See my other note on the folly of putting tags inside attribute values.

The <lb> element has an ED attribute precisely so as to enable you to
specify which edition the line breaking you're recording pertains to.

So why not do <lb ed="orig" n="1"/> at the place where the original line
break was?

As to the regularization you encode as
>         <reg orig="regular-<lb>ization">regularization</reg>

how's about:

         <reg>regular<lb ed="orig" rend="-"/>ization</reg>

(Note the use of the REND attribute to indicate that the linebreak was
originally rendered using a hyphen)


On Tue, 5 Jun 2001, Ingo Mittendorf wrote:

> al.
> TEI is about content, not style, which becomes sort of a problem with
> texts the first edition of which is an electronic one, but with other
> texts perhaps as well.
> In a number of disciplines it is quite common to quote from source texts
> by page and line numbers.
> As the display of a TEI document largely depends on some processing
> software, and may vary considerably, there's a problem here. How can you
> make sure that someone who uses an electronic version of a text is able to
> quote it in the usual way, line numbers included? And every user in the
> same way?
> To "freeze" the line breaks, as it were, does not seem very appealing.
> Does anyone have some experience with this sort of problem, and found
> answers?
> What I can come up with right now myself is to document the original line
> breaks but "neutralize" them by regularizing them away, ie mark them up
> like this:
>         <reg orig="<lb>"></reg>
> Provided this is possible.
> If a word was separated at the end of the line, it could look like this
> (to insure the word is findable):
>         <reg orig="regular-<lb>ization">regularization</reg>
> Better ideas extremely welcome!
> Ingo Mittendorf
> Department of Linguistics
> University of Cambridge
> Sidgwick Avenue
> Cambridge CB3 9DA
> UK