December 17 Draft minutes

[Dave Marvit <dmarvit@pacbell.net>]

Minutes
OASIS LegalXML eContracts Technical Committee
December 17,  2003Conference call

The following were present:
Daniel Noll
Dr. Leff
Charles Gillam
Dave
Steven Pemberton (from the W3C)
Rolly Chambers
Shane McCarron (also from the W3C)
Zoran
Dan
Jason
John McClure

Dan proposes to defer administrative matters since we have some guests 
on the line:
[No objections heard]

Dan: I’d like to start by thanking our guests from the W3C XHTML 
standards drafting effort.

By way of setting context, Jason and Peter have provided some 
questions. This should provide some context. Would you folks mind 
introducing yourselves, and then we can have an open dialog about how 
your efforts might fit with ours…?

Steven: I’m Steven Pemberton form Amsterdam. I’m chair of the HTML2 
group as well as of the forms group. XHTML2 is only one of the 
standards we are defining, but everything we are doing is leading to 
XHTML2. One of our aims is to use generic technologies whenever 
possible. When they didn’t exist we tried to create them. Generic means 
of use by as many as possible.

At the moment we are leading up to last call for XHTML2. I imagine that 
your process is somewhat similar and that you have something like last 
call and know what it means when we talk about last call.

So we are approaching last call. We imagine that we will get a lot of 
comments when we get there. Our aim it to have a candidate 
recommendation (CR) within 12 months, and possibly more than that.

Dan: OK, that gives us a good sense of where you are. In OASIS we have 
2 layers. At The TC layer we have TCs that put out a TC specification. 
The committee can call it there and companies are free to implement it 
there. Or they can go on to the second level and put it forth as an 
OASIS standard. This is a longer and more formal process.

We are working on our committee process. We will have three layers, a 
structural markup which seems to have a lot in common with XHTML2, a 
semantic layer, and an envelope specification of some kind.

In the semantic layer, in addition to having semantics for a variety of 
clauses we expect to have triggers for a variety of workflow issues.

The issue of the month, and the last couple of months, is the 
structural layer. We are trying to see how X2 will fit there.


Could the other members introduce themselves?

Shane: hello. I’m Shane Mcarron. I am lead editor and “conformance god”.

Steven: How do you want to structure this?

Dan: I know that Jason and John have some questions. If you want to 
answer questions that were emailed…?

Steven: One thing that popped into mind is that something has made it 
into the standard that is not in the draft, and that is how to deal 
with meta-information. That’s not only RDF, but…

What we have noticed is that lots of different interest groups have 
been asking us to add different elements, person, date, time and so it 
goes on. While there is a lot of interest there, we recognize that this 
is an endless process. We have to attach those elements to the 
structure. This gives us the best of both worlds.

Dan: When you say attach them, we have the element attribute model. We 
also …

Steven: If I can flesh it out a bit. The solution came from the 
accessibility world that needs to be able to understand the purpose of 
the element. We have an attribute called ROLE that specifies the role 
of an element. So, for example, we might have a paragraph and it’s role 
would be ‘indemnity clause’ Role =: OASIS indemnity clause.

Dan: I’m sure that Jason and John have lots of questions. Do you have 
any that you’d like to pose now?

Steven: Jason’s first question was if XHTML2 was intended to be used 
for business documents. Our intent is that it be a broadly usable 
document type without being all things to all people. We are trying to 
find a balance between usability and size.

We can’t require, or rather we can’t say ‘we’ll solve that problem 
using tools’. We need to keep the folks using text editors in mind.

Jason: How we got to where we are today in the WG is that we have been 
on the quest for a structural markup for contracts. The idea has been a 
subject of considerable debate.

Fairly recently the current draft of XHTML2 came to our attention. 
There is a view that since it is XHTML, and since it is intended to be 
a standard in the marketplace, there is a sense that it will sweep 
everything in the marketplace before it. So, what is the point of our 
making our own if XHTML2 is going to be good enough…?

