Charlie Wiecha, IBM
Erik Bruchez, Orbeon
John Boyer, IBM (chair)
Leigh Klotz, Xerox (minutes)
Nick van den Bleeken, Inventive Designers
Rafael Benito, SATEC
Keith Wells, IBM
Sebastian Schnitzenbaumer, Dreamlabs
John Boyer: We had a full to
over-capacity crowd at the conference. I believe the room had about
20 people sitting on floors, not bad for a 7:30-9:30PM slot. The
chair of the conference said there wasn't even standing room. I got
feedback as did Dave Megginson that many people felt it was the
best thing they'd seen at the conference. They liked the format of
two-hour focused session with the 15-minute talks. They were hit
with the idea that XForms gives you a bit of a thicker client
mentality, combined with Web 2.0 technologies pushing the server
tier more towards it. One of the highlights was Erik Bruchez's
presentation, where XForms cut out the entire middle tier. Of
course, blogs, wikis, and social collaboration software.
Leigh Klotz: We're shipping
that.
John Boyer: Yes, I mentioned that from
Xerox.
John Boyer: I got to talk to Michael
Sperberg-McQueen, who was interested in editing HTML in a textarea
and see a display of the HTML. I think we have that with XForms 1.1
mediatype on output.
Leigh Klotz: We don't have it on
input.
John Boyer: Yes, I put that in the
future features triage I've done. Push the underlying content as
XHTML markup or at a minimum as an encoded string so we can get the
content back to the output.
John Boyer: The keynote speech was by Elliotte Rusty Harold. The demos leading off to it came off well. Elliotte's main point was XForms was a "Cambridge" technology, and what's succeeding are "New Jersey" technologies. His three following points
we need native support, for example in Firefox (not as a plugin)
Leigh Klotz: [irc] The name comes from Dick Gabriel's paper; here's a link http://naggum.no/worse-is-better.html
John Boyer: The first app was word
processing but the first "killer app" was a spreadsheet, which
doesn't do anything. So I pitched the idea that XForms is the
killer app because of what it does to middle tier coding. So that
can coalesce the first two points.
Charlie Wiecha: So what can we do in
the network other than transforming to Tag Soup and
JavaScript?
John Boyer: ...
John Boyer: On the browser side,
...
Leigh Klotz: And Google is using
Webkit (KHTML, Konqueror) in their mobile system.
Charlie Wiecha: Android.
John Boyer: Then on browser vendors,
the web is defined by more than what the browser vendors produce.
As we see from Yahoo and Google toolkits the web is bigger.
John Boyer: And there's some
discussion of adding XForms to GWT by Jerome van Rotterdam.
Charlie Wiecha: That's a bit of a
hack.
John Boyer: Can we make a record of
what we did? Charlie, your version is was somewhat dynamic.
Charlie Wiecha: There's a slide deck.
You can edit your profile and there is a place to upload.
John Boyer: Is everybody here two
weeks from now, the 19th? I propose we have a three-week break
starting then.
Charlie Wiecha: Steven is on vacation
to the end of the year.
John Boyer: Perhaps we should only go
for a two-week break.
Sebastian Schnitzenbaumer: I'll be
back on January 19, but I can make December 19.
Nick van: We probably don't have a
call on January 2.
John Boyer: So three weeks hiatus,
starting on December 26th. So no calls December 26, January 2, and
January 9.
Charlie Wiecha: We can meet the
9th.
Nick van: Yes.
Keith Wells: Yes.
John Boyer: OK, but that's the first
of the 90-minute calls. We can still get Steven to make the
changes. So, two weeks with some absences and we cancel December 26
and January 2.
Resolution 2007-12-5.1: We take a two week teleconference break and we cancel December 26 and January 2. January 9th is the first of the 90-minute calls, with some absences.
John Boyer: What do you think about
having our first virtual day on the day before everybody flies
out.
Erik Bruchez: February 1, fine for
me.
Charlie Wiecha: That works.
John Boyer: We could meet by phone and
decide timezone and work out the details. So it sounds like we're
amenable.
Leigh Klotz: I forgot. I will write
about xforms-1.1.
