| 
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Files spread between Dropbox, Google Drive, Gmail, Slack, and more? Dokkio, a new product from the PBworks team, integrates and organizes them for you. Try it for free today.

View
 

FrontPage

Page history last edited by PBworks 12 years, 9 months ago

 

 

Dear all,

 

As per the request, I have created at another wiki for wordprocessingml

 

regards,

 

Dr.P.Madhav

Institute for Electronic Governance

Government of Andhra Pradesh 

 

 

Dear All,

 

All minutes of the meetings and related documents have been uploaded in the page "Minutes of the Meeting".

 

Regards

Jaijit Bhattacharya

 

 

 

WordProcessing ML Sub-Committee

 

 

 

Center for Excellence in E-governance

Department of Management Studies and Department of Computer Science and Engineering

Indian Institute of Technology, Delhi

 

 

 

Minutes of Meeting held on April 04, 2007

 

Topic

First meeting of the WordProcessing ML Sub-Committee

Venue

Syndicate Room, Department of Management Studies, Vishwakarma Bhavan, IIT Delhi, New Delhi 11 0016

Date and Time

5.00 PM – 6:00 PM, April 04, 2007

Attendees

  • Ms. Geetika Sharma, TCS

  • Mr Jaijit Bhattacharya, Center of Excellence in E-governance, IIT Delhi

  • Prof. Rajanish Dass, IIM Ahmedabad

  • Ms. Anuradha Sen, Sun Microsystems

  • Sourav Suman, DMS, IIT Delhi

  • Dr. M. P. Gupta, Department of Management Studies, IIT Delhi

Regrets

  • Ms. Reena Garg, BIS

  • Mr. Vijay Kapur, Microsoft

Agenda

5.00 -5.05 PM Introduction to WordProcessingML

5.05 -5.15 PM Suggestions from attendees on way forward

- Definition of Open Standards

- Process for conducting the sub-committee

- Any other issues

5.15 -5.45 PM Detailed analysis of Part 1,2,3.

5.45 -6.00 PM Resolution of Issues and Division of Work

 

Minutes

 

  • Ms Sharma noted that we need to ensure that the proposed standard is royalty free in perpetuity and cost-effective to implement if India has to support it. She also emphasized that it should be an open standard and that all its supporting standards should also be open standards, i.e. It cannot be using an IT standard that is not open standard.

  • Mr Bhattacharya emphasized that any standard needs to have atleast one verified implementation before it can be adopted as a standard otherwise the standard will be an academic activity with no value addition to the country.

  • Regarding the process of conducting the sub-committee, Ms Sharma proposed the following:

    - the list of the core committee members should be defined.

    - in a meeting if 50% of the core committee members are present then the quorum should be recognized.

    - notice of the meeting should be given be given atleast 48hrs before the meeting and regrets should be communicated atleast 24hrs in advance.

    - confirmation of participation will be considered as attendance and hence the quorum complete.

    - no response will be considered as confirmation.

Ms Sharma's proposals were unanimously agreed by the sub-committee as such a ground rule is required in order to analyze the 6000+ pages of the proposed standard within the 5 month period.

  • Prof. Gupta suggested that the meetings should be held in the BIS premises.

  • Regarding the process Prof. Dass commented in case if some member of the committee cannot be present in the meeting then he can send his comments through email or send a representative to be present in the meeting.

  • Mr Suman expressed concern over the size of the actual document. Including the referenced standards in the ECMA OOXML the total documentation will be over 10, 000 pages he commented.

  • Prof. Gupta therefore suggested to include other people like Suchitra Pyarelal from NIC, Prof. Pankaj Jalote from IIT Delhi. He also emphasized that NASSCOM should also be included.

  • Mr Bhattacharya suggested to include representatives from CSI and Prof. Shiva Kumar from IIT Bombay.

  • Prof. Gupta also added that the document should be divided and given to different members for analyzing it. A techno-legal expert like Pavan Duggal should be a part of the committee.

  • Prof. Gupta highlighted line 8 of page 12 in Part 1 of the ECMA OOXML document and questioned about the meaning of "at large" and why a product name is mentioned as the purpose of creating a standard.

  • Prof. Dass supported his comments and emphasized that no product name should be mentioned while describing a standard.

  • Mr Bhattacharya expressed his concern over a OOXML compliant application's ability to open the existing large set of documents that has been referred to in the Introduction part of proposed standards

  • The committee went through the document and had the following concerns which is listed in the table below.

PART 1

 

Page No.

Line No.

