W3C home > Mailing lists > Public > public-xml-processing-model-wg@w3.org > February 2006

Re: The iterator case

From: Richard Tobin <richard@inf.ed.ac.uk>
Date: Thu, 16 Feb 2006 22:48:16 +0000 (GMT)
To: Norman Walsh <Norman.Walsh@Sun.COM>, public-xml-processing-model-wg@w3.org
Message-Id: <20060216224816.E73935A4DD3@macintosh.inf.ed.ac.uk>

One approach would be to allow a component to say that it produced a
sequence of infosets.  This would be done in some kind of component
description file; the API by which it actually outputs the sequence
would be implementation-dependent.  The pipeline compiler would then
be able to know when a sequence output was fed to a "normal" component
that only accepted a single infoset, and insert code to repeatedly run
the second component.  Similarly a component could say that it
accepted a sequence of infosets. 

(This kind of component description file will be necessary anyway to
specify the parameters that a component requires, how many inputs and
outputs it has, and perhaps whether the inputs and outputs are PSVIs
or other augmented infosets.)

-- Richard
Received on Thursday, 16 February 2006 22:48:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:21:47 GMT