(has a temporary acl, ask rms@ansol.org to be added to editing group)

CT-173 meeting of 2007-07-16 By Rui Seabra

Contents

CT-173 meeting of 2007-07-16 By Rui Seabra

Intro

This document is in english because being in english may be of interest to ISO, many stakeholders and at least one participant of the meeting.

Participants

Initials (by order of participation):

MSD: Miguel Sales Dias, Microsoft, President of TC 173 "Lin guagem de Descrição de Documentos"
CR: Carlos Romero, ONS and TC member. ONS is II (Instituto de Informática)
GH: Gustavo Homem, Ângulo Sólido, Vogal
MV: Mário Valente, Director of ITIJ, ODF Alliance member, Vogal
MS: Marcos Santos, Microsoft, substitute for MSD
RS: Rui Seabra, ANSOL, Vogal
JPM: José Miranda, MULTICERT (portuguese certification authority), proposed Vogal, not verified, invited as expert by ANSOL
ME: Miguel Espiga (not sure of name, surname is correct, I think), log OpenSource Consulting, proposed Vogal, not verified, invited as expert from ITIJ
MCM: Maria do Céu Mendonça, Primavera (MS business partner), Secretary
PQ: Pedro Quintas, Jurinfor (MS business partner), Vogal
JM: Joaquim Machado, INE, Vogal
MJM: Mário J. Madeira, (Instituto de Informática da ?) Segurança Social
MM: Mário Morgado, Caixa Mágica (portuguese GNU/Linux distribution), Vogal
JN: João Neves, Intraneia, Vogal
SM: Stephen McGibbon, Microsoft, OOXML expert
LS: Luís Silveira, CIIST, Instituto Superior Técnico, proposed Vogal, not verified, invited as expert from INE
PS: Pedro Seabra (no relation with me AFAICT) Via Tecla (MS business partner)
RB: Rui Barroso, ADRAL, Vogal
MC: Manuel Cerqueira, ASSOFT (BSA alike)

FIXME: get names and affiliations of all participants.

Total Participants

Initials (by sitting order):

CR: Carlos Romero, ONS and TC member. ONS is II (Instituto de Informática)
MSD: Miguel Sales Dias, Microsoft, President of TC 173 "Lin guagem de Descrição de Documentos"
MCM: Maria do Céu Mendonça, Primavera (MS Business partner), Secretary
LS: Luís Silveira, CIIST, Instituto Superior Técnico, proposed Vogal, not verified, invited as expert from INE
RB: Rui Barroso, ADRAL, Vogal
RM: Rui Mendonça, jp sa couto
AC: Alexandre Correia, Safira (MS Gold Certified Partner)
HO: Henrique O'Neill
RC: R. Coelho, ASSOFT (BSA alike)
MC: Manuel Cerqueira, ASSOFT (BSA alike)
SM: Stephen McGibbon, Microsoft, OOXML expert
MS: Marcos Santos, Microsoft, substitute for MSD
PS: Pedro Seabra (no relation with me AFAICT) Via Tecla
PQ: Pedro Quintas (not sure of the name), Jurinfor, Vogal
JM: Joaquim Machado, INE, Vogal
MV: Mário Valente, Director of ITIJ, ODF Alliance member, Vogal
JN: João Neves, Intraneia, Vogal
GH: Gustavo Homem, Ângulo Sólido, Vogal
MM: Mário Morgado, Caixa Mágica (portuguese GNU/Linux distribution), Vogal
RS: Rui Seabra, ANSOL, Vogal
ME: Miguel Espiga (not sure of name, surname is correct, I think), LOG OpenSource Consulting, proposed Vogal, not verified, invited as expert from ITIJ
MJM: Mário J. Madeira, (Instituto de Informática da ?) Segurança Social
JA: Jorge Alves, ACIDI
FB: Fernando Baptista, ONS and TC member. ONS is II (Instituto de Informática)
JPM: José Miranda, MULTICERT (portuguese certification authority), proposed Vogal, not verified, invited as expert by ANSOL

Notes

(RuiSeabra) I'm not a particularly strong note writer, so some information may still be missing, please help make it complete.

ITEM 1

14:37 MSD

Opens meeting by starting with item 1, review and voting of the minute of the previous session. The minute was approved by consensus.

