On May 5, 2007, at 10:47 PM, John Boyer wrote:
Maciej,
I think I understand now why we have been in disagreement.
You articulated that using something (XForms) as the basis for
review pretty much means adopting the text of the thing. And that
adopting both WF2 and XForms as the bases for review means
something along the lines of "concatenating" or merging them.
This is very literal and not at all what "basis for review" means
to me. A document to be used as a basis for review is one that
will be carefully scrutinized for content that should be used to
influence the creation of something new.
My understanding of "basis for review" in this case is:
- A copy of the text that become a basis for review will be checked in
- This text becomes a W3C Editor's Draft
- The text is modified by the editors in response to review comments
from the group and others
- Once sufficient review has been done, the text is published as a
W3C First Public Working Draft
Note, however, that the text being checked in and edited further does
not imply a committed decision to anything in the text.
This is what the Apple/Mozilla/Opera proposal on which Dan based his
proposed resolution clearly stated. Although it's true that the
resolution itself is less clear on this point.
I would appreciate if Dan could clarify whether indeed it is expected
that the review text will be checked in and edited as the basis for a
future FPWD.
Parts of the new thing *may* (in the RFC 2119 sense) look very
similar to parts of the documents used as the bases of review, but
they may also not. In this case, using XForms and WF2 as the bases
for HTML5 Forms means (to me) that in cases where the two took
different paths to satisfy the same requirement, then the
underlying requirements which caused the divergence would be
examined to pick the best way forward, and if at all possible that
way would include the ability to map the solution onto the XForms
architecture.
This would imply that we'd start editing with a document that had an
empty Forms section. It also seems like a request to rehash every Web
Forms 2 feature from scratch, despite the existing design and
implementation experience with WF2, and even if there is no technical
objection to a given feature as such. I would find that unacceptable
as it would cause needless delay and I do not see how it would
improve the quality of the finished product. Instead I'd like to
focus on changing the parts of WF2 that do have some technical
objection.
XForms is really a misnomer since it is a Turing-complete XML data
processing language. Therefore, alignment with XForms architecture
is theoretically possible as long as the chosen solution is
computable. However, the semantics of "alignment" to XForms
includes a degree of practicality (e.g. superlinearity of
transformation is not really practical).
I'm not sure what this means exactly.
Anyway, given that my definition has a lot to do with your #2, it
seems that we are in violent agreement on how to proceed with
XForms, quite apart from the difference in how we understand the
words "basis for review" and its implications for what we think
will be happening with WF2 as a result of this questionnaire.
It sounds like we now agree (roughly) on what should happen with
XForms, but not on what should happen with WF2. Do you have a
compromise to propose?
I think it would now be unreasonable to complain further about the
author of our charters for any ambiguities that have arisen given
how this has been going.
And finally, it does seem that Dan should be made aware that there
is a fairly serious divergence of understanding on the meaning of
"basis for review" which affects the answer to the questionnaire.
Indeed. I hope he can clarify.
John M. Boyer, Ph.D.
STSM: Lotus Forms Architect and Researcher
Chair, W3C Forms Working Group
Workplace, Portal and Collaboration Software
IBM Victoria Software Lab
E-Mail: [EMAIL PROTECTED]
Blog: http://www.ibm.com/developerworks/blogs/page/JohnBoyer
Maciej Stachowiak <[EMAIL PROTECTED]>
05/05/2007 05:24 PM
To
John Boyer/CanWest/[EMAIL PROTECTED]
cc
[email protected]
Subject
Re: [off] Re: Forms Task Force Charter Requirement
On May 5, 2007, at 4:41 PM, John Boyer wrote:
>
> Maciej,
>
> mjs> You've stated that you are unwilling to drop your Formal
> Objection
> unless you get exactly what you asked for. Since I disagree that
what
> you asked for is good, if you are unwilling to budge, I don't see
how
> compromise is possible. Unless you think compromise consists of me
> just agreeing with your existing position.
>
> I think what you want is that WF2 be used as *the* basis for
> review. What I want is that XForms also be
> used as *a* basis for review along with WF2.
I don't think it makes sense to adopt XForms as a basis for review
and further editing for the following reasons:
- It is a W3C standard being actively maintained by another working
group.
- It defines a wholly separate language in a different namespace.
- It defines elements and attributes which, if placed in the HTML
namespace, would be incompatible with existing HTML elements with the
same names.
For these reasons, it cannot as-is be a basis for anything in HTML.
It would be like adopting SVG or MathML as a basis for review - even
if we wanted to add graphics or math capabilities to HTML, it would
not be right to start by appending those specs to the end of the HTML
spec. Instead, selective features would be moved and recast as parts
of HTML, instead of separate languages.
We also know based on past experience that XForms is unacceptable to
browser implementors. However, the XForms working group declined so
far to take that feedback into account in the XForms spec. Adopting
something as a starting point that is known to be unacceptable to
many of the likely high-volume implementors seems like an
unproductive way to proceed, as we'd have lots of work to do to
remove the problems before we had anything publishable.
My original position was that the HTML WG should completely ignore
XForms and the Forms Task Force. However, I have moved to a
compromise position that I think is a reasonable middle ground, as
described below.
Possible sound ways to proceed to include more XForms capabilities in
HTLM Forms would be:
1) Adopt as a basis for review something based on XForms but that is
rewritten to be a set of HTML elements, and with incompatibilities
with classic HTML removed. However, no one has written a document
like that, and we don't have time to wait for it to be written. So I
don't think this is a practical option.
2) Consider XForms as a potential source of designs for additions to
HTML Forms, without literally adopting the text. WF2 already did this
to a large extent, but I am seeing it happening even more, such as
with the review and likely revision of the repetition feature.
3) Define more clearly what we mean by architectural consistency, so
regardless of the starting point we know that the endpoint satisfies
the requirements.
I expect #2 will happen regardless of official decision by the group,
but I would not object to making it a formal resolution of the group.
I already proposed a way to do #3. I think these are reasonable
middle grounds between adopting the text of XForms and completely
ignoring it. Do you have other possible compromise proposals?
Your other Formal Objection was to Dave and Ian as editors, and your
proposed alternative is to add someone from the Forms WG as a co-
editor for the Forms section. I also disagree with that, as I think
it would lead to a lot of arguing and slow down work; and because I
think editors should be chosen based on their qualifications and
experience, not based on being members of an external Working Group.
My proposed compromise for that is that the HTML WG and Forms WG
together in the Forms Task Force co-edit a Forms Architectural
Consistency Requirements document which XForms and HTML Forms both
then satisfy. I think this will accomplish the desire for closer
alignment with less likelihood of conflict delaying progress. Do you
have a different proposed compromise?
> Your writings imply you believe I am being rigid and exclusionary;
> this makes no sense given that what
> I want includes what you want and what you want excludes what I
want.
I think what you want is impractical - concatenating WF2 and XForms
does not result in a single specification that we can edit further.
However, XForms should definitely be mined as a source for feature
ideas that can be cast into a more HTML-friendly form.
> You can feel free to publish this to the list if you like, but it
> seemed unimportant to bore the whole group
> with this until we can sort out between us why you and I are not
> seeing eye to eye.
I cc'd www-archive for the record.
Regards,
Maciej