Comments

12

5-8

Will a OOXML compliant application be able to open the existing implementations of MS office documents and all previous versions as indicated in this line

14

26

Complex Language, statement is not clear

11

It is not clear as to what files & applications are referred to in this line

31

This is an interoperability issue as a standard converter cannot be built due to the presence of some non-standard embedded objects.

16

6

This action can include addition of non-standard elements in the documentation.

428

17

A standard should not use proprietary naming convention.

 

PART 3

 

Page No.

Line No.

Comments

13

6-8

Contradicts Line 3, page 12/178, Part 1.

14

15-20

Need to confirm if the terminologies used are ISO compliant

15

25

Additional resource required to study Section 2.3.1 preferably from NIC

16

8

Check if the RGB color code proposed is a ISO standard.

34-37

The treatment of trailing white spaces is not clear and it does not specify what it is designed for.

17

4

pPr & rPr elements not defined

5

What is glyph and if glyph is part of WordProcessing ML.

33

It is not clear as to what is the source of “preserve” since its not a tag.

21

14-35

It is not clear as to where the value of the parameters used are coming from and if they are ISO standards, including “single”, “auto”, “4”, “0”, “10296”.

22

1-4

How is 100% of page width getting defined.

2

Is “pct” an ISO standard metric ?

12-13

Explain

23

8-11

Is “twips” an ISO standard metric ?

 

 

  • The group unanimously endorsed the above points and agreed that it is not possible to analyze the document without more experts.

  • Ms Sharma's remarked that is necessary to further divide the work of analyzing the document and it should be done in the next meeting.

  • The group also decided that the above results to be presented to BIS for their help in resolving the concerns.

  • Ms Sharma also remarked that as per the previous BIS meeting, she was not supposed to be the convener of the WordProcessing ML group and that it was decided in the previous meeting that Sun Microsystems will be the convener of the WordProcessing ML group.

  • The next meeting was tentatively scheduled on 10th April at 2 pm at BIS premises.

     

     

     

     

    WordProcessing ML Sub-Committee

     

     

     

    Center for Excellence in E-governance

    Department of Management Studies and Department of Computer Science and Engineering

    Indian Institute of Technology, Delhi

     

     

    Minutes of Meeting held on May 07, 2007

     

     

    Topic Second meeting of the WordProcessing ML Sub-Committee

     

      Venue Conference Room,Manak Bhawan, BIS, New Delhi

     

    Date and Time 3.00 PM – 5:00 PM, May 07, 2007

     

    Attendees Dr. Jaijit Bhattacharya, Sun Microsystems (Convener)

    Dr. Geetika Sharma, TCS

    Dr. Nagarjuna, FSF

    Mr. Amit Chadda, MAIT

    Ms. Reena Garg, BIS

    ● Mr. Surendra Gupta, TCS

    Mr. Vijay Kapur, Microsoft

     

    Regrets Prof. Rajanish Dass, IIMA

     

     

    Agenda

     

    3.00 – 3.05 PM Recap of the Action Points of First Meeting

    3.05 – 3.15 PM Suggestions from attendees on inclusion of new members

    3.15 – 3.30 PM Finalizing Division of work

    3.30 – 4.30 PM Detailing analysis of Part 1, 2, 3

    4.30 – 5.00 PM Resolution of Issues and Division of work

     

    Minutes

     

    • Mr. Kapur wanted to clarify who were the members of the sub-committee. He suggested to include representatives from Sonata & Novel. The convener requested for the name of individuals from Sonata and Novel to be sent to the convener.

    • Dr. Bhattacharya suggested to co opt Prof. S.K. Gupta from IIT Delhi, Dr. Nagarjuna from Free Software Foundation, Prof. Pankaj Jalote from IIT Delhi, Dr. Surendra Gupta from TCS. They were co opted unanimously.

    • Mr Kapur raised the query if any other formality needs to be done to include the propsed new members or will they be considered accepted as part of this sub-committee. Ms Reena Garg confirmed that no other action needs to be taken and the proposed new members are now accepted as part of the committee, as long as the members themselves agree. She also mentioned that BIS will send formal letters to the proposed new members.

    • Dr. Bhattacharya also requested to include representatives from NIC and Delhi Science Forum.

    • Ms. Garg commented that BIS would be sending request letters to the Institutional Head of NIC and DSF to nominate their representative.

    • The committee also decided that the quorum of the sub-committee would be deemed to be met if atleast 5 members are present or have sent in their inputs.

    • The committee also decided that all future meetings need to be notified atleast 7 days in advance.

    • Dr. Nagarjuna commented that only the first 120 pages of the ECMA document has to be initially studied and if the doubts are clarified then only the rest of the document should be looked into.

