a 'mooh' point

clearly an IBM drone

The actual work we did in Prague

I thought I’d try to outline a bit what we actually did and what constituted our work in Prague.

The agenda framing our work throughout these three days was this:

  1. Opening 2009-03-24 09:00
  2. Roll call of delegates
  3. Adoption of the agenda
  4. Schedule for publication of reprints or Technical Corrigenda
  5. Defect reports
  6. Future meeting (face-to-face and teleconferences)
  7. Any other business
  8. Extension proposals from member bodies and liaisons
  9. Conformance testing
  10. Closing

The vast majority of our work was in item number 5 on the agenda and each and every single minute was used discussing the defect reports – including in lavatories, on our way to work, on our way back from work, during lunch, dinner, breaks and drinks … in short – we discussed DRs 24/7. This was as it was supposed to be – this was really the reason for all of us being in Prague.

The initial list of DRs we discussed was this (just to give you an idea of what we talked about):

08-0001 — DML, FRAMEWORK: REMOVAL OF ST_PERCENTAGEDECIMAL FROM THE STRICT SCHEMA
08-0002 — PRIMER: FORMAT OF ST_POSITIVEPERCENTAGE VALUES IN STRICT MODE EXAMPLES
08-0003 — DML, MAIN: FORMAT OF ST_POSITIVEPERCENTAGE VALUES IN STRICT MODE EXAMPLES
08-0004 — DML, DIAGRAMS: TYPE FOR PRSET ATTRIBUTES
08-0005 — PML, ANIMATION: DESCRIPTION OF HSL ATTRIBUTES LIGHTNESS AND SATURATION
08-0006 — PML, ANIMATION: DESCRIPTION OF RGB ATTRIBUTES BLUE, GREEN AND RED
08-0007 — DML, MAIN: FORMAT OF ST_TEXTBULLETSIZEPERCENT PERCENTAGE
08-0008 — DML, MAIN: FORMAT OF BUSZPCT PERCENTAGE VALUES IN STRICT MODE EXAMPLE
08-0009 — WML, FIELDS: INCONSISTENCY BETWEEN FILESIZE BEHAVIOUR AND EXAMPLE
08-0010 — WML: USE OF TRANSITIONAL ATTRIBUTE IN TBLLOOK STRICT MODE EXAMPLES
08-0011 — WML: USE OF TRANSITIONAL ATTRIBUTE IN CNFSTYLE STRICT MODE EXAMPLE
08-0012 — SCHEMAS: SUPPOSEDLY INCORRECT SCHEMA NAMESPACE NAMES

I think it’d be fair to say that we have come a long way since the time we were discussing if it was possible to use XSLT to simulate bit-switching or if an OOXML-file was “proper XML”.

For each of the DRs we covered we discussed if the DR was a technical defect or an editorial defect, what the possible implications of the DR would be to existing documents and existing implementations and if the DR belonged in a corrigendum (COR) or if it was an amendment (AMD). It was quite tedious work, but we managed to cover quite a lot of ground in the three days.

Corrigendum or amendment?

One of the first things to accept when working in ISO is that there are quite the number of rules to comply to. As it turns out, it is not our prerogative to decide if a DR goes into “the COR bucket” or if it goes into “the AMD bucket” – there are rules for this. The ISO directives section 2.10.2 state that

A technical corrigendum is issued to correct [...] a technical error or ambiguity in an International Standard, a Technical Specification, a Publicly Available Specification or a Technical Report, inadvertently introduced either in drafting or in printing and which could lead to incorrect or unsafe application of the publication

If the above is not the case, the modification should be handled as an amendment.

Still, there are quite a lot of DRs that fall into the more gray outskirts of this definition. So to facilitate our work we made some guiding principles, and these principles were discussed at the SC34 plenary in Prague:

[…] in the interest of resolving minor omissions in a timely fashion, WG4 plans to apply the following criteria for deciding that the unintentional omission or restriction of a feature may be resolved by Corrigendum rather than by Amendment. All of the following criteria should be met for the defect to be resolved by Corrigendum:

  1. WG 4 agrees that the defect is an unintentional drafting error.
  2. WG 4 agrees that the defect can be resolved without the theoretical possibility of breaking existing conformant implementations of the standard.
  3. WG 4 agrees that the defect can be resolved without introducing any significant new feature.

Unless all the above criteria are met, the defect should be resolved by Amendment.

Of course we will still have to do an assessment for each and every DR we look at, but it is our view that these principles will help us quite a bit along the way and to have a more expeditious workflow. Notice also the wording “WG4 agrees”. A very small number of DRs falls clearly into the COR- or AMD-bucket, so it is not possible to regard these principles as a mere algorithm with a deterministic result. The principles requires WG4 to agree to the categorization of DRs so we’ll actually have to sit down and talk everything through.

