See also: IRC log
<trackbot> Date: 21 February 2013
<kford> http://lists.w3.org/Archives/Public/w3c-wai-ua/2013JanMar/0026.html
<kford> http://lists.w3.org/Archives/Public/w3c-wai-ua/2013JanMar/0041.html
<kford> Thanks,
<kford> Scribe: Kim
<kford> Jim's latest draft http://lists.w3.org/Archives/Public/w3c-wai-ua/2013JanMar/0055.html
Kelly: conformance – two ways to approach this. Jim sent another draft of an attempt to capture all the feedback. Take five minutes to read it. If we still think there's a lot of debate I'd like to step back and see where the points of disagreement are.
Eric: I did read through Jim's
document. I made a few minor comments to it – I thought it made
some good progress. And then I also sent those comments with
track changes, and I also took a crack at writing a conformance
section that addresses – builds on Jim's document it doesn't at
all the detail but sketches out an outline for it and reflects
some of the advances that I think that Jim's...
... document had.
... Those were sent to the list about 10 minutes ago.
... I was building on Jim's work, not adding all the detail,
but also adding a couple of concepts I think might be helpful
and ironing out the conformance section
Kelly: who hasn't read Jim's document?
Greg: so use Eric's document because that has everything
Jeanne: I would rather not modifier definition of user agent – that introduces a very big black hole. I would rather see a statement at the beginning of the conformance section that says for purposes of conformance the user agent is the software – however we want to phrase that. And therefore we don't need to touch the general definition of a user agent.
Eric: what I put in my email approaches that point
Jeanne: I'm happy with that
language
... can we resolve that were not going to change the definition
of user agent per se, but put a statement at the beginning of
the conformance statement that says for the purposes of
conformance the user agent is the software that's being
available for conformance
Jan: I do see value in changing the subtypes
Jeanne: I think it's confusing that Jim has it in the conformance proposal
Jeane: the first sections of it – the definition of user agent, what qualifies as a user agent. It's currently in the introduction, it's not in the conformance section today.
Greg: as part of the introduction
is it normative? we could have a normative section that
differentiates what we think is a user agent versus what is not
to guide other agencies. Theoretically in the end it's all
going to come down to the purchasing agency deciding what a
will do to comply with UAAG. So in that we are more
advisory.
... purchasing – that's where the decisions will be made
whether an operating system needs to comply with UAAG, whether
a telephone does etc.
Eric: in the second email that I
sent I essentially took that general approach – that a user
agent conforms when it meets all the applicable requirements
specified by the document. So there is some flexibility about
what kind of software you wish to conform, but it does
enumerate two major kinds of conformance. This is what I
discern from Jim Allen's and the structure of our last
discussion....
... A couple of new terms – core conformance, which is for user
agents to have broad capabilities for retrieving rendering and
user action, and the classic example is a web browser.
... and the other major categories when I referred to as
auxiliary conformance, and that's for user agents that have a
narrower set of capabilities– don't render any content. So
within core conformance it appears that Jim's document has a
couple of major subtypes. In my rewrite I didn't elaborate on
those major subtypes, but that would be a detail under the core
conformance.
Greg: my comment was just where does our discussion of user agent belong – normative for non-normative section.
Eric: at the end of the day
conformance is about what you define as your user agent. It
begs the question what about all that earlier determination you
make about one of the major categories of things, or how strict
you want to be about the kinds of user agents that would be
eligible for conformance. I don't know that I know fully the
answer, but I did in my rewrite put some of that...
... information in the conformance section.
Kelly: Jeanne, back to your comment about the first part – you were saying just eliminate that?
<jeanne> http://www.w3.org/WAI/UA/2013/ED-UAAG20-20130108/#intro-def-ua
Jeanne: In the January 8 document– putting it in IRC. Here is Jim's part about the definition of user agent, in the introduction. And if you look at the implementing document that's where the second section is.
<jeanne> http://www.w3.org/WAI/UA/2013/ED-IMPLEMENTING-UAAG20-20130108/#intro-def-useragent
Jeanne: that's all
informative
... so that first definition of the user agent is informative.
Jim quoted it straight – when Eric made comments, he added
comments of the definition of the user agent that really only
belong in the conformance section, not in the introduction
<jeanne> http://www.w3.org/WAI/UA/2013/ED-UAAG20-20130108/#conformance
Eric: making the distinction
between those two definitions may be valuable even early in the
document. That is what is really intended within the success
criteria. That is the definition that is being used there. So
maybe useful to make that distinction prior to people reading
the actual requirements or especially, success criteria.
