User initiative for improving OOXML integration in LibreOffice/Apache Open Office
2 likes3,080 views
Dr. Matthias Stürmer at Ernst & Young discusses an initiative to improve OOXML integration in LibreOffice and Apache OpenOffice, addressing the current shortcomings in handling .docx files to facilitate migration to open-source office suites. The collaborative project involves public institutions and community organizations, aiming to mobilize resources and develop solutions to collective action problems faced by stakeholders. The project's phases include initialization, funding, and implementation, with the goal of enhancing document interoperability and functionality within open-source office applications.
User initiative for improving OOXML integration in LibreOffice/Apache Open Office
- 1. User initiative for improving OOXML integration in LibreOffice/Apache Open Office Dr. Matthias Stürmer, Ernst & Young ApacheCon Europe, 8 November 2012
- 2. Short bio of Matthias Stürmer ● Matthias Stürmer is Manager at Ernst & Young based in Bern, Switzerland. He focuses on open source advisory, open data management, and social media governance. ● He studied business administration and computer science at University of Bern until 2005 and finished his doctoral dissertation at the Chair of Strategic Management and Innovation at ETH Zürich in 2009. ● His research focused on open source communities and firm involvement. The title of his doctoral thesis was "How Firms Dr. Matthias Stürmer Make Friends: Communities in Private-Collective Innovation". Manager ● He then worked for one year at Liip AG, a Swiss software Ernst & Young AG company creating agile Internet solutions based on open Belpstrasse 23 source technologies. 3001 Bern Switzerland ● Matthias Stürmer is working group leader “Office Interoperability” of the Open Source Business Alliance, matthias.stuermer@ch.ey.com member of the board of Swiss Open System User Group Work: +41 58 286 61 97 /ch/open, member of the Swiss association Opendata.ch, Mobile: +41 58 289 61 97 secretary of the Swiss Parliamentarian Group for Digital Sustainability, and leader of www.opensource.ch and other open source initiatives. ● Since 2011 he is member of the parliament of the city of Bern.
- 3. Ernst & Young and Open Source ● Open source has one major weakness: marketing and PR ● Top management vendor-neutral brochure from Ernst & Young: Why and how professionals use open source software ● In English and German ● Content: ● Benefits, risks and good practices ● Professional application of open source software ● Legal aspects of open source ● Background information on open source software Download as PDF on www.opensource.ch
- 4. Open Source in Switzerland ● New study of October 2012 ● Introduction by CIO of UBS ● Results of survey with 202 firms and public institutions ● Around 20 articles on open source topics and use cases ● Portrait of open source companies Download as PDF on www.opensource.ch
- 5. Agenda ● Motivation for the initiative ● Phase 1: Initialization ● Phase 2: Funding ● Phase 3: Implementation ● Discussion
- 6. Why promoting a proprietary format? To increase adoption of open source office suites!
- 7. Motivation of the initiative ● Currently Microsoft's Office Open XML (OOXML) format is not well integrated in LibreOffice/Apache Open Office ● However, this is necessary to successfully complete and maintain migrations towards open source office suites ● Today various public institutions have switched to LibreOffice/Apache Open Office on the desktop or use head-less open source office suites for business applications. ● Nevertheless citizens, businesses, and other public stakeholders continue to send OOXML documents, especially .docx files. ● These external stakeholders expect that the public institution are able to handle such files.
- 8. Major technical issues ● Opening OOXML formats is possible with current versions of LibreOffice/Apache Open Office ● Unknown if visualization is conform to Microsoft Office 20XX ● Most relevant OOXML support issues: – Layout of .docx documents regarding frames, images, tables, enumerations, header/footers etc. – Fonts embedding (resulting in wrong or different layout of documents) – Comments layout and editing in .docx and .xslx – Change tracking within .docx files
- 9. Collective action problem ● “Scratching a developer's itch” does not always work for public administrations because they have different itches. ● While open source projects bring great advantages for their users, they inherently bear the problem of „collective action“. ● Although it would be beneficial for all users to implement enhancements in open source projects, no single stakeholder is willing to invest sufficient resources to accomplish this considerable development effort alone. All the other beneficiaries would free-ride on the solution. ● Therefore often status quo does not represent the optimum. ● In order to solve this challenge coordinated action among the stakeholders is necessary.
