raff
>>
>>