John Boyer: And I have the RSS feeds
on our public site, so when the report goes to www-forms, provide
the RSS feeds for public-forms. The exact feed links are on the
site already. I noticed that the feed links for www-forms and
www-forms-editor are incorrect so I replaced them with feeds that I
have verified actually work.
Leigh Klotz: They are hosted by
W3C?
John Boyer: Yes.
Leigh Klotz: Why is dialogs on 2.0?
What process did you use to put things on 2.0 instead of 1.2?
John Boyer: It's just an initial
triage. I knew someone would find that surprising.
Leigh Klotz: How are we going to
validate the 1.2/2.0 split?
Charlie Wiecha: I thought 1.2 was
mostly simplification. I'm surprised to see stuff other than 20
there.
John Boyer: I knew someone would find
that surprising as well. So this is mostly a list of reasonable
things that had the door closed out of 1.1. So for example,
model-driven switch with switch. Then look at #2, multi-case
switch; it could be 2.0.
Nick van: If we do this triage,
will the outcome be a requirements document? If so, will we be
required to do them all? Can we drop some once we do the
requirements document?
John Boyer: I know we did a number of
versions. We resolved we had completed 1.1 requirements. These
others seem more like point features.
Sebastian Schnitzenbaumer: I'd suggest
not doing a 1.2 or 2.0 features list. I'd recommend just a list of
the features as XForms 1 .2 WD, all on one document, and let it
develop from there, rather than splitting it out. I don't think the
modularization is a good approach. It sounds great but you lose a
lot of energy in small documents on different timelines.
John Boyer: I've more or less done
that, except there is one division. Consolidated Recalculate and
Revalidate is bigger than a sum-product function. We can move stuff
out of the 1.2 list if they're problematic.
Charlie Wiecha: That doesn't mean we
can't still use that 1.2 document in a modular way.
John Boyer: Yes, submission is now
self-contained in chapter 11.
Sebastian Schnitzenbaumer: I wouldn't
put multiple specs through the W3C process.
John Boyer: I would agree on
throughput. How much effort could we put into an instance-only
module document which is really just a chapter?
Charlie Wiecha: We could write it as
an almost standalone.
John Boyer: We're closer on
submission. There's a tight coupling of instance and model, which
seems to be needed in order to get the ease-of-authoring thing off
the ground. So we'll probably stick with one document but try to
get them more separate.
Charlie Wiecha: New types of
instances, such as JSON.
John Boyer: Over time. It would be
nice to say that we have a bucket of data called and "instance,"
maybe not called instance, and couple that with a referencing
scheme. I don't know how far down that path we want to go in 1.2 vs
2.0.
Sebastian Schnitzenbaumer: I also
tried to express that the best thing to get public involvement
would be one spec.
John Boyer: Or two specs now.
Nick van: Or a requirements document,
a draft version of both requirements documents.
Sebastian Schnitzenbaumer: It's more
pain to have a real requirements document; I was editor of the
XForms 1.0 requirements document.
John Boyer: We have been able to use
the requirements document to finish last call for our
recommendations as we can point to it when we get new requests.
John Boyer: Back Leigh's original
question on dialog.
Leigh Klotz: Mine was about the
1.2/2.0 process, not dialog, which was a test case.
John Boyer: My question was do I have
a clear and present idea of how to do it right now without shifting
other features. So I wasn't clear on dialog. If someone has a
proposal, like a tree control.
Charlie Wiecha: Try to New Jersify
this. We're still in spec-elegance mode. I can relate to a business
driver for #20. I don't have a sense for what the rest of the stuff
is about in the marketplace. So, as Leigh was saying, why are we
doing this 1.2? What's the higher-level purpose.
John Boyer: I can get a sub-triage
going of ease-of-authoring. Like #3 if you hit enter.
Charlie Wiecha: Some of these are
about business priorities such as granularity, ease-of-use,
etc.
John Boyer: We need to get the
discussion going.
Nick van: Leigh started last week, and
I edited many of these on the wiki.
John Boyer: Due to the craziness of
the network, they may have been deleted due to network problems.
The wiki software crashed on the w3c site and some stuff may be
missing. I don't have any network right now.
John Boyer: The direction we need to take on the list is for each person to think about how to elaborate bottom-up on these lists and express the business-case drivers for requirements to help us decide what really should do.