This is an open letter which I would like to be considered in the Standards Australia review of the ooxml proposal. I have not been sent information by Standards Australia about the required format and would be happy to adapt this material to a format when provided.
While I appreciate the original scope of the process around technical concerns I also am writing about the proposal process and feel that the primary problems with the proposal are process related. I am writing more broadly because I feel the broader concerns are important to our Standards organisation and would avoid similar proposals using the same process.
Overview: Double standards are counter productive
In this issue there are two kinds of double standards. The proposal and its counterpart existing ISO standards, and the open or closed processes which the different projects have followed.
Microsoft's OOXML replicates functions covered by the Open Document Format standard and other standard XML formats. Duplication of standards reduces the value of the original openly developed standards.
Microsoft supporters in the Australian meeting suggested that "OOXML was enough of a standard, that we should expect there would always be proprietary material in Microsoft formats, and that their clients understand this."
The MS OOXML proposal is not 'enough of a standard' to make it useful for all developers and users because it is undefined both legally and technically. This will devalue a standards based approach in information formats.
The process for the OOXML format has been closed in its development phase and structured to reduce meaningful review. The result is a vendor centric document which is inward looking in its implementation.
Standards usually are developed to enable consolidation of current best practices and they need to be written and legally framed so that they are reliable and safe to use. ODF provides this function. OOXML undoes that work by approximating a standard but being sufficiently noncompliant to provide data compatibility problems with material
matching existing standards.
I feel that standards organisations need to take a stand to recognise the value in the work they have done on ODF XML SVG and to stand firm to support them.
In each case where there is a multiple standard there is a cost to the wider community in accommodating multiple approaches to the same function space. Forking the market into two separate standards to cover the same material is not a step forward.
The market in documents has been vendor oriented for many years. Open Document Format was developed as a common format two years ago. The purpose was to agree on a way to make information accessible regardless of vendor.
The effect of the OOXML proposal is undo that commonality.
Document specification is a core requirement in our information and computing based economies and therefore the cost of multiple standards in this fundamental space is an expensive outcome.
In any case where an information format is as fundamental as this there will be increased temptation for vendors to try to find ways to gain control of the space. ISO has approved a standard which has worked through this space and is continuing to develop and refine functions. Introducing a vendor based standard at this point would
compromise all that work in negotiating common ground. While the market might include many products which match specific vendors, the role of the standard in that space is to share information regardless of vendor. A single vendor standard on the other hand operates as a kind of franchise over a sector of the market.
From Andy Updegrove:
"The best reason for not approving OOXML/Ecma 376 as a global standard is that it will encourage other vendors to push for multiple, unnecessary standards rather than achieving consensus on a single standard that will best serve the needs of all stakeholders, and not individual proprietary vendors."
The Australian Commonwealth memorandum of understanding with Standards Australia is certainly written around these kinds of values related to public benefit and national interest. I understand that in negotiating a useful and meaningful standard there will be competing and challenging choices to make and that there has been a move toward accepting plural standards for different vendors, possibly to reduce
contention in these processes. However I feel this move shifts the contention into the marketplace and reduces the value of standards to the wider community.
Existing standards which are contradicted by the material included in the OOXML
- ODF (document xml)
- OOXML doesn't use the stablished SVG (vector based xml) but on the
contrary two private and exclusive formats, one of them legacy: VML and GraphML
- OOXML doesn't use the stablished MathML (mathematical markup) but its proprietary format (already rejected by scientific editorials which make deep use of that feature for Nature and Science)
- OOXML doesn't use the International date and time (ISO 8601), the basic of all economic, legal and political transactions in world. On the contrary uses two calendars, one limited to start recently, in 1904, and other the one used by the Roman Empire thousand of years ago, the Julian Calendar.
Rather than passing a proposal which conflicts so fundamentally with standards practice Microsoft should be encouraged to participate in constructive collaboration on the development of those projects. This would better serve the standards process than initiation of a second set of materials which non-Microsoft people would have trouble interoperating with, and which would undo our ability to share information using conforming standards.
OOXML is not functionally open
The name of the proposal OOXML is Office Open XML, this is a misleading name as the proposal is not fully open, its development has not been open, process for new versions is suggested to be through ECMA's closed process, the legal access to "enough rights to implement the standard" are not clearly provided. Wording
inside the proposal includes 'wordwrap like Microsoft97' which is subjective and unimplementable language and would be unlikely to pass in a proper ISO process. The format includes proprietary binary material which is not open.