- 10. Solving the collective action problem Phase 1: Initialization a) Mobilize attention and interest of several open source software users b) Create clear and common understanding of the issues, ask the experts c) Aggregate and specify the requirements, prioritize and cluster them Phase 2: Funding a) Publish Request for Proposal (RfP) b) Evaluate and decide for best proposal(s) c) Do fund-raising for the implementation project Phase 3: Implementation a) Organize project management, sign contracts, start implementing b) Do testing among the open source software users, finalize development c) Publish resulting source code, pass it upstream to the open source project
- 11. Agenda ● Motivation for the initiative ● Phase 1: Initialization ● Phase 2: Funding ● Phase 3: Implementation ● Discussion
- 12. Setup of the initiative ● Workshop in October 2011 in Zürich ● Tasks were: – Identification of common issues (users) – Technical challenges and innovative solutions regarding OOXML (developers) – Specification of use cases regarding requirements and technical aspects (users and developers in groups) – Discussion and prioritization of use cases (all) – Funding models of implementation phase (users)
- 13. Initiating organizations Public Institutions ● City of Freiburg i.B. ● City of München ● City of Jena ● Swiss Federal Court ● Federal Steering Unit for IT (ISB) ● Canton of Vaud ● Another Swiss federal agency Community organizations ● Association Swiss Open Systems User Group /ch/open ● Association Freies Office Deutschland e.V. (former association OpenOffice.org Deutschland e.V.) ● Open Source Business Alliance OSBA
- 14. Workshop participants Public administration: Developers: ● Rüdiger Czieschla, City of Freiburg ● Thorsten Behrens, Novell/SUSE ● Andreas Kawohl, City of Freiburg ● Florian Reuter, self-employed developer ● Andor Ertsey, City of München ● Svante Schubert, self-employed developer ● Dr. Jutta Kreyss, City of München ● Thomas Uhl, Open Source Business Alliance ● Dr. Martin Reinhardt, City of Jena ● Dr. Frank Siebert, Freies Office Deutschland e.V. ● Dr. Harro Rosner, City of Jena ● Lothar K. Becker, Freies Office Deutschland e.V. ● Marc von Weissenfluh, Swiss Federal Court ● Mario Wendt, Microsoft ● Daniel Brunner, Swiss Federal Court ● Stefan Stüring, KO GmbH ● Tobias Hintze, KO GmbH ● Jos van den Oever, KO GmbH Ernst & Young: ● Dr. Matthias Stürmer, workshop leader ● Mario Aeby, assistant
- 15. Challenges ● Stakeholders – Users: public institutions – Developers: LibreOffice and OOXML filter programmers ● Challenges – Huge knowledge gap: terminology, standard specification, structures and processes within public administrations etc. – Different perspectives: input oriented (=developers) vs. output oriented (=users) – Different interests: perfect implementation vs. solving problem ● How to collaborate together?