ITEM 2

14:41 MSD

Proceeds to item 2, informing the TC that ONN approved TC 173 "Lin guagem de Descrição de Documentos"

ITEM 3

14:43 MSD

Proceeds to item 3, new vogals, almost all were accepted.

Provides official form for registration as member of TC.

MSD & CR

Propose new vogals fill the form while discussion proceeds to next item.

Demands this item to be kept open for there are doubts about new members that beg the questions:
1. Sun wasn't admitted entry for lack of space. Does ONN maintain this decision, and is it sensible?
2. Is this "rule" sustainable when there is more than one person per entity in the room? Is this in conformity with the rules?

States new entities weren't accepted because of
1. lack of space in the room (20 seating spots -- this isn't true, 24 people were seating at the beggining and there were still chairs in the room)
2. representativity had been achieved (this isn't true, eg universities, public administration, libraries, etc...), there were software companies, software associations, and even an individual.
The alternative would be to go to an auditorium where one wouldn't have the same conditions (an U disposition oriented towards the table with the ONS, and President and Secretary of TC).

One shouldn't sacrifice representativity because more than one person per entity would be present.
Considers the decision should be reviwed, demands this request be registered in the minutes.

Repeats his justifications.

States the TC should know who was admitted and who they represent so the TC can verify representativity.

MSD

States representativity is decided by ONN and ONS, so that can be postponed to the end of the day or another session.

MV & GH

Strongly oppose, representativity must be checked *before* decisions are made.

States there are 8/9 from public adminstration, 7/8 from software companies....

Interrupts, it's not just the number that matters, but who and which affiliations.

(lists too fast for me to note down)

Proposes the unknown element he couldn't identify should then be invited to leave the room.

Identifies subject as being Stephen McGibbon, IIRC, an expert in OOXML

Solicits that substitutes and expert leave the room, they occupy places that were best served for representativity.

ONN/ONS has the responsability, and ONS assumes it.

Asks what was the criterium for acceptance.

MSD

Proposes to discuss this at the end or at another meeting.

Clarifies that the criterium for acceptance was order of arrival of requests for participation as Vogal.
Agrees (with himself) that 20 is a reasonable number of participants, and that this stands to reason.

Demands that minutes record 24 people are currently present.

JPM

(enters the room as RS ends phrase, meaning 25 people present, now)

He didn't receive any answer directly from ONS, but since MV had previously listed him as accepted, he came.

If no answer was sent from ONS, then he wasn't accepted.

Justifies that adherence procedure wasn't clear at the begginning, previously though all was OK and communicated acceptances.
Then, after the confusion was known, failed to warn everyone.
As a compensation, nominates ME as expert for this meeting. Asks if the number of 20 people is fixed.

Clarifies that companies weren't chosen by their size.

Alledges (because it's not in the minutes, can't be proven) that in the last meeting he asked Microsoft to bring an OOXML expert.

JPM

Didn't receive answer from ONS either.

Is sorry for failing to reply to him.

Invites JPM as an expert.

Invites LS as expert for INE

Demands that minutes record 25 people are currently present.

Angrily states that only by delicacy they are not sending people away (not sure if he means substitutes and experts or if that was a veiled threat from ONS).

ITEM 4

15:00 MSD

Proceeds to item 4, validation of the following items of the agenda: 5 discussion of OOXML 6 vote on OOXML 7 other subjects, next activities

Asks for vote clarification: if one has to leave, can he delegate his vote? Also, since it's such an extensive subject, is it sensible to vote already? Proposes postponing vote to another session.

MSD

Vote by delegation is not excluded by CNQ directive, which is clear on the subject, one should have a substitute...

Without substitute, can one delegate?

Considers to be too premature to vote. Current doubt could be clarified later if vote was postponed.

Agrees with postponing.

MJM

Also agrees that no conclusions should be taken this day.

Agrees. Adds it would give the opportunity for new memebers to be proposed.

States ANSOL has 6 questions that need answers, and almost 100 technical comments, even giving 1 minute for each of them would result in almost two hours for one entity alone to state technical problems.
Believes it isn't reasonable to vote today.

MSD

Proposes to postpone items 5 and 6 (discussion and vote) to another meeting.

Proposes only item 6 (vote) should be postponed.

MSD

States items of discussion were approved on the previous meeting, and the decision was to have discussion and vote today.

Has 78 pages of summary points that were evaluated in the USA, doesn't think it is sensible to vote today.
Only knew he'd participate on the previous wednesday.

