[sc34wg3] N358 and N372 Requires synchronization: changes in workplan, names, etc.

[Steve Pepper]

Hi Mary,

Thanks for your posting. Let me try and clarify the situation for you.

At 02:44 01.01.2003 -0800, Mary Nishikawa wrote:
>I guess it serves me right for leaving early before this was settled :)

Don't be too hard on yourself: no-one can be present all the time! But as 
long as *some* people from WG3 are present, things continue to be discussed 
and moved forward... which I guess is a good thing, really.

>On Saturday we had all agreed to a multi part standard.

That was the proposal as of Saturday, yes.

>Then this decision was discussed in private and another proposal was 
>submitted, it seems.

Actually the discussion started during the official WG meeting on the 
Monday, specifically around my suggestion that the Reference Model more 
properly belongs in a standard of its own. (I won't repeat my arguments 
here. I think I discussed them with you at least briefly before you left 
Baltimore.)

The Monday meeting failed to resolve the issue, which seemed to me to be 
largely one of marketing rather than anything technical. Charles Goldfarb 
offered his help as mediator, so I took the initiative for what turned out 
to be an Editors' Meeting on the Tuesday evening. Those present were Steve 
N and Michel (RM editors), Lars Marius (SAM editor), myself (convenor), and 
Charles.

With Charles' help we reached consensus on several issues, in particular 
that it is important to get the marketing right! Topic Maps mustn't appear 
to be more complicated than they are: Different kinds of users of the 
standard should be able to get at the information they need easily, without 
having to wade through oceans of what to them is largely irrelevant. In 
that respect, a series of small and compact standards (like the W3C's XML 
specifications) is better than a large and cumbersome one (like HyTime), or 
one consisting of multiple parts (like STEP). We already have four projects 
approved: Three IS's (13250, TMQL and TMCL) and a TR (Conceptual Model). 
Adding a couple more would result in a neat little 'family' that would be 
easier for most users to relate to (and avoid giving the impression that 
topic maps are impossibly complex).

In a sense, this was SC34 learning the lession of HyTime...

Following the meeting, Michel and I drew up a plan for a new "road map" 
suggesting a disposition of material across several standards. This was 
then communicated to all the WG3 members we could find in the time 
remaining before the SC34 closing plenary and received the approval of 
everyone who saw it. On behalf of Japan, Naito-san was fully informed, and 
I also had a brief conversation with him and Komachi-san to explain the 
thinking behind the proposal.

The new road map was discussed by WG3 during the closing plenary and 
modified slightly. The resulting consensus is given in N372, which was then 
approved by SC34. In my opinion, there is no conflict between N372 and 
N358. N372 simply represents the current consensus on how the material 
listed in N358 will be distributed among documents.

>The structure of the Topic Map Standard as described in N372 does not seem 
>to agree with the proposal in N0358.

N358 is not to be construed as showing a final disposition of content 
between what was originally expected to be multiple parts (and is now 
planned as separate standards). It simply shows the kind of content that is 
expected to be included.

>The national bodies need to vote on this by March 20, 2003. There is 
>conflicting information and we need clarification to consider this properly.

I will try to answer each of your points in turn. As you will see, I don't 
think there is any conflict (though the fact that you think there is shows 
that there is a need for clarification).

>In N358 we have
>
>1. A guide to the relationships between the topic maps standards
>2. The Standard Application Model (Data Model)
>3. The Reference Model (Information Aggregation Model), etc.
>4. XTM Syntax
>5. HyTM Syntax
>6. Canonication of Topics Maps
>
>For the program of work there will be 3 international standards and an 
>amendment or amendments to ISO 13250.
>
>Which ones are the standards and which ones are the amendments?

As noted above, nos. 1-6 are *not* a disposition of content, so there is no 
direct correlation between them and the new standards/amendments. Note the 
use of the word 'expected' (twice) under Programme of Work. In this way a 
New Work Item Proposal (NP) leaves room for adjustments, which almost 
always turn out to be necessary during the development of a standard.

>In N372
>
>1. ISO 13250  plus basic concepts (1 above?) and Annexes for XTM and HyTM 
>(4 and 5 above)? These annexes are the amendments mentioned above?

This is the amendment, yes, to 13250, which will not change substantially 
from what it is today.

>2. ISO XXXXX Topic Maps Data Model (SAM)  So this  name will be changed?

As far as I am concerned, none of the names are cast in stone yet.

>(Annexes for deserialization and mapping of SAM to RM .... new information 
>that is not included above? Why is this in the SAM?)

There is nothing new here. It was always the intention to include a mapping 
from the SAM to the RM. (This is part of the "relationships between the 
topic maps standards", see No. 1, above.)