Steven: Early in the process we developed the idea of modularization. 
What we wanted to be sure of is that if one used tables the others 
would use the same table. Now X stands for extensibility, but that’s a 
fine-grained extensibility. We wanted a courser-grained extensibility 
so we had the idea of modularity.

Our own rule is that if you want to use the name XHTML anywhere you 
need to have certain modules included.

Jason: There is a view in our group that XHTML2 will take over the 
world. I’d like to hear the ambitions of your group.
The thing that concerns me is the ease of authoring for non-technical 
authors. That’s why I was wondering about certain elements. What we are 
envisioning is having this on the desktops of attorneys and others and 
having these elements makes it harder for them to use it.

Steven: One of the primary thrusts is thinking about the end user. Part 
of the success of html is that it didn’t require tools.

We are up for discussion on some elements. We kept them because they 
are low implementation cost and possible benefit.

Dan: Do any of you have Peter’s questions? They would be good to 
address.

Shane: I have it. The first has already been answered. It can be used 
for a variety of applications, depending upon your timeline.

Steven: One of the aims of XHTML2 is to use generic technologies as 
much as possible. There are many XHTML2 documents that work in existing 
browsers. The things that don’t work in a generic way are forms and 
events. The rest of the stuff you can do in browsers already without 
having to re-write the browsers.

Jason: I like the x-forms a lot. From a conformance standpoint, do you 
have to keep that to be compliant?

Steven: No, You can remove that module and still use the XHTML2 name if 
you add something to it. XHTML2-legal, or something.

Shane: You can still be a host language if you want to be.

[Shane cites Peter’s question about conformance levels.]

Shane: The choice of conformance level is to some extent philosophical. 
It also has to do with adoption. If you are a host language then you 
are more likely to be interpretable by XHTML2 family.

Dan: What are the types of features and available elements that we want 
in our dream contract specific markup that we want to be in our 
contract specific language? Does that impose too many…

Shane: I was hoping that we could get through the existing questions 
first.

Dan: Jason?

Jason: If you would describe the major things that you are and are not 
allowed to use for host language conformance?

Shane: You need to use the required modules – and it is a short list. 
You need to use the basic structure, lists, and hypertext links. Also, 
if you pick up a non-required module you need to pick up all of the 
functionality. You can extend the module but you can’t subset the model.

There are some picayune details about white space but those should not 
be relevant.

Jason: Are we allowed to add elements to a required module?

Shane: Yes. You add elements to your own module. You define your own 
module, add elements to your module, and then attach  it.

Jason: So it is an XHTML 1 model.

John McClure: Steven and Shane, I thank you for joining the call. I am 
an old RDF hand so my perspective comes form that direction. In 
contracts in particular the predominant structure is the clause. As we 
look at XHTML2 we intend to use the section element to represent the 
clause. It would need to have a heading and a caption number. It looks 
as though the XH elements are going away. What is the current status of 
the nr elements?

Steven: you are talking about numbers that need to be explicitly in the 
doc and not generated by a style sheet?

John McClure: Yes

Steven: It is there as an idea, but it is not surely in or out.

John McClure: It would be great to see that element standardized on 
your side of the fence.

Steven: If you want to give us that as a data point  that would be 
good. We need to review our current semantic elements and see if we can 
improve them or remove them. We need to iterate again.

John McClure: A contract is divided into the main doc, but has 
schedules and appendices attached. Clearly we would not want to see the 
section applied to those documents. Is there any discussion at all in 
the W3C about things like appendices?

Steven: You started by saying that clearly you wouldn’t and it is not 
clear to me why saying section with a sufficient marker would not be 
good enough.

John: I mean without specifying a role

Steve: We already have two ways of dividing up a document. One is the 
DIV and the other is the section element. Adding new elements that 
essentially do the same thing but add a bit of semantics would be 
harder to convince the group to accept. That is not to say that we 
wouldn’t do it. We are always open for suggestions.

John McClure: If we are relying on a marker such as role to indicate 
that it is an attachment rather than another clause in the document 
then it tends to hide the info further in the document. Consider that 
as a possible data point – having a document with front and back matter 
would go a long way. We would want this architecture at the level of 
the W3C because it would impact our ability to be conformant

Jason: Would you be happy with a head, body, tail elements?

