I fully agree, James, of course.

Your stellar reputation as a lover of xml:ids definitely precedes you. You’ve been mentioned several times in this context at the xQuery Institute at Vanderbilt… :) 

All best,

Toma Tasovac
Belgrade Center for Digital Humanities 

On Jun 12, 2014, at 11:32 AM, James Cummings <[log in to unmask]> wrote:

Hi Toma,

Enabling this kind of work, even if done later by someone else,
is why I often recommend providing lots of useful ids for texts.
My preferred granularity is words so I sometimes recommend
automatically adding @xml:id to <w> elements, or indeed
automatically add <w> elements around orthographic words, even
when they are not being used by the current project. This would
allow others to point into the text and comment (or markup)
portions of the text through using these ids.

You are right to point out that checking this kind of thing is a
good application of schematron. (And for those that don't know
you can also define those project specific schematron rules right
in your TEI ODD file.)


On 12/06/14 15:52, Toma Tasovac wrote:

Dear Antonio,

I've used your first option (two separate files). We started with
a modernized spelling version of an 18th-century text (which we
had from the publisher) and were then changing the spelling back
to the original spelling.

So we started with two identical files, automatically created
xml:ids in each (adding a suffix in the id to differentiate
between the old spelling file and new spelling file) and the
@corresp pointing to the exact same word in the other file.

Because old and new spellings sometimes don't follow the same
word segmentations, ids would eventually get out of sync (i.e.
one word in one file would be pointing to two files in the other
or vice versa), but there I wrote Schematron rules to make sure
that pointers are correct in both directions (i.e. to make sure
that an id from one file is pointing to existing id or ids in the
other file). It was up to the encoder to get things right, of
course, but Schematron was helpful underlining cases of
mismatched ids and corresp ids in two files.

All best,

Toma Tasovac
Belgrade Center for Digital Humanities

On Jun 12, 2014, at 7:41 AM, ANTONIO ROJAS CASTRO
<[log in to unmask]> wrote:

Hi Martin,

Both old and modernized spelling "versions" (or transcriptions)
are based on one manuscript - which I call "base manuscript"
because I used it to collate other manuscripts in order to get
substantive variants.

Many thanks

2014-06-12 14:34 GMT+02:00 Martin Holmes <[log in to unmask]
<mailto:[log in to unmask]>>:

Hi Antonio,

Is your base text one of the two versions you mention, or
is it a third text?


On 14-06-12 04:07 AM, ANTONIO ROJAS CASTRO wrote:

Dear TEI list members

I am currently encoding a long poem in Spanish and I
have come across
some difficulties. I have been encoding both an old
spelling version and
a modern spelling version of my base manuscript in
order to present them
as parallel texts. In addition, after collating several
scripts, my aim
is to encode a critical apparatus attached to the
modern spelling version.

I have been reviewing the Guidelines but I failed to
find what should be
the "best" method:

1. I could treat these transcriptions as different
texts and encode

them in different XML files - although I'm using
the same script as
a base text?
2. I could use the <choice> element and combine the
<orig> and <reg>

elements to encode both transcriptions?
3. I could link both transcriptions as parallel texts
using the

<linkGrp> element as described here?

I have seen that some projects follow the first
methodology but I don't
know how they connect both files so I remain suspicious
(browser?); I
would prefer not to follow the second one because I'm
already using the
element <choice> to encode abbreviations and expansions
and I believe
there would be a conflict with the apparatus. Finally
the last method I
think is meant to be used for encoding one text in
different languages
(original and translation) and also it entails I'll
have to identify
each <l> in order to link them, which I began to do and
it is very tiring.

This would be a native digital edition so I don't have
to stick to any
model apart from my base manuscript.

Any advice? Any other method?

Many thanks.

All the best

Antonio Rojas Castro
650 767 335

Antonio Rojas Castro
650 767 335

Dr James Cummings, [log in to unmask]
Academic IT Services, University of Oxford