- From: mozer <xmlizer@gmail.com>
- Date: Thu, 2 Apr 2009 09:52:49 +0200
- To: James Fuller <james.fuller.2007@gmail.com>
- Cc: Norman Walsh <ndw@nwalsh.com>, public-xml-processing-model-comments@w3.org
Jim, To be fair, not really The main difference between p:group and p:declare-step is that for p:declare-step you have to EXPLICIT every input and output Which is not the case for p:group (because sibling-components can see each other i/o) For me it makes a lot of differences which means as a user p:group is more user frendly (but probably it makes life harder for implementers) Xmlizer On Thu, Apr 2, 2009 at 8:45 AM, James Fuller <james.fuller.2007@gmail.com> wrote: > I guess my point is that we can declare a p:group in terms of a > reusable p:declare-step e.g. > > <p:declare-step type="my:group"> > <p:output port="result"/> > </p:declare-step> > > wouldn't this achieve the same exact behavior ? > > cheers, Jim > > On Thu, Apr 2, 2009 at 3:26 AM, Norman Walsh <ndw@nwalsh.com> wrote: >> James Fuller <james.fuller.2007@gmail.com> writes: >>> it occurred to me that p:group is almost like p:declare-step and more >>> of a syntax shortcut along the lines of p:pipeline ... should we not >>> frame it in these terms within the spec ? >> >> It really doesn't feel that way to me. The most significant features of >> a p:declare-step to me are the fact that it can declare a type that can >> be called as an atomic step and it can have arbitrar inputs and outputs. >> >> A p:group can't have declared inputs and can't declare a type. >> >> I guess if you think of p:group as a semantic-free wrapper and >> p:declare-step as an extension of p:group that adds inputs and >> semantics, I can sort of see where you're coming from, but it doesn't >> feel natural to me. >> >> Be seeing you, >> norm >> >> -- >> Norman Walsh <ndw@nwalsh.com> | As a general rule, the most successful >> http://nwalsh.com/ | man in life is the man who has the best >> | information.--Benjamin Disraeli >> > >
Received on Thursday, 2 April 2009 07:53:31 UTC