... might be helpful if I quote from the second paragraph
User Agent. In general, a user agent is any software that retrieves, renders, and fosters user interaction with Web content. However, another, subtly different meaning of user agent is: the software for which conformance to UAAG2 is being sought; as described below, the user of this document has considerable flexibility in defining the exact nature of the user agent for which conformance is...
scribe: sought.
Eric: it generally touches on
both definitions – it might be useful that only in the
conformance section but maybe earlier so people get the idea
that there's at least a two-part process. One is figuring out
what are the kinds of things – software entities that should be
considered for UAAG conformance and then second is defining the
user agent – what you mean by it with all the...
... inclusions...
... and exclusions – and in the third step would be to apply
all the applicable requirements to it and see if it
works.
... it might be possible to do that rather simply early in the
document but I agree with you that it is critical to have that
second definition understood when people are getting into the
conformance section.
Greg: by the phrase second definition you are referring to anywhere where conformance is sought
Eric: within those broad
constraints or guidelines and there are these details that can
be defined by the user of the document, and that's what
ultimately you are testing against. Guidelines about what kinds
of document, what kinds of software is relevant – it is
relevant to conformance because the set of requirements that
are likely to be applicable very depending on which type
of...
... conformance who
are seeking. Full conformance, which I called core conformance, then a large number of the success criteria would be mandatory. They cannot be non-applicable. On the other hand if it's just auxiliary conformance and you're dealing with just a plug-in or an extension, that you might have a large number of success criteria that are non applicable
Kelly: any reactions?
Jeanne: quick reactions – the
first is I don't like the word core because to me core is a
small subset of the requirements that would be core and what we
are really saying is the full conformance is everything – you
comply to everything not a smaller subset. The concept is fine,
I just wanted to comment on core. But I do like the word
auxiliary – I think that addresses concerns of other...
... groups that we not use the word partial.
Eric: that makes sense. Regular performance versus auxiliary? I was not so much in favor of full conformance because it seems like even something for which you are seeking auxiliary conformance it can be fully conformance. But regular conformance would be another possibility
Greg: getting lost with the three documents
<jeanne> Regular conformance, Comprehensive conformance, for terms for Full conformance
Jan: I'm not exactly sure what
the document is proposing to take out
... full isn't the greatest word, but it's what atag uses and
is much as possible I'd like to coordinate between those two.
partial don't see that it's a problem, auxiliary hard to
pronounce
Jeanne: problems with partial
Jan: the way I see partial conformance is the user doesn't care – they have certain interactive needs and are they or are they not met – partial just as well as part of the puzzle – meeting them and this is why
<jeanne> Partial, Limited, Auxiliary, conformance
Greg: bopping between bits –
better to have here's a document we're looking at, here's a
list of issues would be easier
... do we have a list
Eric: I have one thing that
should be on the list. To put it in perspective the issue of
whether – I basically like Jim's distinction between full and
partial, whatever it's called is a secondary matter, there is
one thing that I thought might be useful in thinking about
conformance that was not really in Jim's document. And that was
the notion of mandatory requirements. I don't think
that...
... was really mentioned at least in quite that way in the
document. It might be useful for the group to consider it.
<Greg> The term Partial Conformance isn't used in Jim's doc.
Eric: depending on the type of user agent there may be certain requirements of the document that are mandatory
<jeanne> I agree with Greg. We need a list of issues where we can say we agree and disagree.
Jeanne: example?
Eric: my main point is list of
requirements that are mandatory. For auxiliary may be different
list. once you have defined the mandatory requirements for each
type then you are left with the nonmandatory requirements for
each type. Those are the ones for which you read need a
reasonto not have them be applicable. It can be useful in terms
of making clear what people need to actually do – to...
... distinguish between mandatory requirements and
non-mandatory with the idea that it is the non-mandatory
requirements for which you need to determine whether they are
applicable or not
... you can actually get to the point of enumerating, for this
type of user agent here are the mandatory ones that are always
going to be applicable and the others are nonmandatory, and
then iif you decide they're not applicable you have to have a
good reason for it.
<Greg> If Eric's suggesting a matrix of row being level (A,AA,AAA) and column being "mandatory" and "can be N/A", that sounds like a huge change, and one that I'm skeptical of.
