W3C home > Mailing lists > Public > www-forms-editor@w3.org > September 2002

XForms comments

From: <paul@prescod.net>
Date: Thu, 5 Sep 2002 15:02:00 -0700 (PDT)
To: <www-forms-editor@w3.org>
Message-ID: <Pine.LNX.4.30.0209051400500.11201-100000@latte.ActiveState.com>

I know how hard the XForms group has worked on the technical side of
XForms and it might show through better if the specfiication were
organized in a way that made XForms a little easier to learn. Perhaps
third parties will step in with books and so forth, but it would still be
nice if the specification itself were pedagogical as well as normative.

The tutorial section is useful but IMO it needs to be paired with an
"overview" section that outlines all of the major concepts. Is there any
section that describes how everything fits together at a high level?
Chapter two doesn't introduce important concepts like "module", "event",
"exception", "calculation", "actions", "repetition", etc. XForms is
complicated enough that I feel I need the moral equivalent of a high level
design document (perhaps without UML).

When I read section 3, I feel like I've been thrown in feet first. The
chapter is called "document structure" which to me suggests that it would
say something along the lines of "here is how an XForms-containing
document is structured. It has a model. It has some bindings, etc. etc."
But instead we do attributes first which I find a little bit disorienting.
I can't really understand what the attributes are *for* until I understand
what elements are available. I'd reorder that stuff. MustUnderstand and
Extension are things I would be happy to learn about later rather than

Other comments:

XForms says that exceptions are more serious and "fatal" than errors.
This is the reverse of the terminology in many other languages, most
notably Java.

Is there something wrong with the HTML rendition of this page:


I don't see any method names.

This paragraph has two XPaths: "The context node for outermost binding
elements is the top level element node, or the single node returned by /*.
A binding element is any element that is explicitly allowed to have a
binding expression attribute. A binding element is "outermost" when the
node-set returned by the XPath expression ancestor::* includes no binding
element nodes." If I'm not mistaken, one is an XPath into the instance
data. The other is an XPath into the model(?).

"The context node for non-outermost binding elements is the first node of
the binding expression of the immediately enclosing element." -- should
that be the "first node of the node-set returned by the binding
expression..." Also, is this "first node" stuff valid in the face of

7.4 typo: "In particular, there restrictions"

I think it would be clearer to use absolute XPaths for the top-level form
elements in examples. e.g. /bookmarks/section.

 Paul Prescod
Received on Thursday, 5 September 2002 18:02:32 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:25:04 UTC