John McClure: Well, head is typically reserved for metadata so head, 
body, tail, may not go far enough.

When is the next draft?

Steven: Shortly. The middle of next month.

John McClure: Let’s talk about the grammatical paragraph. Is the 
intention to walk away from the typographical structure that is the 
hallmark of XHTML1 and move to lists?

Steven: We have dome that already. That is exactly the idea. A 
paragraph has been a very simple block of text and it is not what you 
think of a paragraph is. This is because of user demand and user 
analysis. People thought of a paragraph as a conceptual unit. Since a 
paragraph can contain a list, people should be able to represent that.

When you have a mental model of a paragraph then it has some semantic 
meaning for the person writing it and you can tag it for that semantic 
meaning much more easily.

John McClure: Any consideration for inline lists as compared to block 
lists?

Steven: Those concepts disappear. We have noticed a confusion with 
users when using block in HTML and CSS. If something is called a block 
that references its content model and not how it is presented. If you 
have a list and you want to present it inline then that is perfectly 
acceptable.

Shane: You can distinguish between flowing and block lists through 
style.

Steven: The class attribute is the place where you would reflect that 
kind of information.

John McClure: Signatures? This is an issue that is apart from digital 
signatures (which is an encapsulating technology). What we want to do 
is be able to determine where a signature is located. It could be 
related to a button, for example. Has there been any discussion about 
establishing a module for signatures?

Steven: Not that I know of. There has never been any discussion. I am 
not aware of any requirements or discussion.

Shane: When we have talked about things like this, we have said that 
the ‘role’ attribute is good for this sort of thing. Saying Role is 
OASIS: signature then this is an opportunity for you to define exactly 
what you mean by signature.

John McClure: Is role a way of attaching semantic meaning to any 
element?

Steven: Yes.

John McClure: How does this relate to the Ruby module?

Steven: The Ruby model was created by a different group. That is a very 
interesting comment on your behalf. There has been no discussion about 
using Ruby for that purpose, but I agree that this is something that 
Ruby could be used for. Ruby will be a module that is included.

John McClure: Do you see a conflict between roles and Ruby?

Steven: I don’t see why you might see a conflict there.

John McClure: It appears that Ruby was a method for annotating an 
arbitrary string of text.

Steven: Yes. There is no requirement of the Ruby markup to use Ruby 
presentation. I personally think that there is a lot of possibility for 
use of Ruby in specialized use cases.

John McClure: I am finished asking about.. well I am prepared to get 
into semantics but if you would like to get into structure..

Jason: Well, we have only just started to mark up contracts with XHTML2 
to see how hard or easy it might be for end users. I am interested 
in... in a section you can have mixed content…? Why are you allowed to 
put PC data in a section if it is most convenient to think of it as a 
container?

Steven: I agree with you. But I think it was to make it as generically 
useful as possible. If you want to use it as a container then that’s 
all right. We are granting the freedom to let people use it if they 
really want it.

John McClure: So it is for ease of authoring.

Steven: We didn’t want to restrict the way people would use it. We 
didn’t want to unnecessarily restrict people.

Jason: It does mean that an authoring environment would assume that a 
section is a paragraph rather than a container for a paragraph.

Shane: In addition to the role attribute being available on all XHTML2 
elements, it is also available on all of your elements. XHTML2 
guarantees that the semantics are available on your elements.

John McClure: How would the XHTML2 data stream would point at the 
relevant RDF data stream that describes the semantic content of the 
document?

Steven: These are details that we are hammering out now. Hopefully by 
mid January…

John McClure: What is the relationship between qname in the role 
attribute which is to be defined for a given namespace, and the 
namespace that is used within?

Theoretically one could have those qnames inside an RDF description 
element that references that...

Steven: We intend to specify how an XHTML2 document gets interpreted by 
an RDF processor, but we haven’t done that yet.