    • Dr. Bhattacharya expressed concern about the feasibility to analyze the whole ECMA OOXML document due to its extremely large size.

    • Mr. Kapur argued that the size is due to the large number of features present.

    • Dr. Nagarjuna stated that a standard is not supposed to cover the maximum number of features possible but should cover the minimum number of points of standardization in order to make it implementable. Features are introduced by applications and not standards.

    • Dr. Nagarjuna pointed out a deficiency in the ECMA OOXML standard that it had no reference to the prior standards of its kind in the same area, i.e. ODF.

    • Mr. Kapur defended the objection by stating that ISO does not mandate the mention of a past standard while describing a new standard. He also commented that ODF has a lot of shortcomings.

    • Dr Nagarjuna pointed out that all standards, including OOXML will have shortcomings and hence the shortcomings needs to be ironed out in the existing standards rather than creating new standards.

    • Dr. Nagarjuna pointed out that ODF does not allow arbitrarily inclusion of new semantics, tags but ECMA OOXML standard does. This hampers the purpose of open standards as it restricts interoperability.

    • Mr. Vijay Kapur clarified that one of the important design goals of the ECMA OOXML was to bring the existing corpus of documents into an open standards without the loss of its fidelity.

    • Dr Bhattacharya requested Mr Kapur to clarify if by existing corpus of documents he meant the Microsoft format based documents to which Mr Kapur responded in the affirmative and stated that the design goal of OOXML, as mentioned in the OOXML documentation, is to support existing Microsoft format based documents.

    • Dr. Nagarjuna reiterated that existence of such a provision which may be non-intentional, defeats the purpose of interoperability and can be done away with.

    • Mr. Kapur pointed that ODF cannot include formulas.

    • Dr Bhattacharya queried if OOXML explicitly defines formula schema or are they referenced to non-open segments. Mr Kapur said that he will get back with the clarification in a detailed note to the convener, within 7 days from the date of the said meeting.

    • Dr Nagarjuna also volunteered to put in his issues in detail and send it to the convener within 11 days of the said meeting.

    • Dr. Geetika Sharma questioned the necessity to have a separate standard when ODF already exists.

    • Dr.Nagarjun questioned Mr. Kapur to state the functionalities which OOXML has in excess to that of ODF.

    • Mr. Kapur responded that it is mentioned in the Document having the ECMA response of the contradictions.

    • Mr. Gupta stated that two standards can co exist and it is up to an organization to adopt the standard of its choice.

    • Dr. Nagarjuna pointed that as OOXML and ODF function in the same way there is no need for two standards. As ODF is already a standard therefore any other standard should not be considered.

    • Mr. Surendra Gupta suggested that BIS has to identify what it is looking for and identify some parameters for it.

    • Ms. Garg agreed that BIS has to decide the parameters in which a standard has to be chosen in the Indian context.

    • Dr. Nagarjuna's remarked that it is very clear that the compatibility of the existing Microsoft documents was the goal to have ECMA OOXML and not an open standard. He pointed out Page 2379, Section 2.16.19, which refers FLD data and it can be embedded in .doc, .xls etc He also questioned the assumption that ODF cannot open the existing corpus.

    • Mr. Vijay Kapur mentioned the existence of test suites by Univ. of California which states that Open Office and KDE Office have no conformance. It states no implementation of ODF is more than 50% compliant. He also commented that the current implementation of ODF cannot bring back the existing corpus with full fidelity.

    • Dr Bhattacharya asked Mr Kapur if he would like to stand by the figures on compliance or would he like to recheck the facts. Mr Kapur replied that he will get back to the convener with the actual figures after double checking.

    • Dr Bhattacharya also queried if compliance of the single implementation of OOXML (i.e. Microsoft Office) has been checked by any independent 3rd party in the manner the multiple ODF implementations have been checked.

    • Dr. Nagarjuna added that ODF cannot support formats that are proprietary and not disclosed to the public.

    • Mr. Gupta added that as ODF is a standard and has some limitations in supporting some formats presently and can always improve and support them in future.

    • Mr. Kapur said that ODF and ECMA OOXML are both standards and the the lossy process of translations is a problem of the editors. He also said that the argument that ODF and OOXML cannot co exist is hollow. Again he stressed that ISO does not mandate the reference of an existing standard.

    • Mr. Gupta inquired if OOXML have more features.

    • Mr. Kapur emphasized that we could do much more with ECMA OOXML than with ODF. He stated that it has custom XML schema's that connect documents to back end databases.

    • Dr Nagarjuna pointed out that these custom XML schema's are precisely the reason why OOXML cannot be considered to be an Open Standard as Open Standards cannot have custom XML schema. All macros, scripts, OLE objects and formula are in the closed domain and any reference to these by a standard automatically renders the standard as also closed.