This is perhaps a work which contributes to better documentation of existing Microsoft formats in terms which are useful for Microsoft related developers but it is not outward looking or well integrated with ISO standards which it is likely to interface with.
Clients who believe they are moving from a closed format to an open format will find it is only comparably more open but is not drafted in a way to make it transparent or easy to avoid vendor lock in.
Making this a standard in my opinion makes the standards bodies complicit in duping customers into adopting a format which is not truly representative of the criteria which are normally recognised as ISO.
The value of formal standards
Standards are increasingly important as our economies and information becomes more transnational. We cannot afford to be imprecise with data working at the scales which
are now possible. Working with live transactional data would be a situation where
explicit single outcome processes would be most important. Mapping those sensitive projects to a format which is built on a single vendor retrospective is not a recipe for safe eCommerce.
"[A] document, established by consensus and approved by a recognized body, that provides, for common and repeated use, rules, guidelines or characteristics for activities or their results, aimed at the achievement of the optimum degree of order in a given context NOTE Standards should be based on the consolidated results of science, technology and experience, and aimed at the promotion of optimum community benefits." -- ISO/IEC Guide 2:2004, Definition 3.2
The value of formal standards is to find a best practice approach to common agreed functionality. To use that as a unifying reference point in order to ensure
developers of systems have a way to map to an exact and accessible reference. And that customers of products are able to choose from solutions by a range of vendors and to move their information from one to another without prejudice. This makes it possible for multiple providers to participate in a market and for consumers to make
flexible choices about the formats and applications which suit their needs on an ongoing basis.
I feel that Australian and other standards bodies need to take this opportunity to review their function in society and to consider that the point of difference they offer is to contribute a means for finding negotiated best practice. This is a craft in the truest sense. It involves people developing skills in understanding systems objectively and in working through issues to find a useful and reliable outcome.
Companies which have a large market share already enjoy a kind of default standard status through cost to change, skill base or in some cases through data lock-in. Formal standards processes need the ability to attract quality participation in collaboration of standards development by using processes which do not marginalise collaborative processes and goals in favour of competitive processes. ie. Without a
focus on common interest and collaboration the formal process will only be able to mirror market share and will not be able to move Australia forward in a flexible and broadly useful manner.
ECMA fast track process
The ooxml proposal is a very large document(6000pages) and has been pushed through in 6 months. It has been suggested in the Australian meeting that this made it impossible to review. The volume/time compares poorly with usual standards which are 50 pages and take approximately 2 years to review thoroughly. ODF is 1000 pages itself and is considered large for a standards document but at least it was processed in an open way and there was room for all parties to contribute to its definition as well as a standard process of review.
The ECMA fast track process is structured in such a way that it does not honor aspirations of public benefit and national interest. The value that ECMA offers its vendor customers is to "offer a path which will minimise changes to input specs"
The process is a product offered to vendors by a provider who benefits from being able to push through many projects. This means the original relationship between the vendor and the ECMA organisation is based around momentum and single source material with no benefit to either party in providing real and rigorous review or participation in development. This together with the idea that multiple vendors offering vendor specific 'standards' is a workable approach are the underlying
assumptions which derail the rest of the process.
While it is expected that there may be different approaches to technologies and products, it is the role of a standards group to identify points of best practice around which an industry can integrate and offer flexible choices to consumers. This relies on access to a process where the craft of standards writing is core, where the considered and 'hands-on' input of many participants in the given standard is facilitated and where the proposal is thoroughly tested for its interoperability with existing standards.
Open review and participation in developing the standard is the means by which unifying and consolidating outcomes are made. This is the role which all groups interested in long term safe access to knowledge look to standards bodies for.
Australian process and responsibility
I sense a 'good faith' spirit in the efforts of the Standards Australia staff.
However I feel that the fast track process applied to this proposal, both in
Australia and internationally, is structurally not able to deliver outcomes in the Australian and public best interest. We need to address these procedural issues prior to processing this proposal so that it can be considered in a way which does address the goals of the Commonwealth in applying these standards processes to proposals.
Difficulty in finding out information about the process has not helped, including meeting information and formats for responses.
Australia as a country with an important role in these proceedings should not take an uncritical role in these processes. There needs to be a method for identifying when a new approach to processing standards is disfunctional. It is in Australia's national interest to understand the core function that open standards play in enabling all businesses and organisations to safely participate in working to or adapting from a standard format.
Accessibility and accuracy are prominent requirements within the education sector.
As James Dalziel pointed out in the Australian meeting, Australia leads on technologies for the education sector. Access to knowledge is an increasing piority and being able to share information developed with public funds is becoming an important issue. Organsiations which are able to develop once and distribute the value across the sector offer efficiencies which are valuable.