Jeanne: strong inclination not to
do this – it's adding another level of complication. I think
that's making it excessively complicated
... it may cause more confusion for people who are trying to
follow the document – not closing my mind to it completely, but
would rather we find a solution that is adding complexity.
Greg: it sounds like that would be replacing three levels with the matrix of six – I would be very hesitant – perhaps it would be useful if you come up with examples that we could put into those levels. Perhaps there's a way of denoting that in the document without making the overall structure more complex
<jeanne> This is the current Conformance section in the existing draft <- http://www.w3.org/WAI/UA/2013/ED-UAAG20-20130108/#conformance
Eric: how many categories do we have – partial and full, and under full, web-based, and another subtype.
Jeanne: that's part of what we're debating
Eric: that's what's in the
current document – maybe instead of mandatory, softer approach
here's an example of one of each of these, and for this regular
one because it's doing this thing there's no reasonable excuse
for this not being applicable
... show how the set of applicable requirements is related to
your taxonomy.
... I don't see any point in elaborating on this distinction of
different types of user agents and then not have them impact
directly your definition for your specification of what
requirements are applicable
Kelly: I am concerned that we are
making it too complicated and some of this would be better as
guidance and in implementing document – if you're going to try
and do this these are the absolute things you should do versus
making them absolutes in – this document is already pretty
onerous.
... how do we want to move forward on this, because we probably
are on four hours of discussion on this and I don't feel we are
any closer to resolution. Eric's taken a stab at it, Jim's
taken a stab at it. We have taken stabs in meetings. I am open
to suggestions – we aren't done because we're still talking
about this. I'd like to stop the discussion and I'd rather talk
about ideas on...
... how we close this down.
Jeanne: our current conformance
statement is actually pretty simple. I'd like to suggest that
we take a look at the existing conformance statement and each
of us make a list of what is wrong or unaccessible or needs to
be improved in the current conformance statement – the one that
is in the document now. I don't think Jim's intention was to
move the definition into the conformance section....
... I think some of our confusion is building and building and
building on previous proposals, and we've lost where we started
from.
Kelly: how do you want to do that – make a survey?
Jeanne: yes – I can do that
<Greg> As I said, it seems to me that we're jumping around and not making much progress.
<Greg> My suggestion is to get back to working in a structured way, working from the the editor's draft and the databases of issues/suggestions people raise regarding it and actions assigned. Vote on suggestions, put them into the next editor's draft, and if necessary and then go on with issues and changes to them there.
<Greg> I like using the Issues database, which helps by giving titles and ID numbers to issues. Having separate versions with different people's suggested changes makes things difficult for me.
<Greg> I think I'm saying I agree with Jeanne's suggestion.
Kelly: we are tapping the discussion at 45 minutes next week. If you have things to say put them in the survey so that we can have a closing down discussion next week. Any objections to that?
<jeanne> there is good language in the other proposals, and I think we can use some of it.
<jeanne> ACTION: jeanne to make a survey of the current conformance section, so people can make comments and changes based on that. [recorded in http://www.w3.org/2013/02/21-ua-minutes.html#action01]
<trackbot> Created ACTION-803 - Make a survey of the current conformance section, so people can make comments and changes based on that. [on Jeanne F Spellman - due 2013-02-28].
Jeanne: recharter by June, need to give information March. Milestones – last call, candidate recommendations,
<jeanne> ACTION: jeanne to make a survey of the charter milestones. [recorded in http://www.w3.org/2013/02/21-ua-minutes.html#action02]
<trackbot> Created ACTION-804 - Make a survey of the charter milestones. [on Jeanne F Spellman - due 2013-02-28].
Kelly: survey with those items
and asking people's best estimate of when we are going to reach
them, just month and year. We've all been in this group for a
long time – we all know how quickly we do or don't get things
done. I think that's going to be as informative as anything
else.
... when we get into the mode of using the surveys we do better
as a group. Next week focus and lockdown.
... that will still give three weeks to give Judy what she
needs
... take up Eric's feedback next week.
This is scribe.perl Revision: 1.137 of Date: 2012/09/20 20:19:01 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) No ScribeNick specified. Guessing ScribeNick: KimPatch Found Scribe: Kim Present: Kelly Kim Simon Eric Greg Jeanne Jan Regrets: Jim Found Date: 21 Feb 2013 Guessing minutes URL: http://www.w3.org/2013/02/21-ua-minutes.html People with action items: jeanne[End of scribe.perl diagnostic output]