MSD

Predicting no consensus, proposes vote: "Vote on OOXML happens today?"
Of the 18 present vogals 5 voted YES, 13 voted NO.

MJM

Restates need to clarify vote in case of sudden absence.

MSD

Directive is omiss, best strategy should be for MV to nominate a substitute, and come to the meeting with him. MV should nominate a substitute soon.

ITEM 5

Reiterates three pillars for his opposition:
- form: proposed by one entity alone
- function: conflicts with ISO 26300, doesn't stand to reason to have more than one standard in a focused domain such as this.
- content: doesn't consider correct that an XML format contains examples that don't validade as XML in circa 10% of the cases, shouldn't contain date and math errors

Claims to have studied OOXML hard since last meeting (June 26th), there was more than enough time to study OOXML since then.
If ODF is an ISO standard than maybe at first there shouldn't be two standards, but OOXML is aimed at backward compatibility.
So what if OOXML has some errors, ODF also had errors, namely with SVG and MathML, but TC shouldn't focus on details.
Claims to have invited Microsoft to bring an expert in OOXML for finding 3 items of relevancy.

(NOTE from RuiSeabra to readers: PQ claimed previously to have requested the expert's presence in the previous meeting, supposedly before his "study" time)

First, VML on section 15.2.17 (Fundamentals). Spec says it's a legacy format. Having found DrawingML specification, is VML a minimum requirement?

VML is there because it's needed for millions of documents that use it.

Asks if MS Office 2007 can use VML.

MS Office 2007 uses DrawingML, which is fully XML.

Page layout, why does the spec separate it in partes, does any of it depend on the OS?

The spec splits page layout in two parts in order to handle independence of media display (print, screen, mobile phone, etc...), there is no dependency of registry information or anything like that.

What about autoSpaceLikeWord95, will MS Office 2007 use this in a new document?

No. It will not.

Why isn't it in a separate part, then?

It is an optional tag, only needed for backward compatibility.
One can get an older version of Word and find out how it is done.

States that there's nothing prohibiting new implementations from using those tags.

Agrees, but implementors could just warn about that tag being present, asking the user what to do.
OpenOffice also has similar problems, and you can only look at it to find out how it is done.
Claims to be a requirement from Library of Congress and the British Library.

This relies on information not easily available, which places Microsoft in a priviledged position, hurting fair competition.
Microsoft hasn't been up to today a good member, historically.

That's an emotional argument, Netscape had the tag <BLINK>. Since ODF is underspecified, Microsoft would need to make proprietary extensions.

They don't have to be proprietary. Why are proprietary extensions the only alternative for extending ODF for Microsoft's legacy support?

You can always check behaviour with older applications.
You can measure (makes an air drawing of a ruler) the spaces in Word95 and other behaviours.

Unlike the dismissal from the expert, IST believes the optional tags are a guarantee that they will be used and considers this a serious problem against adoption of the standard.
If this represents an unsolvable technical problem, then the standard itself is stating that it is neither open nor can it ever be fully open to legacy/existing documents (hence it is a bad standard as far as its openness). If there is intention to use the tags, then this is evidence that there is currently an intention to subvert the openness of the standard.

Take it to JTC1, they can drop those tags if they see it like a big issue.

There is no guarantee that these two bodies will be in agreement and work in consonance. Hence any promise from ECMA/OOXML regarding what JTC1 will do is necessarily not binding and thus irrelevant.
IST strongly objects such tags.

MSD

Proposes that a speaking list is used, so things go more fluid. Next in line PS, then JM, then JN, then GH, then RS, then ?3?, then JPM.

Feels difficulties arise from emotional arguments, asks for more moderation.

Points out that a speaking list prevents unencumbered discussion.
People may remember important things for the topic at hand that might be lost.

Congratulates Microsoft for the effort, for returning to the 80's when there were published specifications of Microsoft software.
His experience with standards tell him to worry about things like COMMON USAGE, where you need a spec that allows implementation.
Is estranged that a specification requires reverse engineering of old software which might not even be available in the market any longer.
Thinking this tags are not problematic because they are not required is irresponsible.
There are also other failures, namely in the respect for ISO standards.

You're from FSF, right? (probably advised by the Microsoft guy at his side, MS)

Refuses, states representation of his own company, Intraneia, which has support issues due to lack of respect from standard from Microsoft. Expresses worries about MS patent promise.

