TEI Stand-Off Markup WG 2002-08-22 Conference Call Notes

Initials Used for People

  • SB Syd Bauman
  • LB Lou Burnard
  • CC Chris Catton
  • HC Hamish Cunningham
  • DD David G. Durand
  • JH Jessica Hekman
  • NI Nancy M. Ide
  • AI Amy Isard
  • CL Christophe Laprun
  • LR Lloyd Rutledge
  • FV Fabio Vitali

Contents

Commenced 12:10 UTC w/ DD, SB, JH, CL, FV, LB, and CC.

XPointer and TEI Pointer

1.1

Action 1: DD write TEISOW01, a delineation of differences between XPointer and TEI extended pointer mechanism 2002-08-26

what's up at W3C

Originally 1 standard, XPointer: an extension of XPath to include ranges, basically. Full pointing language, had the notion of a scheme, but there was only one: "xpointer:". Recent result is to have the scheme for pointers broken out to separate standard.

Thus there are now three official documents (full xpointer will follow):
  • XPointer Framework (names separated by spaces before colon tried in order, first one that works is used)http://www.w3.org/TR/2002/WD-xptr-framework-20020710/
  • Simple Element-Style Addressing (child sequence numbers or idref) to point directly to a fragment http://www.w3.org/TR/2002/WD-xptr-element-20020710/
  • Utility for declaring a namespace within the pointer attribute that are not declared elsewhere http://www.w3.org/TR/2002/WD-xptr-xmlns-20020710/
One good side effect for TEI: a few things exist in TEI that don't exist in XPointer
  • different (better?) pattern matching
  • foreign()
  • reference systems
If we need to keep them, we may be able to add mechanisms by means of this framework.

LB asks if using scheme mechanism could be used to maintain TEI extended pointers as they are. SB & DD answer ‘yes’.

NIST troubles

NIST is using XPointer in the Atlas project http://www.nist.gov/speech/atlas/, a framework for linguistic annotation. Interchange format is XML based. It possible to reference elements in other corpora, and, e.g., create new annotations for 'em. In order to support that NIST uses XLink to identify elements that can be referenced. Referenceable elements have an id=; pointing them involves an Xlink, "#", the ID (Simplest possible XLink).

The problem is that of URLs: if you create a copy that uses absolute references, you have the full URL before the # and ID, you end up with broken links (which cause parser failure?) Seems should use URNs (or FPIs). Pearls work only via OCLC. Once a corpus is published, it cannot be modified. Thus one that references another that is on the web, its links may break. LB & SB: entities (at least in SGML)

DD notes that TEI is not in a position to resolve the persistent identifiers problem. However, we should create working examples (code snippets & instances) to demonstrate various things; and perhaps this problem belongs in list of things to demo. Should also consider catalogs for XML.

quick review of plans for the next few weeks

Everyone should read (in rough order of importance) chapters
  • 14 Linking, Segmentation, and Alignment http://www.tei-c.org/P4X/SA.html
  • 11 Transcriptions of Speech http://www.tei-c.org/P4X/TS.html
  • 22 Tables, Formulae, and Graphics http://www.tei-c.org/P4X/FT.html
  • 18 Transcription of Primary Sources http://www.tei-c.org/P4X/PH.html
.
Action 2: Everone Read chapters 14, 11, 22, and 18 (URLs above).

Action 3: CL write short description of NIST's problems and post it to list

Today's call kinda long; let's try to get more done on list, less on phone, esp. as expensive for Europeans.

Xlink and its relation to many TEI features

Item tabled until next call.

The issue: XLink and TEI linking elements all have the same kinds of attributes. TEI has historic names.

Action 4: DD with help of JH post some Guidelines' examples reworked in XLink 2002-08-28

LB: What kind of issues do new media impose here? DD: the diffference is that the world of outside formats that we refer to has changed since we started. Practically people would like advice as to which standards to look at. E.g., pointing into a rectangle of a video.

LB: how to extend linking annotation etc for text into time-based or multi-dimensional media. DD: CP's posted a list of URLs on this subject to the list.

Discussion of format of WG's output. DD asks if WG needs to rewrite prose or just point out problems. SB & LB defer answer until we know more about the size of the problem. Note that WG should not feel constrained by the chapter structure of P4 in their discussions, as for P5 we are not married to the same strucutre.