RE: Alternate editor's draft of proposed CR draft (posted to web)

Hi,

I think that replacing "binding" for inputs and outputs with "connected
to" everywhere works for me. I really have had problems with
understanding whether they mean the same, or what the difference between
them is.

When I read section 2.2 (Inputs and Outputs) in the new draft, I think
it is much easier to understand now (but it may be just because I know
what the behavior is supposed to be - or at least I think I know :).

So +1 from me.

---

Three small remarks about the new draft:

1. Perhaps the sentence: "If a (non-primary) output port of a compound
step is left unconnected, it produces an empty sequence of documents
from the perspective of its siblings." should emphasize it more that it
applies only *within* the compound step , not when connecting the
compound step to its sibling steps.

2. Also, whether a primary output port is connected *within* a compound
step or not has nothing to do with err:XS0005 ("primary output port not
connected"); perhaps the text should make it more clear, too. Error
err:XS0005 always applies only 'on the outside' of the steps.

3. If we decide to replace "binding" with "connected to" for input and
output ports, are we going to fix the schemas (DTD and XSD) so that they
don't use the term "binding" for inputs and outputs?


Regards,
Vojtech


> I don't think it's intrinsically wrong that that our spec uses
> "binding" for about three or four different things, but I do think it
> invites confusion.
> 
> 1. Connections between inputs and outputs
> 2. Variable and option bindings
> 3. Namespace bindings
> 4. Parameter bindings (Maybe; we never use the phrase, but it could
>    reasonably be inferred as a parallel to variable and option
>    bindings.)
> 
> After some discussion and thought, I came to the conclusion that
> there's no good alternative for "variable binding". I also came to the
> conclusion (perhaps erroneously) that variable and namespace bindings
> are pretty distinct. That just left connections. We call connections
> "bindings" and speak of ports being "bound", but what if we just
> called them "connections" and spoke of them as being "connected".
> 
> I gave it a try and I think it works. What do you think?
> 
>   http://www.w3.org/XML/XProc/docs/alternate/
>

Received on Thursday, 8 October 2009 08:09:40 UTC