    • Dr. Bhattacharya commented that in India we do not know any implementation of OOXML. He also questioned if anyone has tested Microsoft Office to be OOXML compliant. Mr Kapur did not respond to this query and said that as part of the documents with his detailed views that he will send to the convener within 7 days of the said meeting, he will will cover this query.

    • Mr. Gupta again stressed the need to decide upon the features to consider and then to decide which standard should India adopt.

    • Dr. Bhattacharya commented that the Chinese sponsored WAPI has not accepted as a standard as a wireless LAN standard as a prior standard (IEEE 802.11) is already existing. Hence there is a precedence of standards bodies avoiding duplication of standards.

    • Finally Dr. Geetika Sharma commented that reference to prior art has significant implications in the scientific world. She expressed concern that we should not be wasting time to judge what is not good enough.

     

     

     

     

    Action Points:

     

    • Dr Nagarjuna to give his detailed opinion on the issue by May 21st, based on his experience in document format standardization.

    • Mt Vijay Kapur to send his detailed inputs on OOXML by May 15th, based on Microsoft's background in creating the standard. The document will clarify the issues raised in the discussion from Mr Kapur's perspective and should include the design principals of ECMA OOXML.

    • BIS should come out with the parameters on which the other members of the committee base their decisions by 15th May.

    • The minutes of the meetings to be sent out after receipt of the above documents.

    • It was also decided that BIS and Microsoft will set up a 3 day workshop in Delhi wherein Microsoft will explain the ECMA OOXML standards to the sub-committee and any other interested groups. At the end of the 3 day workshop, on the 4th day, eminent academicians familiar with the standards issue can deliberate and come out with their opinion on India's response to the voting on OOXML as an ISO standard. The dates of this session was fixed for May 29 to June 01.

    • The convener pointed out that he will be on vacation which he will cut short and come back by May 27th and therefore he will not be able to followup with the members to do their action points. He therefore requested the members to strictly follow the deadlines and not expect followups.

    • BIS to send invitation letters to Delhi Science Foundation and NIC as well as to the newly nominated members.

       

     

     

     

    COMMENST FROM DR NAGARJUNA, Chairman FSF India

     

     

    Why Ecma OOXML cannot be regarded as a free/open document standard?

    Note submitted to the Working Committee, Board of Indian Standards on WordprocessingXML, a component of OOXML

    Nagarjuna G.

    16/05/07

    nagarjun@gnu.org.in

    Table of Contents

     

    1Introduction.......................1

    2What is a Free/Open Document Standard?.................................................................................................1Decodability..................2

    2

    2Use of preexisting standards.....................................................................................................................3

    3No Private Language Components...........................................................................................................3

    4Implementation Independance................................................................................................................4

    5Transparent Collaborative Production Process..........................................................................................................................................................4

    6Rationale for deviating from the preexisting standards incommensurability...........................................................................................................4

    7Demonstrate 4

    8Freedom to implement the standard.......................................................................................................5

    3Does OOXML meet the above conditions?....................................................................................................5

    1OOXML is not decodable in all cases........................................................................................................5

    2OOXML does not use many of the existing standards..........................................................................6

    3OOXML allows insertion of private language components.................................................................7

    4OOXML is clearly implementation dependent.......................................................................................8

    5OOXML is not produced collaboratively among the stake holders....................................................8

    6OOXML specification does not provide rationale for not using or deviating the existing standards.........................8

    7OOXML specification does not demonstrate that it is incomensurable with the existing standards.........................free......................................................................8

    8It is not clear if OOXML implementation is royalty Response.......................................................................................................................8

    4Response to EMCA's 9

    5Concluding Remarks......11

    1Introduction

     

    Let me briefly state the objectives of this note. During the working committee meeting on 7

    th May 2007 at BIS it was decided that I draft the main arguments for, why free software community thinks that OOXML cannot be regarded as a free / open standard. This draft though is based on a number of resources already published, it is an attempt to bring together a few of the most important objections raised against OOXML along with a few Why OOXML is not a FDS? Nagarjuna G.

    other additions that came to my notice while studying the 6000+ page proposal submitted by Ecma

    All the comments expressed below are based on what the author's understanding and assumptions of what is a free document standard. I will therefore provide such a statement in section 3 before listing the arguments. The rationale for using the term 'free' in place of 'open' is also to bring to the notice of the standards body the opinion of Free Software Foundation. Briefly stated, the rationale is:

    1. Some of the details already published (for example in footnote 5.) are not included in this document. opening is not enough, without the freedom to implement.