On the first day (or was it second?) we also touched briefly upon the subject of modifying decisions made at the BRM. The delegates at the BRM were nothing but normal people, and due to the short timeframe of the meeting, errors likely occurred. At some point or another, someone will discover we made a mistake and put a DR on our table. At this point we will have to figure out if we think the decisions made at the BRM are now cast in stone or if they should be treated by the same criteria as the other DR we receive. As I said, we just touched upon the subject and didn’t reach any conclusions to this. If you have any thoughts regarding this, please let me (and us) know. My personal opinion on this subject is, that we in WG4, at this point in time, should be extremely careful when thinking about reversing decisions made at the BRM.

And finally, I thought I’d give you some pointers about what is in the pipeline of blog entries (I don’t have a sophisticated system as some people, so I’d just give you a small list of topics at the top of my mind these days:

  • Markup Compatibility and Extensibility
  • Conformance class whatnots
  • Namespace changes and the considerations about doing it or not
  • Why should we care about XPS?
  • Why I like the ISO model
  • Maintenance of IS26300 in ISO

 

Maintenance of IS26300 in SC34

The streets of Prague are buzzing with rumours coming out of the work in the working groups of SC34 and SC34 itself as SC34 is currently having its plenary meeting in Prague.

It seems that SC34 has done the only clever thing to do - to create an Ad Hoc Group (AHG) to have responsibility of maintaining IS26300. I applaud the decision to do so, and it has in my view been a long time coming.

The details and scope of the group is yet to be seen, but I am glad that SC34 has chosen to create it. There is only one entity responsible for maintaining ISO standards, and that is ISO. Maintenance of IS26300 has falling between two chairs at the moment, where WG1 was initially responsible for it, but it has been preoccupied with other tasks. Also, I think the maintenance agreement of IS26300 has been mentally prohibiting any work being done.

The upside of this is that there is now a group in SC34 responsible for receiving defect reports submitted by NBs. One group is responsible for preparing reports to OASIS and the get the responses back in the ISO system.

This is a clear improvement and it is a sign and a statement that we believe that IS26300 is too important to not have a group responsible for its maintenance in ISO.

Smile

WG4 meetings in Prague

Wow – this has been a tough week. I arrived at the hotel here in Prague (I am currently waiting in Prague Airport for my flight back to Copenhagen) at around 21:00. I met Doug in Copenhagen and flew with him to Prague and in the airport we ran into Kimmo. After 15 minutes in my hotel room I went down to the bar to get a “welcome to Prague”-beer. After another 15 minutes I crawled back to my room completely devastated due to a flu I hadn’t been able to get rid of. 5 seconds later Florian called and ordered me to get my ass down in the basement wine-bar where he was having drinks with Doug and Megan. I went back to my room when the bar closed at around half past midnight, did some last-minute updates/tweets and almost cried myself to sleep because of near-death-like fatigue.

… and the meetings hadn’t actually started yet.

The next morning the meetings started with a joint session between WG4 and WG5 at the Czech Standardisation Institute. A total of 31 delegates attended this initial meeting. Apart from the SC34 officers (SC34 chair, SC34 secretariat, WG4 convener), there were delegates from Canada, China, Czech Republic, Denmark, ECMA, Finland, France, Germany, Korea, Norway, South Africa, UK and USA. We had quite a lot of work on our table for these three days, and we immediately got to work after the initial pleasantries. A rough list of categories to be dealt with was “Defect reports”, “Rules of engagement” (or “Prime directives”), “Future work”, “Roadmap for future editions/corrections” and “Planning of future meetings and tele-conferences”.

If you’ve been following my twitter-feed (and the ones of Alex, Doug an Inigo) you’ll already have a notion of the insanely interesting things we talked about. But for those not following me (and you should!!!) we talked about sexy things like whether “named ranges” in spreadsheets were defined on the workbook-level or the worksheet-level, whether a reference to Unicode 5 implied dependencies of XML 1.1, whether xml:space applied to whitespace-only-nodes or just to trailing- and leading whitespace in element content, whether font-substitution algorithms in OOXML had a bias for Panose-fonts and if “Panose” really meant “Panose1” and suttle differences between the Panose-edition of Hewlett-Packard and the one of Microsoft (as far as I understood it, anyway)

Can you imagine all the fun we had?

And you know what? We didn’t stop talking about it during lunch, dinner nor brakes. As Doug noted in one of his tweets, the only difference between session and breaks was that during session, only one person talked at any given time.

Well, apart from all this fun, we made an enormous amount of progress. A total of about 169 defect reports have been submitted to us until this point, and we processed almost all of them. We didn’t close all of them, but we managed to process the most important ones and prepare ourselves for our first tele conference in mid April. We laid down some ground principles upon which we will make decisions in the future and we talked about a set of “Prime directives” to form a mental basis for our work (think: The three Laws of Robotics).

