[tei-council] reviewing SA again;what is standoff?
Syd_Bauman at Brown.edu
Wed Oct 11 12:07:46 EDT 2006
> Perhaps there should be a section earlier in chapter 14 to discuss
> this more generally, as well as discussing the use of xinclude
> basically to include boilerplate text, chapters, etc. This is the
> way people are going to want to use this (in preference to
> entities), so I feel that we should at least have a brief
> discussion of it.
I think, and I think Lou agrees, that this discussion would probably
be better situated in chapter 2, "Gentle Introduction to XML".
> However, I don't think anyone would suggest that XInclude become
> part of the TEI -- that the Xinclude namespace be incorporated into
> the TEI. Exactly in the same way that the SVG namespace shouldn't
> be incorporated into the TEI. In both cases we should give
> explanations and examples, and an easy acceptable way for
> customisation to allow it, and indicate that this is the
> recommended method to do this kind of thing.
In general I agree, but this discussion started when people balked at
the thought that you don't need to (and perhaps shouldn't) customize
TEI to use XInclude -- you can (and perhaps should) validate your TEI
document after performing XInclude processing.
> I would suggest that we want to discuss entities as one possible
> (and IMHO now dated) method [for including other bits of text or
> documents], and XInclude at the same time. (As well as any other
I agree, although I'm not really familiar with any other mechanisms
in general use. Is anyone else? (And no, I don't think we should be
discussing using the keywords of your version-control system. :-)
> > identifiers and be happy. Heck, that's what I do!
> Well, one of the things the TEI has claimed in the past is that the
> Guidelines stand as an evolving record of contemporary recommended
> practice. Thus the things we do should be reflected in there if
> central to the purpose of the Guidelines.
Right, so I'll continue the analogy to bare name identifiers. I claim
most people (like me) will generally use bare name identifiers and
not the more complex type of XPointer. And indeed, the use of bare
name identifiers is ubiquitous throughout the Guidelines. But that
doesn't mean the Guidelines should not address the (somewhat rarer)
need for more fine-tuned pointing. So that's what we do in section
Similarly with XInclude: just because there are lots of uses for it
other than for stand-off markup, uses that are certainly more popular
than for stand-off, doesn't mean we shouldn't suggest using it for
> I don't think 14.9 should disappear (otherwise I would have said so
> when I read it and reported earlier), but I do think if it exists
> then the other more common uses of XInclude should be described
> somewhere, possibly in this chapter.
Well it seems we agree, if for different reasons. I see no reason why
mentioning XInclude for stand-off obligates us to discuss it for
other purposes (e.g., chapter management). On the other hand, I
really think that XInclude is a viable alternative for those other
purposes, and thus should be discussed.
More information about the tei-council