W3C home > Mailing lists > Public > xproc-dev@w3.org > September 2008

Re: compound steps and primary inputs

From: Norman Walsh <ndw@nwalsh.com>
Date: Fri, 26 Sep 2008 08:39:10 -0400
To: XProc Dev <xproc-dev@w3.org>
Message-ID: <m2iqsjnksx.fsf@nwalsh.com>
James Garriss <james@garriss.org> writes:

> I understand what a primary input port is.  What I don't know is how I know
> that p:for-each has only one input. 

The single (and hence primary) input to p:for-each is spelled
p:iteration-source (instead of p:input). Similarly, it's spelled
p:viewport-source on viewport and p:xpath-context on p:choose/p:when.

Why, you may ask, did we change the spelling in all those places? Why
not just use p:input there? Well, the answer is, because p:input has
all sorts of attributes (port, sequence, primary, kind) that don't
apply in those contexts. For a long time, we tried to make p:input
work in those places but it was hopelessly confusing.

Our decision to use a different name obviously introduces some
confusion too, but I'm pretty confident it's less confusion than if I
had to explain why some attributes were allowed in some places and not
others and why some attributes could only have fixed values in some
places and not others.

I hope that helps.

                                        Be seeing you,

Norman Walsh <ndw@nwalsh.com> | There might very well be nothing; nor
http://nwalsh.com/            | anyone. No one to notice that there is
                              | nothing, and to consider that natural.
                              | But that there is something, and,
                              | whatever it may be, the strange thing!
                              | I shall never cease being amazed at
                              | this.--André Gide

Received on Friday, 26 September 2008 12:39:53 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:03:03 UTC