In short – it was a good week. I’ll post a series of blog posts in the next weeks outlining the results we achieved (and did not achieve) including both the extremely boring ones as well as the more controversial ones. So Watch this space …

PS: I almost forgot. Microsoft sponsored a dinner/buffet for the participating experts on Wednesday. But what was even cooler was that they had lined up a bunch of Ferraris and Lamborghinis for us outside the restaurant, and we could just take a pick to choose a car to take home. Mine was red! Is that wicked or what?

To the nitwits from <no>ooxml.org: Take it home, boys!

OASIS to JTC1: Bye, bye ...

Ever since the hoola about OOXML-approval there has been quite some discontent in the ISO community regarding how ODF TC has fulfilled its obligations after IS26300 approval. A few meetings have taken place to "amend the harsh feelings" and now some preliminary results have been sent to the NBs for consideration. For those with ISO privileges the documents [1], [2] can be found in the SC34 document repository.

There has been a lot of debate as to where maintenance of ODF should take place, be it in OASIS via ODF TC or via some construction as with OOXML, where the originating TC is included (assimilated) into SC34 and maintenance and development takes place there. I really don't care where these activities take place. I just want the best qualified people to do it.

Now, the documents deal with a definition of principles and a more specific definition of "who takes care of what?"-items. When reading through the documents, I couldn't help getting the feeling that what OASIS was essentially telling JTC1 was "It's my way or the highway".

JTC1 and OASIS have come to the following agreement around maintenance: 

  • OASIS ODF TC takes care of maintenance and development of ODF. 
  • National body participation in this work is encouraged to take place in ODF TC by either direct membership, via the "Comment mail list" or via TC Liaison (I didn't know JTC1/SC34 had one of those in ODF TC)
  • OASIS will submit each approved edition of ODF to JTC1/S34 for approval to make sure that approved standards are equivilant.

I completely agree on item 1) and 3) above, but item 2)? In the paper there is not a single sentence on how the procedures in JTC1 fit into all this. Why are there no wording regarding voting procedures in SC34? If ODF TC comes up with something new and "substantially different", it should be submitted using the "PAS submitter status" of OASIS (similar to the Fast track procedure ECMA used with OOXML). But a PAS submission requires voting in SC34 and if the vote fails (or substantial concern is raised), a BRM is scheduled. If the comments are fixed, the result of the BRM will be an "errata-sheet" and a new vote takes place.

Suppose the post-BRM vote approves the submitted ODF edition

  • what will OASIS do with the errata-sheet?
  • what are the principles for getting them back into the OASIS-approved edition of ODF?
  • what is the time frame?

Is the truth really, that OASIS doesn’t want JTC1/SC34 to do anything to ODF but rubber-stamp it when it comes our way?

When the original ODF 1.0 was submitted to JTC1, a maintenance plan was agreed upon. It had two small but really important words in it: "as is". The maintenance agreement said (AFAIR) that JTC1/SC34 was expected to approve future editions of ODF "as is". In other words, what OASIS managed to get JTC1 to agree to was essentially: "Don't look at it, don’t' open it, don't flip through it, just - don't touch it. Get a hold of the ISO-approval stamp, stamp it and send it back to us".

The only possible conclusion is that OASIS does not want any direct ISO-involvement in development of ODF.

That is fine - the ODF TC should do what they find best. But I am wondering if that also means, that OASIS will not send future editions of ODF to JTC1 for approval? Surely, OASIS can't live with the reputation of having their standards simply rubber-stamped by ISO? 

You may also ask why it is not good enough for JTC1-members to contribute to ODF through ISO. Well, OASIS is a vendor-consortium and the interests of the vendors seem to be somewhat different than the interests of the national bodies. If you look at the contributions of Murata Makato and Alex Brown through the ODF Comment list, it is clear that their interests in quality in schemas, constructs and the specification itself was not prioritized in the TC at all. To me a mix of vendor interests and national bodies is the best way to ensure high quality in any specification, but the proposed agreement between JTC1 and OASIS seems to cut out the national bodies acting as "national bodies"

I think it is a good idea to ISO-approve ODF in the future. But JTC1 needs to send a clear signal to OASIS saying, that is it fine that they want the “Seal of ISO” and we welcome them. But in order to have the cake, OASIS must eat it too. The ISO package must come with two items, 1) the ISO quality stamp and 2) national body involvement. You cannot just have the stamp! It should be emphasized that it is the prerogative of the national bodies to process the standards that come their way and that cutting them off and have them do nothing but rubber-stamping the specification is completely unacceptable.

The proposed maintenance proposal will be discussed at the JTC1/SC34 plenary in Prague on Friday, and I hope all national bodies have understood the ramifications of approving the maintenance agreement. I suggest the plenary responds by saying to JTC1/OASIS: "Thank you for your suggestion for a maintenance plan for ODF, but come back again when we as  national bodies have a solidly founded role in the maintenance of the specification".