    Let me begin by appreciating Microsoft for taking the responsibility by providing a data exchange format in which several million documents could now be converted into an encoding that is not apparently private. This action can save a lot of public as well as private documents from vanishing. But, before a nonprivate encoding becomes open and finally free, it is a long way. Though Ecma and MS pushed this nonprivate format to ISO to get a standard status, the existing situation does not seem to be in their favor.

    It is however very important to keep in mind that ISO is a standards body and not an exclusive 'open standards' body. A standard granted by ISO does not imply that it is free or open. Therefore satisfying the ISO conditions is not enough for free software community.

    2What is a Free/Open Document Standard?

     

    I present here the following requirements, which according to me make a proposal a free document standard (hence forth FDS). In order to qualify as an FDS a proposal must meet all the following applicable conditions. In the following section I will try to demonstrate that OOXML does not pass through this filter, while ODF does

    2. The criteria, obviously, are not invented for excluding OOXML but based on sound reasoning keeping in mind the requirements of a good digital society.

    1Decodability

     

    One of the

    first condition to be met by a FDS is that the data is encoded by any application according to an explicitly stated and published convention such that any document saved in such a format can be decoded (interpreted) as per the published specification. This

    1 Ecma 376, 1

    2I am deliberately using the term 'Free' in place of 'Open', since a few of the open standards pass all but not the condition stated in 3.7. 2Why OOXML is not a FDS? Nagarjuna G.

    st Edition December 2006

    condition ensures that the data created by the users in an electronic format is accessible (interpretable) not only in the application in which it is created, but in all future versions of the same application, as well as in all other applications that implement the standard. This ensures not only the endurance of the data, but also its use (interpretation) in future. Since an official document created by any agency for egovernance (Government or any public/private body) is legally required to have such endurance, any proposal must meet this condition.

    The most widely used term for this condition is 'interoperability'. The term 'decodability' is used in place of 'interoperability' since the latter is not possible without the former. It is scientific to use an operational term, for operational term does not have multipleinterpretations.

    2Use of preexisting standards

     

    The proposed standard must use existing standards of the similar kind. In the context of a document standard it is important to remind ourselves that a document is a collection and composition of

    SGML (ISO 8879) is a good example of a standard markup, since such a technical requirement was felt by the industry at that time. When XML was proposed as an open standard, it did not throw away the existing standard, though it was possible to create an arbitrary new language, instead they made use of SGML, included an additional contraint to it, and also adopted another open standard Unicode (ISO 10646). This is how they went ahead by creating an exemplary standard to serve the purpose of multilingual document exchange. HTML was rerepresented in XML and became XHTML. This example may be considered as a paradigm case of how standards are created and extended to meet technical and practical demands. Let us keep this example while handling the case at hand.

    codes. All code is arbitrary to begin with, till it becomes a socially accepted convention. Standardization is the process by which we arrive at such social acceptance. It is possible to invent several distinct arbitrary codes to represent the same thing. This latter possibility is good for expressing creativity, but is not desirable to indulge in such an activity if reusability and endurance of electronic documents is our primary practical objective. If reencoding is required for technical or practical reasons, it is important to invent one.

    3No Private Language Components

     

    The proposed standard must not have private language components embedded within it. A private language is an encoding, decoding of which is not declared publicly. Though this sounds like a very important requirement, it is redundant, because by adhering strintly to the first condition we already precluded this possibility. I am committing redundancy in order

    3Why OOXML is not a FDS? Nagarjuna G.

    to make the argument complete and explicit, particularly to handle the case at hand.

    4Implementation Independance

    One of the most important objectives of agreeing to some standard is to ensure its implementation by multiple vendors. The objective of the standard should not be defined to conform to a single implementation.

    5Transparent Collaborative Production Process

     

    The standard must be developed by involving and inviting all the stake holders in a transparent collaborative consensus driven process. This is to ensure that the standard is not dominated by any one interest group. There must be a room for inviting contributions, suggestions, criticisms and improvements from any interested agency. Since a standard is required to be upheld by several agencies, it is necessay to take everyone into confidence.

    3

    6Rationale for deviating from the preexisting standards

     

    If there are technical or practical reasons for not using an existing standard, it is necessary to specify the relation or lack of such relation with the existing standards, particularly when the existing standards are already adopted by several agencies. This possibility ensures two things. One it allows inventing new ways of expressing, and two it helps in establishing relations to known canons of understanding. This is the way modern science and technology manages to innovate as well as relate to history.