Claims Larry Rosen says it's the best patent promise there is.

What about the "clarification"? It says it's only for the "required" parts that are described in detail.

The required parts of optional parts are also covered.

Points out that it's not just Microsoft's sowftware that is referenced, but also other software. There's no patent promise for those.
Assuming that all additional information is available in some form, one need to consider the time to obtain it, as well.
States it's the common problem of information that is not in a specification, it can put "Time to Market" at stake, giving the opportunity for the dominant player to offer the solution faster and at a lower price when he's competing.
Points out that the patent promise can't be an ever changing document (last updated on the 9th of July), but it needs to be, according to ISO regulations, referenced in the beggining of the specification and safely stored at ISO offices.

Reminds TC that Microsoft claimed the optional tags are a requisite from Library of Congress of British Library. This means that for all practical purposes, this renders the tags mandatory if you want to do business with entities such as these, so argument of "being optional" shouldn't be taken in account.
Asks how many in the TC work with Dots per Inch, different types of fonts, different screen resolutions... The mere idea that one can go back to an older version of software and measure behaviour is absurd, a simple fluctuation of the aforementioned parameters creates different measures.
It was mentioned by the expert that OpenOffice has similar problems, that you also need to measure older versions.
This is a misleading argument, since Microsoft Office is fundamentally different from OpenOffice, in that Microsoft Office can't be measured.
In lack of a specification, the only other way to properly analyse software is by studying at its source code, which is impossible for Microsoft Office.
It's only possible in special circunstances, but then you can't implement what you learned in a cometing office suite, which renders this possiblity into an impossibility from the point of Free Software.
OpenOffice, however, uses the GNU GPL, the most used Free Software license, which gives you the right to study the software, and provides source code availability.

16:43 MSD

Calls for a break, since the scheduled break is 15 minutes overdue.

DURING BREAK:

SM talks o RS Asks why RS is against a format that can be implemented with Free Software.
RS answers that it is a closed standard when it depends in practice of hidden information.
He asks for one example, answer is autoSpaceLikeWord95.
The problem with tags like these is easy to disappear, all Microsoft needs to do is to add to spec: it is done like this ... just that.

BACK FROM THE BREAK:

MSD

Speaking list is now: MJM, JM, JPM, SM

Is OOXML proposed by only one company, namely Microsoft?

Refuses, it's proposed by ECMA, with lots of companies, only IBM voted against.
Claims Novell is a competitor and is implementing OOXML.

MJM

Asks what the advantages are, from the user point of view?

A whole new range of apps, you can save directly to on-line services, etc...

JPM

One other worry from the point of view of busineses is the dematerialization of documents.
In Portugual, whoever deals with Electronic Receipts needs to store them from 10 years, which has high costs for companies.
Does OOXML help in reducing these costs for AP and companies, and does it allow them to take advantage of this? Still has no answer for this.
Also, OOXML uses proprietary hash algorithms, which is also a problem, since they can't be proved safe.

In order to get a full implementation, one needs the Covenant no to Sue. Has it been analysed in light of Portuguese law?

No need to sign it.

Why is it there, then?

The Covenant and the OSP are there because of Microsoft's defensive patentes.
Microsoft basically copied SUN's promise.

Proposes that TC requests a juridic evaluation of these promises.

MSD

Agree that a juridic evaluation should be requested.

ECMA holds copyrights and patents over the specification, but the promises don't cover future implementations.

The promises are critical since ECAM dosn't cover future implementations.

Points out that there are no proprietary hashes, that OOXML uses SHA-1, SA-256, etc...
Points out that optional tags are optional, so there shouldn'be an issue.

If it's not an issue then it has no place in the specification.
However, SM said it was a requisite issue from the Library of Congress and the British Library.
How can smaller entities like us consider it isn't an issue if it is a big issue for them?

Points out that not proposing OOXML because there is another standards is a weak reason since OOXML is much more complete.

OOXML doesn't respect ISO standards like countries and languages.
If a new language is added, two standards would hae to be reviewed.
It's not the only example, there are many simiar to this, like colors or mathematic formulas.
Question to Stephen: Why not extend ODF?