Dan: I see that we are running up against our traditional hour for the 
call. I wanted to see if there are any objections to running a few more 
minutes.

Shane: I have to run.

Dan: Thanks for your time Shane. I hope that this can be the beginning 
of a more consistent liaise between our efforts.

Dan: Do we have any outstanding obvious obstacles in the mind of any TC 
participant for using XHTML2 as a host language? I did hear a few 
things about lists...

John McCure: As far as the structural model, the issue of the NR tag is 
hanging out there. The issue of siblings to the body tag remains.

Dan: Jason, do you have any issues?

Jason: I think it helps to know what we are going to aim for. There is 
nothing more I’d like to ask (thanks very much). In terms of deciding 
if we can run with XHTML2. I want to sit down and see how it will work 
in an editor from the perspective of our target user base.

Dan: Of course. I wasn’t trying to push for consensus yet…

Dr. Leff: Just quickly... Could XHTML2 be inside anything else, such as 
a conditional clause? If something might happen then something else 
might happen?

Steven: No, there is no reason.

Dan: I’d like to suggest that as a follow on the TC ruminate on and 
digest what we heard today and develop a list of questions or topics 
for follow on questions. Also, is there a formal process for dealing 
with other standards bodies, or is informal email exchange between 
chairs appropriate?

Steven: I don’t think anything formal is required. Just sending email 
is fine.

Dan: Great. You mentioned that you are trying to apply an 80/20 rule. 
Is there some process by which we can make our sticking points know to 
you

Steven: www-html-editor@w3c.org. That is the official entry point for 
questions. We have a formal process that every comment that goes there 
gets logged and gets and issue number and gets replied to.

DG: Is there someone from MIT in the working group?

Steven: OASIS is a member of the W3C so feel free to join. You would be 
more than welcome. We have a 1 hour phone call 1 time per week.

DG: Is Fujitsu represented?

Steven: Yes in the W3C, not on the WG

Dan: More questions?

Zoran: Some of the functionality you have described is in X-forms. What 
is the alignment between the two?

Steven: X-forms is an integral part of XHTML2. It was always intended 
to be. In fact the X-forms group was a subgroup. But it required such 
specialized knowledge that it became a separate group rather than a 
subgroup.

Zoran: So you answer to Dr. Leff was that it comes from the X-forms 
part.

Dave: What are your intermediate milestones?

Steven: We plan one more working draft, then last call, and CR in this 
coming 12 months.

Dan: As of the next working draft, assuming that there are …

Steven: The next draft will be a working draft. Comments on that will 
change it into the last call working draft. And comments on that will 
turn it into a candidate rec.

The further you get in the process the less change there is. All I can 
say is that we are converging. The RDF part is quite a big change, but 
that doesn’t affect the structural  part. That hasn’t changed in a 
while.

Jason: The RDF part that you mentioned is a big change. Will that be 
required stuff, or will we be able to opt out of the RDF stuff if we 
chose to do so.

Steven: We haven’t decided yet.

Dan: Thank you very much Steven for your time.

Steven: Thanks to you all…

Dan: Would the group still stay on the line?

What about the timeline? How can we base something on a standard that 
is going to change over the next 12 months?

Dave: Well, the structural part hasn’t changed much…

Jason: I agree with Dave. I don’t know to what extent the TC wants to 
decide if it is a sensible approach or not, but we are in a position to 
say that if we wanted to go for post-language level conformance we 
could decide what that would be. It would be possible for any member of 
the TC to sit down and decide. We have some solid ground to work form…

John McClure: I think the next step is to develop the set of talking 
points. It might still be addressed in even the January draft. If they 
are undecided on something such as the NR …

[discussion about the need for front and back-mater and how to 
represent that. Should it be in their representation, or can we do it 
with our module]

[discussion about scheduling]

January 7 for our next meeting

Dan proposes that requirements should be the first half of next 
meeting, then working out the list of questions to pass to W3C.

[Discussion that when representing the group in messages to the W3C the 
information should be on the list… That way members are not left ‘in 
the dark’ as to what the group is doing.]

Meeting adjourned