    7Demonstrate incommensurability

     

    If the innovation is so novel that it is incommensurable with the existing set of conventions, then it is indeed a celebration time, for scientists and engineers indeed look forward to listen to such radical innovations. However, the inventors do have the burden to demonstrate that the new innovation is incommensurable with the existing canons, and why it should be considered for a standard, specifically if their innovation is to be regarded as a standard.

    In any case, incommensurability is a very rare case, and such events are exceptional. Exceptional cases are better not handled by standardization process. Because, standardization is about recognizing a procedure, to make a known art into a social convention.

    3Definition of Open Standards, June 2004, National IT and Telecom Agency, Denmark. 4Why OOXML is not a FDS? Nagarjuna G.

    8Freedom to implement the standard

     

    And last, as usual not the least, is the condition that every agency must be given freedom to implement the specification without royalty. If there exist any restrictions in use due to say some patents or other such rights, they must be explicitly exempted before agreeing to confer the FDS status to a proposal.

    4

    There may be many other parameters on the basis of which one may consider if the proposal is a free/open standard. Though I have not made references in formulating them, these are stated keeping the already published ideas and following the several debates on the issue in mind. Given enough time, one can make a comparision to the existing understanding on the subject.

    3Does OOXML meet the above conditions?

     

    Having made the criteria of analysis explicit, let me now see if OOXML meets the above conditions.

    1OOXML is not decodable in all cases

     

    The first condition of decodability is

    After the emergence of semantic web standards, like OWL (Web Ontology Language), which is encoded in another open standard called RDF (Resource Description Framework), and the most recent ISO standard CL (common logic) the need for inserting any arbitrary XML in any document is almost nil. These standards are meant to insert any arbitrary annotations to any component of a document. These are based on the wisdom that any description logic (DL) can be captured by first order logic, and therefore, if one vendor encodes in any standard language that can express DL, any other application vendor can make sense of them, even if the assertions made are arbitrary. The wisdom in this is visionary, and is driving the current revolution called semantic web which is transforming the current web publication into Web 2.0.

    OOXML does allow the inclusion of any of these standard markups inside the OOXML

    apparently met by OOXML. The fact that it is encoded in XML apparently gives one this impression. However, OOXML specification contains certain special provisions to embedd 1. arbitrary custom XML at places, 2. possibility to embedd e.g. OLE objects and 3. possibility to reference to any document encoded in a proprietary format. This enables an an applicatiion vendor to keep proprietary elements within the so called standard.

    4This is also an important condition to call a standard 'open' according to Danish Government definition, European union definition and Bruce Perence definition. 5Why OOXML is not a FDS? Nagarjuna G.

    document. This is because they gave the freedom to insert any arbitrary XML. This is not wise, and can promote proprietary interests. How this can promote proprietary interests is explained in the following scenario.

    Take a vendor who adds a feature which allows the individual users to add annotations in their office documents. The vendor advertises this feature as additional value their customers can get if they purchase their office implementation. Attracted by this nifty feature, users begin to use this feature and add lots of annotations. These annotations are encoded in some arbitrary XML tags in the document, which OOXML allows. The user exchanges this document to another colleague, and to make the matter a little more complicated, to a colleague in another country using another office implementation. The second user cannot make use of the annotations unless the latter also uses the same office implementation. This to me appears a way of inserting the proprietary interests in side an otherwise completely decodable XML.

    This is a good example to show how a vendor can

    As explained earlier, computer science as well as standard making bodies already siezed of this requirement, and created the standard ways of semantic markup that can be used about a document (as in the case of XTM, XML Topic Maps) or within a document by inserting arbitrary assertions by using RDF or OWL or even more expressive CL.

    I therefore argue that no

    abuse XML, an open standard, to generate a proprietary encoding. As experts we must understand that this kind of exploitation is possible within XML. Simply declaring that the encoding is XML is not therefore sufficient to call a standard free or open. This is a common misconception that any encoding in XML becomes a standard. arbitrary insertions neither be allowed nor requuired in a FDS. Thus, though on the face of it, OOXML is apparently open, it is cunningly made not to be a FDS. Therefore the first condition is not met.

    2OOXML does not use many of the existing standards

     

    This is the main reason why the OOXML specification runs to over 6000 pages, and ODF which meets exactly the same goal in about 700 pages. The only open standard they have used (rather abused) is XML.

    Several published comments on OOXML already make this point sufficiently well, therefore I do not want to repeat them here. Not only that OOXML does not make use of ODF, it also does not make use of MathML, SVG, Xlink, RDF etc.

    5

    5 See also ODF Alliance UK Action Group Technical Distinctions of ODF and OOXML: A Consultation Document by Edward Macnaghten, 6Why OOXML is not a FDS? Nagarjuna G.

