Re: I'm _still_ confused about 3.8 Extension elements

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

So my tentative feeling about this is that the problems in the current
draft arise mostly because we adopted the XSLT model too
uncritically.  It only makes sense to make the "has no [semantic]
impact" kind of statement that XSLT does because the control flow of
XSLT is pattern directed.  Our control flow is (in part) linear, so
adding an extension element in a subpipeline just doesn't make any
sense -- it's _extremely_ unlikely for it to be of any use there
unless it _does_ affect the flow of results between the steps on
either side.

So, my conclusion is that for pure extension elements (i.e. not
user-defined steps, which are covered already, or Vnext language
extensions, which we now have a proposal for) the right thing to do is
restrict them to p:pipeline-library and _the prologue_ of p:pipeline,
that is, the part _before_ the subpipeline.

That means the subpipeline rule should be moved to 2.1 and changed to read:

   Elements in a subpipeline are interpreted as follows:

      1. In XProc namespace?
            1. p:documentation? Ignore.
            2. Names a built-in compound step? Check against grammar,
               interpret per spec.
            3. Names a built-in atomic step? Check against grammar and
               built-in declaration, interpret per spec.
            4. Otherwise, error.
      2. Names a declared step type? Check against grammar and supplied
         step declaration, interpret per spec.
      3. Names a defined pipeline? Check against pipeline definition,
         interpret per spec.
      4. Otherwise, error.

The 'subpipeline' pseudo-production in 2.1 should have ipfx:ignored
removed.

ht
- -- 
 Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
                     Half-time member of W3C Team
    2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
            Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                   URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFHR/7HkjnJixAXWBoRAg0xAJ9JXf4Nnd4KHzh9F6vF1OHjyvp0ywCcDc+t
52ZxW+3+7wbbysQOyL2Yugk=
=59s2
-----END PGP SIGNATURE-----

Received on Saturday, 24 November 2007 10:37:03 UTC