ODF doesn't aim for backward compatibility, like OOXML does, with millions of exsting documents.
Doesn't see any problem with co-existence.
OASIS rejected Microsoft's proposals and expelled/excluded the Microsoft representative because the proposed changes would add backward compatibility to legacy documents.

MSD

Points out that only 4 minutes remain for this agenda item.

MCM

Reminds that since item 6 (vote) won't happen, one can add some more time.

MSD

Grants 5 extra minutes.

The existence of a previous standard should not be used as an objection in principle since standards evolve and new versions also eventually become standards. But the fact that a standard already exists raises the bar for newcomers and necessarily implies that any potential new standard proposal must, in addition to all other requirements, clearly show, beyond any doubt, that it brings considerable added value to the table.
Does not see this discussion anywhere.
Nonwithstanding promises for correction of the problems mentioned, the decision must be made based on the currently existing version and not a virtual (promised) version after such changes are made.
If some changes are considered essential, the deliberation should state that clearly and not assume that those changes will be incorporated.

Optional tags must be solved.
If there are times when they are mandatory, that must be clarified.
Between having two standards, where one fails to fulfill the objectives, or one single standard that does it well but less complete, better the later than the earlier situation.
The experience with HTML, ACPI and other standards lead to the conclusion that someting more objective and concrete is preferable.

Still hasn't understood why ODF wasn't extended.

MJM

Why not improve what already exists? We're opening the doors to the risk of creating a multiplicity of standards, which isn't an advantage.
It's good that OOXML is backward-compatible with Microsoft softwae, but what about the rest?

Jests that one can always create a third standard.

MSD

Starts closing item 5.
Refuses to let RS speak, who was asking for permission to speak for some time.

States that a louder voice can be used if necessary to be heard.

MSD

Postpones intervention of RS for end of item 7.

ITEM 7

(remember that item 6, the vote, was postponed)

Atacks RS claiming lack of respect. Recommends ONS to only accept at most 20 elements in TC

Supports RB Claims it's associates are the main stake-holders.

MJM

But user interests are not being taken in account, just the commercial interests of some of the companies here present.

The specification may have some problems, but we don't need to throw it all ways, corrections can be inserted.

Supports RB.
Agrees that a juridical analysis must be done.
Proposes that a definition of the National Interest should be created, since TC is there to discuss taking in consideration the national interest.
Proposes to create this definition on the next meeting.

Doesn't oppose the number of 20. Points out that he opposes the nature of the current 20 elements, it has too many commercial interests, and too little from other important players like universities, libraries, and other public administration.

IST will not comment on the current membership of the CT, but notes that they represent only a very small portion of the interested parties in the matter being discussed and the consequences of the deliberation to be reached.
Such parties include software vendors and developers, but also many other companies that make extensive use of electronic documentation, the public administration, universities, libraries, etc, etc.
Hopes TC will better represent citizens and the state of Portugal, way beyond the commercial interests of the present companies.

MSD

States that TC has representativity, ignores request from RS to speak, after insisting refers to after the vote.

Declares this is the second time the President tries to cut his expression, won't allow it a second time.
Rejects the so called representativity, since important members of society aren't present, like libraries, universities, and other public administration entities.

Asks wether this vote is valid

Declares that the directive is omiss.
States that the constitution of the TC is not closed to new vogals, but that mobility should be achieved by swapping entities, like the two entities that are not present if they don't justify their absence.

MSD

VOTE: maintain only 20 elements? YES 9, NO 7

MSD

Proposes 31st of July as next meeting date, on the same place.

Agrees.

(there is a consensus on this date)

MSD

Proposes the agenda items are:
1. discussion of OOXML
2. vote on OOXML

Proposes hat the President collecs the list of possible votes and their explanation, for the next meeting.

MSD

Agrees. Tries to close item 7. Gives permission to talk to RS on item 5, as "promised", after RS insists.

About item 5, which is now about to speak out of context, has to add that the coexistence of two standard does not add usually add choice.
Cites example of european and US power plugs, which are incompatible, although both of them a standard.
If one doesn't take the care, or even know about this issue, he will suddenly view his "choice" reduced from ONE to ZERO forms of powering his laptop.
Declares that either Microsoft should work to extend ODF, or implement OOXML in a way that it fully includes ODF.

Considers the presence of the expert was important, and invites him back, if possible, for the next meeting.

Agrees.

MSD

Those who don't have substitutes should nominate at least one until the next meeting.
Closes this meeting.