Australian companies and developers would find it expensive to accommodate two complex and incompatible approaches to documents requiring two approaches to each project they deliver. We need to be more sophisticated in our choices because we need to be able to make better value from smaller projects. Australia has many smaller players which means we need to be able to do things in a develop once use flexibly kind of way. Developers who produce technologies which help with accessibility also need to be nimble and flexible to cater effectively to small markets.
The University of Toronto recognises these issues "This paper undertakes a preliminary analysis of the OOXML format with respect to its accessibility, with emphasis on accessibility to persons with disabilities. We will demonstrate that the OOXML format fails to adequately support accessibility of documents."
Work on translating the legacy document formats into an xml based format is a worthwhile project. If Microsoft have a reliable tool for mapping material out of those formats into parsable XML that is a great step forward for their clients and for the ongoing relevance of their applications. Unfortunately it is my understanding that the technique for translating the documents involves embedding binary material in the XML.
OOXML is also a new-'old format', the time, date, vector, math models used in the new format are still reflective of historical choices and do not represent current best practice. In order to truly be facing forwards the material from the closed formats should be parsable XML and to match existing ISO standards. If the OOXML approach is not able to deliver on these outcomes then it is a great step forward from
opacity to translucency but is not the full journey.
Security, transactions and XML with embedded binary material
Transactions are not a point of difference to existing standards for OOXML.
Transactions are the normal function of XML and the specific focus of EBXML.
The OOXML approach to formats includes binary material.
Embedding binary material in transactional secure XML is likely to cause security issues. The transactional material has been developed only to work with MS
Internet Explorer which is the browser of choice for internet based exploits. Designing a transaction solution around embedding unknown content in an eCommerce context, to match only a singler vendor's formats and applications would be unlikely to pass a regular standards process.
Australia's Standards organisation should vote no with comments.
The comments should reflect our concerns with the principles and process of Fast Track standards development using ECMA.
The proposal should be put on hold until a review of the goals, responsibilities and processes for formal standards are completed.
This to include establishing a standard independently known and safe scoping for legal safety of any material which is ISO accredited. We cannot rely on unseen information and vendor assurances. There should be a known legal scope for all ISO standards.
The proposal could then be reconsidered, including whether a second standard in this space is a useful project for good standards practice.
These factors to be considered first before any further work and consideration is given to working through the body of the proposal.
It is possible that after reconsidering the processes and overlap between existing standards and OOXML that it might be possible to use aspects of the translation from legacy formats to map to Open Document Format and to translate from other non standard models to match other ISO standards.
From that point forward it could be also useful to look at the way Microsoft have worked with transactions, to compare those with existing standards and practices and to see if there is any scope for finding unity in those approaches.
6000 pages is too large. Breaking the proposal into
- retrospective work mapping to ODF or other standards.
- new work starting from an ODF and ISO standard compliant base would help to make the size of the standard less of an obstacle for implementation.
Specifications for non standard implementations of vector, math, time
and date, could be included in the retrospective work but should not
need to be a part of the forward looking standard format.
Further to the action on this proposal and on the international Fast Track process I would suggest that Standards Australia, and if applicable other nations, review their own processes and advice around the role of participants in the review process.
The memorandum of understanding between Standards Australia and the Commonwealth is written in terms of public benefit and national interest. This contrast with the participant information on the Standards Australia website which talks about representing specific groups and their own interests. This generates a tension between the overall goal of developing unifying standards in the national
interest, and a process which encourages people to view the process as
a partisan practice.
The process for any standard being considered should be clearly available on the website. The phases and formats required for public or industry participation
should be available throughout the process. Criteria for evaluation of a good standard for a given function should be identified and agreed in order to give a focus to the conversations. Existing standards which may overlap should be identified. If the criteria are listed and available this makes it far more likely
that people will be able to anticipate missing concerns or concerns which are not core. It provides a purpose which enables collaboration as well as competition around meeting known objectives.
I feel it would be useful to provide some workshops or communication around what kind of skills resources and time are required to participate. ie Australia needs to consider how it resources the process of standards development in order to produce outcomes which have vision beyond vendor interests.
I value very much the role that standards bodies have to play for our innovative and social function and hope that Standards Australia feels that OOXML is NOT "enough of a standard" and that while it might be true that Microsoft formats will always have proprietary components, that kind of fancy footwork is not compatible with ISO and Standards Australia branding.
Thankyou for the opportunity to participate.