Using the old terminology (which is on the way to being changed, as you 
will have seen on the mailing list), the "RM" will provide the basis for 
defining "TM Applications". The "SAM" is one such "TM Application" - the 
most important, standard, one. For each such "TM Application" there may be 
one or more interchange syntax (XTM and HyTM in the case of the SAM). Given 
this arrangement, it makes most sense to include the SAM-to-RM mapping and 
the (SAM-related) XTM and HyTM deserialization specifications as annexes to 
the SAM.

>3. ISO XXXXX Topic Maps Information Aggregation Model (RM)
>
>I guess this one has not been decided on yet with the discussions on this 
>list. I hope that whatever is settled is placed into N359 later.

I don't think that would be appropriate. N359 should reflect the position 
as of the Baltimore plenary. New decisions should be formalized in London 
in May.

>4. ISO XXXX Topic Maps Conformance
>
>(This looks like 4 standards to me now, is this correct?)

As of today, it is "expected" that there will be six standards altogether, 
exactly as listed in N372. Given that we already have approval for three 
(13250, TMQL, and TMCL) that means three new standards, as suggested in the 
NP. [I actually proposed that the Conceptual Model TR be "upgraded" to an 
IS and contain the SAM (or the RM). This is only a formal mechanism, but it 
would be a convenient way of cleaning our books. There wasn't consensus on 
this approach so it didn't go into N372.]

>Why all of these name changes all of a sudden?

They are a result of the evolving consensus on how the pieces fit together 
and how best to present them to the world. I can only speak for myself, but 
I have never regarded "Reference Model" and "Standard Application Model" as 
the final names, just as I didn't with "Foundational Model" and "Processing 
Model" earlier. That's one reason why I have been sceptical to too much 
marketing of these names outside the committee. I have always felt that we 
needed to reach a fuller understanding of the role of each before finally 
settling on their names. It seems to me that we are getting very close now...

>It is very difficult to comment on N0358 as it stands or perhaps this will 
>be corrected and resubmitted?

I don't think that is necessary. It is not a final blueprint, but it *is* 
sufficiently detailed that its purpose should be clear.

>At local Japan meetings, we have spent much time discussing what name has 
>become what. Where is the conceptual model, or TMPM4 or this or that data 
>model, etc. We need to translate all of these changes and explain why 
>these name changes have been made. There not only affects ISO but also JIS 
>and IPSJ who are made up of many representatives of Japanese electronics 
>companies and governmental bodies. Naito-san and I will be speaking on 
>Topic Maps at the IPSJ National Convention in March, and we need to give a 
>report on the progress  of the topic map standards. Especially, we should 
>hope for stable names at least.

I agree - but not before we really understand what we are doing: It isn't 
really feasible to name things appropriately before we know what they are! 
I am confident that the Japanese committee will be able to explain what is 
necessarily an evolutionary process in such a way that WG3's flexible 
approach actually strengthens the confidence of industry representatives, 
rather than weakens it...

>Now on this list this week, Steve Newcomb is asking for ideas to replace 
>the name of the RM and he has outlined a new work plan. Will this all 
>supersede related information that is now in 358? I request that this be 
>settled as soon as possible and really decide on the name and not keep on 
>changing it. Once the names and Program of work has really been discussed 
>an decided on, then have all of this in the Proposal for the New Work Item.
>Maybe it is already in the plan to update and resubmit 358, but from the 
>formality of the submission and the ballot, it did not seem this way to us.

I hope that my explanation helped and that you now agree that it would be 
positively *unwise* to cast names and dispositions of content in stone at 
this stage. An NP needs to leave room for maneuver. N358 does that, but 
also contains more than enough detail for the ballot. As to the "work plan" 
that Steve N mentioned, that is merely advance notice of how the editors of 
the "RM" plan to proceed in the next few months. As such it very welcome 
indeed and will help us make progress between now and London.

>I was very encouraged as I left Baltimore knowing so many SAM issues had 
>been resolved, and it was one of the reasons why I came to Baltimore in 
>the first place. I think we should all applaud the hard work and resolve 
>the ambiguities in the New Work Item so that we can go forward as soon as 
>possible.

I agree that the Baltimore meeting was immensely constructive. We should 
all congratulate ourselves (o-tsukare sama deshita). I think we are in very 
good shape and expect even greater things in London!

Thanks for your comments, Mary. I understand your concerns and hope that 
this reply puts your mind at rest. If you need more information, don't 
hesitate to ask, either on this list, or by contacting me directly.

Best regards,

Steve
--
Steve Pepper, Chief Executive Officer <pepper@ontopia.net>
Convenor, ISO/IEC JTC1/SC34/WG3  Editor, XTM (XML Topic Maps)
Ontopia AS, Waldemar Thranes gt. 98, N-0175 Oslo, Norway.
http://www.ontopia.net/ phone: +47-23233080 GSM: +47-90827246