- 16. Example: filter vs. feature ● Important difference between – error in OOXML filter – missing OOXML feature within LibreOffice/ Apache Open Office ● What does that mean? – Little effort to correct OOXML filter – Large effort to implement new office feature ● Example of not correctly displayed OOXML standard: Alternating row colors of tables within .docx files – Little effort to 'hard format' colors, however loss of information in the .docx file – Large effort to program this feature in office application because of missing feature in ODF standard
- 17. Structure of a use case
- 18. Evaluation of use cases 1000 days possible show strategic 100 days stoppers features Estimated effort → defined by developers 10 days low special hanging wishes fruits 1 day today ½ year 1 year 3 years Priority regarding implementation → defined by users (public institutions)
- 20. Agenda ● Motivation for the initiative ● Phase 1: Initialization ● Phase 2: Funding ● Phase 3: Implementation ● Discussion
- 21. Bundle of use cases ● Use case bundle is called “Layout-preserving view of OOXML documents in open source office applications” ● Targeted use cases: 1. Formatting of frames and images within .docx 2. Formatting of tables within .docx 3. Formatting of lists and numbers within .docx 4. Formatting of comments within .docx and .xlsx 5. Font embedding for both OOXML and ODF ● Every use case is specified by test documents
- 23. Specification and RfP ● Specification and Request for Proposals (RfP) published in December 2011 ● Press releases by the OSBA working group Office Interoperability ● Proposals submitted by Lanedo and SUSE in February/March 2012 ● Contracts split up along the five use cases
- 24. Public procurement challenge Funding model “intermediary”: Problem: Public procurement Funding partner 1 regulations requires clear results for spoken money Funding partner 2 Intermediary: Implementing Funding partner 3 Coordination company and contracts Funding partner 4 Funding partner 5 Funding partner X
- 25. Public procurement solution Funding model “cake slices”: Technically there is one Legally there are several contracts job to do: the cake which each define some specific requirements: the slices Funding partner 1 Funding partner 2 Funding partner 3 Intermediary: Coordination Implementing Funding partner 4 but no contracts company Funding partner 5 Funding partner X The implementing company/companies sign several contracts with different funding partners
- 26. Current OOXML improvements Funding by Total: approx. EUR 140k (excl. VAT) ● City of Freiburg i.B. ● City of München ● City of Jena EUR 8k ● Swiss Federal Court EUR 4k ● Federal Steering SUSE Unit for IT (ISB) EUR 13k EUR 50k ● Canton of Vaud ● Another Swiss EUR 13k federal agency Ernst & Young ● French ministry of culture and EUR 14k Lanedo communication EUR 25k EUR 15k
- 28. Agenda ● Motivation for the initiative ● Phase 1: Initialization ● Phase 2: Funding ● Phase 3: Implementation ● Discussion
- 29. Agenda 1: Motivation for the initiative Phase ● Initialization a) Mobilize attention and interest of several open source software users Phase 1: Initialization b) Create clear and common understanding of the issues, ask the experts ● c) Aggregate and specify the requirements, prioritize and cluster them ● Phase 2: Funding Phase ● Funding 3: Implementation 2: Phase a) Publish Request for Proposal (RfP) ● Discussion b) Evaluate and decide for best proposal(s) c) Do fund-raising for the implementation project Phase 3: Implementation a) Organize project management, sign contracts, start implementing b) Do testing among the open source software users, finalize development c) Publish resulting source code, pass it upstream to the open source project
- 30. Relevant for Apache Open Office The specification explicitly says: “License - The contractor is obliged to publish the new OpenOffice/LibreOffice software developed for this project under Apache License Version 2.0. This license permits the widest possible distribution of the development services.”
- 31. Current state of the work Some background details: Current state: ● Use case 1 and 2 implemented ● Use case 1 fully implemented by Lanedo ● Use case 2 fully implemented ● Use case 3 to 5 implemented by SUSE ● Use case 3 fully implemented ● Functionality should go into ● Use case 4 partially implemented LibreOffice 3.7 ● Use case 5 started ● All changes will go as source code tarballs (Apache License) to the OSBA working group ● Technical contacts: - Thorsten Behrens from SUSE - Lionel Dricot from Lanedo
- 32. Work in progress in daily builds
- 33. Agenda ● Motivation for the initiative ● Phase 1: Initialization ● Phase 2: Funding ● Phase 3: Implementation ● Discussion
- 34. Questions ● How can OOXML filter patches programmed by the SUSE and Lanedo be integrated in Apache Open Office? ● In what way do Apache Open Office and LibreOffice collaborate today? ● What could be done to improve collaboration between Apache Open Office and LibreOffice?
- 35. Thank you for your attention! For any feedback, questions, ideas, suggestions please contact: Dr. Matthias Stürmer Manager at Ernst & Young AG OSBA Working Group Leader Office Interoperability matthias.stuermer@ch.ey.com Work: +41 58 286 61 97 Mobile: +41 58 289 61 97