    All that OOXML demonstrated in this massive effort is the possibility of creating another

    If terseness is the innovation of OOXML, it is not necessary to create another massive standard. Give me some time, I will create another terse ODF format, and publish a table specifying the one to one correspondence with the existing nonterse ODF. Whether this exercise has any technical or practical value, I doubt. This is not what they did anyway, they unnecessarily created a massive arbitrary code, with the single minded objective to map their own proprietary documents.

    As already stated earlier, what they did is good for making their proprietary formats open. It does however serve this purpose, and let me repeat:

    arbitrary markup from scratch. Some claims that they made, such as OOXML is more terse and produces more compact documents, is untenable because XML is meant to be a human readable markup so that developers can implement and reimplement support for a FDS easily. XML exists as an intermediary data exchange format, but at the same time it is also a very easy way of communicating in a desciplined natural language. That is the reason why XML is the industry choice for communicating not only with human beings, but also with machines. This is the second case where OOXML abuses XML. this is the only purpose OOXML serves. Let us congratulate them for this feat. But, this was their responsibility, long due. Having completed their task, they are interesting in pushing this as an open standard. I think civil society should not let this model to become a shortcut process to create a FDS. Please see the latter condition, that the process of creating a FDS must also be a collaborative exercise, as is the case in ODF. It is known that the current vendor boycotted that process, though it was very clear that the vendor was one of the major stake holders of ODF. This instance therefore is a clear case of expression of a big brother attitude, and not a socially admirable act.

    3OOXML allows insertion of private language components

     

    The reason why OOXML is not decodable in every case, is already illustrated above. Here I wish to bring to the notice of the possibility of inserting OLE objects, even embedding and referencing proprietary documents within. As stated already, this is a redundant statement. My only justification is that this point should not escape any one's attention. This defeats the whole purpose of creating a FDS.

    They might respond by saying a vendor can add lot of value by embedding objects within a document. Yes, they can. But not without voilating the norms of FDS. This can be achieved by other means,such as in UML. This is technically possible, since UML is based on a standard specification of object orientation OMG.

    7Why OOXML is not a FDS? Nagarjuna G.

    Embedding and referencing proprietary documents within must be barred from the process, for it defeats the purpose.

    4OOXML is clearly implementation dependent

     

    Ecma's statement, fully quoted below in Section 5, makes this very clear. Their objective is clearly not to arrive at a generic document standard. By providing the need to refer to and embedd nonstandard vendor specific components (see the above section 3.3), OOXML makes

    full implemenation by other vendors impossible. If such a specification becomes a standard, it will lead to furthering a monopoly. Since the very objective of a standard specification is to eliminiate this possibility, OOXML can not be regarded a standard.

    5OOXML is not produced collaboratively among the stake holders

     

    The ITUT definition of open standard stipulates this important condition. This ensures that standardization process is not dominated by one single interest group.

    6OOXML specification does not provide rationale for not using or deviating the existing standards

     

    There is no reference to the existing standards in the specification, and also no comparision. This is the tradition of science and technology, and in a sense the ethos of the modern democratic society. Deviation or extension of the existing standards is possible. For example, CL (common logic) as recent ISO standard, is an extension of KIF (Knowledge Interchange Format) and also borrows from the wisdom of a published tradition of Existential Graphs in the form of CG (Conceptual Graphs). There may be several other examples. This to my understanding is a desirable way of deviating and extending from the already existing standards.

    7OOXML specification does not demonstrate that it is incomensurable with the existing standards

     

    This is clearly not an applicable condition, since OOXML did not invent anything new. According to Ecma's own assertion, ODF and OOXML are overlapping standards, therefore no incommensurability.

    8It is not clear if OOXML implementation is royalty free

     

    In the reponse given by Ecma to the comments (Ecma/TC45/2007/006) it is clear that they

    8Why OOXML is not a FDS? Nagarjuna G.

    are trying to meet ISO's requirements on this issue. What they are doing is what is possible by the standards organizations. This may make it a standard, but not a free standard. The rational to keep standards free is well published. See for example Eben Moglen's appeal to W3C that RAND is not enough for a FDS.

    6 European Union's, Danish Government's, Bruce Perence's definition also requires this condition to be satisfied for a standard to become open.

    4Response to EMCA's Response

     

    Ecma published responses after the 30 day review of the fasttrack ballot (Ecma/TC45/2007/006). These responses make one point very clear: Ecma is catering to the special requirements of MS, and the objective is not to produce a standard. Having produced an XML version, they want to push this as a standard.

