TAG -- 18 Jul 2006

<scribe> scribe: dorchard

<Norm> Scribe: Dave Orchard

<Norm> ScribeNick: dorchard

<Norm> Date: 18 July 2006

<EdR> http://www.w3.org/2002/ws/addr/minutes.html

<EdR> (tips for taking minutes)

the previous link has

Decisions and Resolutions

RESOLUTION: [precisely worded text]

<noah> More hints on taking minutes: http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm

future telcons and scribe

Henry to scribe once VQ convinces him

Henry to scribe once VQ talks to him

minutes

RESOLUTION: Minutes of 27-Jun approved

<noah> The minutes we just approved for 27-June-2006 are at: http://www.w3.org/2001/tag/2006/06/27-minutes.html

Agenda

EdR: can't do passwords in the clear

generic resource

<EdR> I'll publish Passwords a week prior to the 8th meeting for public/tag comments.

<EdR> please put it on the agenda for the 8th.

TV: still need more to do

tvr: also addressing public comments
... seems to be benignly interesting

vq: need new/more reviewers
... volunteers?

norm: I can review but won't be present on the 8th

do: no time

nm: date is important so reviewers know when doc is stable

tvr: pls review after Aug 8th

<scribe> ACTION: TVR to publish new version of Generic Resources by Aug 8th [recorded in http://www.w3.org/2006/07/18-tagmem-minutes.html#action01]

vq: after checking into cvs please send to public list
... any more comments

dbo: I think tvr is still working on my comments

tvr: I thought I got them, can you look again?

New and retrofitting new media types

vq: AB asked about using media types that aren't yet registered but used

dbo: think we shouldn't say anything against using unregistered media types

tvr: I was on voice browser wg, and the group was worried about getting LC comments
... can also relate to TAG position

nm: is there a written policy they should know about?

nobody knows of anything

nm: some worry about using unregistered media types

tvr: the widely deployed ones that aren't registered got widely deployed partly because they just went ahead and "did it"

nm: balance between saying there is a reason for registration
... reasons to reuse types

vq: ken laskey suggested something like this.
... then AB could use this and examine wrt w3c process

nm: not sure about how formal to be and even list in w3c process

vq: AB seems to ask for something informal, short email even

<Norm> Mostly I agree with Noah. I was going to point out that there's a distinction (in my mind) between a type being defined by a specification (where we could reasonably insist on registration) and a type used by a specification, where I think it would be sufficient for the spec to note that in a non-normative note.

<noah> Right, what I was trying to say was that it seems heavyweight to me to actually update the W3C process to tell everyone submitting a REC that they have to formally justify use of unregistered media types.

<scribe> ACTION: Noah to draft of something rough on this [recorded in http://www.w3.org/2006/07/18-tagmem-minutes.html#action02]

<noah> ACTION 2=Noah to draft a very short email suggesting that in general its good practice to reference registered media types, but that we recognize that there may be good reasons for occasional exceptions. Note is due 25 July 2006

vq: there was also some discussion from AB on changes to ecmascript and replace other media type

nw: on updated rfcs, specs need to do addendum to get new references

passwords in the clear

<scribe> ACTION: EdR to publish updated passwords in the clear by Aug 8th [recorded in http://www.w3.org/2006/07/18-tagmem-minutes.html#action03]

Versioning

<EdR> Vincent: Dave has updated the versioning document yesterday.

<EdR> Dave: I prefer to walk through and give an overview and then take detailed comments.

<EdR> DO: In general what I've done has expunged much of the xml

<EdR> DO: the sections on compatability, section 2 evolution is the same.

<EdR> DO: then I did some more work on principles.

<EdR> DO: I updated the bullet diagram based on our 'on the board' discussion at the f2f.

<EdR> DO: I've tried to move towards the model of defining compatability based on the definitions of these sets.

<EdR> DO: In section 1.6 I talk about components instead of elements because we need to talk about 'things', however I still need to talk about what a component is.

<EdR> DO: I've added in the notion to avoid the 'big bang' by offering multiple versions of the system, where either there are multiple versions or one version maps to another.

<EdR> do: this cant work in all cases (where new data fieldws are required) however.

<EdR> Noah: I think is largely improved, however I'm only 5 pages into it.

<EdR> Noah: In the sections I've read I think the foundations are much stronger.

<EdR> Noah: I feel something needs tuning up.. it seems your looking for a very crisp notion of forward and backward compatiblity but this tends to be a matter of degree.

<EdR> Noah: I think there are times where 99% of the instances have compatability but 1% would be in error. I think people do this all the time...

<EdR> Noah: I'm trying to seperate the notion that nobody should be able to every come by an unidentified compatability.

Noah: one important point is whether software will make mistake in interpretation

<EdR> Noah: We want to talk about the case where we plan for evolution, and in some cases we want to lock it out.

<EdR> DO: I point out that if you want to enforce it, you need to point out a way to the users that you have a defined incompatability.

<EdR> Vincent: any other comments?

<EdR> Norm: I have not had a chance to read it, but would like to for next week.

<EdR> Ed: I havent yet either, but would like to.

<EdR> Vincent: We can review in more detail next week.

<EdR> Vincent: I've put it on the top of the agenda for next week.

<EdR> Vincent: any other topics?

<EdR> Vincent: meeting is adjourned.

- DRAFT -

TAG

18 Jul 2006

Attendees

Contents