    The existence of alternate standards (ODF, HTML, PDF) for office document, and CGMOpen and SVG for vector graphics, JPG and PNG for rastor graphics, RelaxNG and DTD for schema specification in XML document, and TIFF and PDF for press ready graphics format are quoted as existing precedents that there are multiple standards for the same task. In all these cases, there are some special features one format supports, and the other does not. The users of each of these formats make good decisions as to which format is good for which context/purpose. This itself indicates that if the purposes are different they need to make an appropriate decision.

    For example, if one wants to display an image on the screen, any one of JPG or PNG will do (FSF recommends PNG for it is a free standard). But if the user intends to animate only PNG can be used. Thus, in the former usecase there is overlapping, while in the latter usecase only one of them can do the job. This kind of usecases can be shown in all the examples given above.

    The issue at hand, the comparison of OOXML and ODF, is not of this kind, for both of them overlap mostly, and their functionality is identical. The few elements that can be found in one and not in the other constitutes the real problem. The effort of the industry should be to identify these elements and find a technical solution.

    ODF cannot include arbitrary XML to the best my knowledge. If I am right, then ODF is a purer FDS since by design it does not make undesirable elements to be part of the documents. Without such contraint it is not possible to achieve interoperability. Ecma says ODF cannot take care of these elements. But why take care of them when this is against the goals of a FDS.

    6http://lists.w3.org/Archives/Public/www-patentpolicy-comment/2001Sep/0650.html 9Why OOXML is not a FDS? Nagarjuna G.

    Therefore all we can see in this analysis is that OOXML and ODF though have a lot of overlapping, they do not overlap in the following features: 1. 100% interoperability which ODF has, while OOXML does not have it by design, and 2. OOXML can embedd private language elements while ODF does not.

    Ecma's argument that several of MS's clients would not want to loose important data, while transferring from proprietary format into OOXML, and that is why they want to have other means of embedding private data. This clearly demonstrates that existing proprietary docuements cannot be efficiently and completely converted into interoperable parts of OOXML. This may be either a failure on their part to produce a fully expressable office document standard, or they do not want to free all elements of their proprietary encoding.

    A few things ODF cannot do which OOXML can do. One example is, in the current version of ODF it is not possible to insert a table in a slide. Therefore, if one is making a filter from OOXML to ODF, there will be loss. It is in the interest of the community that we should identify all such problem areas. All such problems can be solved. There is, to the best of my knowledge, nothing inherently wrong in ODF structure that prevents it to solve such problems. The free software community could identify such problems and solve them in the forthcoming releases of the standard. If the features that are missing in ODF are genuine, and are required, Ecma would have proposed their addition in ODF, which I am sure ODF community will welcome.

    In Ecma's response document the truth of the matter comes out very vividly:

    OpenXML is designed to represent the existing corpus of documents faithfully, even if that means preserving idiosyncrasies that one might not choose given the luxury of starting from a clean slate. In the ODF design, compatibility with and preservation of existing Office documents were not goals. Each set of goals is valuable; sacrificing either at the expense of the other may not be in the best interest of users. (p.6 Ecma Response)

    This is the fact of the matter. This clearly shows that one of them is trying to preserve the existing data created by a single vendor, while the other is to provide a generic encoding standard for office documents. It is true therefore that their purposes are different. Since there is a difference in purpose despite the overlapping with ODF, Ecma argues, OOXML can also exist with ODF.

    But the issue is: providing a way of preserving a vendor's old documents is the service that a vendor is expected to do. This must happen. This can happen by converting the documents into ODF. Ecma did not prove that this is impossible.

    We therefore think, that Ecma has the burden to prove that proprietary documents made by them cannot be converted into ODF. It is very likely that there can be a few elements that

    10Why OOXML is not a FDS? Nagarjuna G.

    cannot be translated, since ODF was not made to serve a particular vendor's requirements. Once such elements are identified, Ecma can propose a model of extending ODF so that the possible problems are sorted out. This is the desirbale way, so that every office document is available in a version of FDS and free software community's preference is ODF format.

    5Concluding Remarks

     

    1.

    Based on the analysis, OOXML is not a FDS.

     

    2.

    It is in the interests of all the users that all the proprietary documents be converted into the existing free document standard, ODF.

     

    3.

    Ecma or MS may identify the elements that ODF cannot take care, and propose extensions to ODF, or create a standard if necessary that is an extension to ODF, ensuring compatibility. 11
                      At the time of uploading the above minutes on June 2nd, 2007, the comments from Microsoft that were supposed to be sent on May 15th as per the May 7th meeting, is still awaited.

                    Comments (0)

                    You don't have permission to comment on this page.