Re: OSBA Office Interoperability Workshop last week
On 11/17/13 9:19 PM, Pedro Giffuni wrote: Hi Juergen; I am just catching up with some old mail but thank you for the interesting report. Just as an external observer and reminding that IANAL, the LO/TDF doesn't seem to be complying with the Apache License section 4 subsection (d) and is misinterpreting on purpose subsection (b). The so-called rebasing was obfuscated on purpose, probably with some level of paranoia, to avoid making clear which code is under an Apache License and which is not. It was not surprising to find that at least one CWS, not included in AOO, was relicensed and mixed with other arbitrary license swapping. I personally share your views and it makes me sad to see that so called open source advocates tolerate and support such things. Where as individuals might not have interest in a clear and clean code base with a proper and appropriate licensing. But I would have expected that companies who are involved take it more serious. I am still liking the open source idea and I believe it is good to reuse code that is already implemented. But I also like the idea of upstream changes to the original code base because it is the only way how open source can work. Using public funding for opensource projects is great but it involves all the parties working responsibly. I just think that given the situation of uncertainty of the results and the general lack of collaboration from one group it just doesn't make sense for both projects to work together. the first project was in the special situation that OpenOffice was in it's transition to Apache and that some people tried hard to claim that OpenOffice is dead and LO is the new and shining successor. Means the situation at this time was not clear and the project constraints were probably not defined detailed enough. I think it was not easy and the involved people tried their best. I think future projects can make it better and can ensure that the changes are integrated in the lowest common code base. And I disagree and think that collaboration make a lot of sense because both projects can benefit from even more developers and wasting time with useless merging of code is stupid. It's not all bad though: While LibreOffice seems to be a place to run experiments, AOO is setting the new standards that get widely adopted. well if we take the license issue out and look on the LO project only I like a lot of things that work quite well. And I would love to see many of them in place at Apache and AOO as well. Juergen Just my opinion though ... I am always glad to be proved wrong. Pedro. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: OSBA Office Interoperability Workshop last week
Hi Juergen; I am just catching up with some old mail but thank you for the interesting report. Just as an external observer and reminding that IANAL, the LO/TDF doesn't seem to be complying with the Apache License section 4 subsection (d) and is misinterpreting on purpose subsection (b). The so-called rebasing was obfuscated on purpose, probably with some level of paranoia, to avoid making clear which code is under an Apache License and which is not. It was not surprising to find that at least one CWS, not included in AOO, was relicensed and mixed with other arbitrary license swapping. Using public funding for opensource projects is great but it involves all the parties working responsibly. I just think that given the situation of uncertainty of the results and the general lack of collaboration from one group it just doesn't make sense for both projects to work together. It's not all bad though: While LibreOffice seems to be a place to run experiments, AOO is setting the new standards that get widely adopted. Just my opinion though ... I am always glad to be proved wrong. Pedro. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
OSBA Office Interoperability Workshop last week
Hi, I would like to inform you about my attendance of a workshop from the OSBA [1] to discuss a potential future project to improve the interoperability of OpenOffice/LibreOffice with the also standardized file format OOXML. But also with focus on an improved and a standardized change tracking proposal. I attended as individual AOO member and IBM representative. I made clear that I am no official spoke person for AOO and explained once more that we don't really have a hierarchy or formal leaders. One part of the workshop was to review the first project [2] which was a success for open source but not directly for OpenOffice. We know all that the patches are not yet integrated in AOO. I reported that I have informed the AOO project/community about the availability [3] of the related patches but that nobody has worked on it so far. And that it is not easy without having access to the test documents. Svante Schubert gave a good overview presentation about the change tracking proposal that is currently discussed and proposed in the related OASIS sub committee. All attendees agreed more or less that it is important to have it more formalized and be part of the ODF standard. Funding to work on the ODF specification is one aspect ... During the workshop new problems were reported and feature requests communicated. This will be me worked out in detail and new use-case specifications will be prepared similar to the first project. When they are available I will share them with the community. Interested developers and companies can give an offer to work on the implementation later, similar to the first project. A further important point was the potential collaboration between OpenOffice and LibreOffice at least on source code level. Some of the sponsors of the first project were not 100% satisfied because they can't benefit from the work they have paid for which I can understand. The availability of patches under ALv2 is not enough to have them integrated. The integration work have to be done and ideally from the people who were paid for. Or at least in time and in collaboration with other volunteers. Anyway something that will be probably improved in the future. Jan Holosevsky from Collabora and a developer on LO and me as a developer from AOO were asked about a proposal/idea how such a collaboration can happen. We all know that it is not so easy to answer and that it comes quite fast to an ideological and political discussion. I simply tried to explain the situation we already have today. In detail I showed the code flow from AOO to LO and the dependency of LO to AOO since their rebase. They mirror our svn repos and merge fixes and features on a regular basis into their code. And most of their source code is under the ALV2 because you can't remove the license. You can only add additional licenses for significant changes you made in a source file. As one possible way for collaboration I proposed to work more directly on the same code base. And that the TDF could continue to provide binaries and could continue with their community work they are doing today (I like of course many things they doing). The only requirement would be to work together on the same code and contribute the changes upstream. I believe this would make most sense and the resources in both project would be used more efficient. And the most important point from my point of view it would reflect the main idea of open source and would benefit the open source spirit. Jan Holesovsky with backing from Simon Phipps proposed that we could use LO code which is under MPLv2. As a reminder the additional MPLv2 is the result of their rebasing work against the AOO code base after our first official release AOO 3.4.1. Well I found not very much information about the exact licensing on their webpage, mainly LGPLv3. And no reference that at least major parts of their code is under ALv2 today. At least to me it looks quite confusing and I am happy that we have it much clearer today. But back to the proposal I have to confess that I don't really understand how this should work in detail. MPL is category-b and we can link against it but we can't host any MPL code in our repo. And it would work on completely new code only that is quite well encapsulated and modularized. It can be potentially an option for some of their new filters but that have to be checked in detail and is only one aspect. We talk about million lines of code mainly. It was also mentioned that mixing of ALv2 and MPL code is possible in general and that it is more a problem of the ASF and the processes applied to projects here at Aapche. I was thinking what it should mean, confuse people even more or an indirect recommendation that OpenOffice should be hosted somewhere else? I stopped thinking about it because it's out of scope I think. If people think I misunderstood things or summarized it incorrect, please feel free to correct me or add missing information. I shared this with you
Re: OSBA Office Interoperability Workshop last week
On Mon, Nov 4, 2013 at 7:53 AM, Jürgen Schmidt jogischm...@gmail.comwrote: Hi, I would like to inform you about my attendance of a workshop from the OSBA [1] to discuss a potential future project to improve the interoperability of OpenOffice/LibreOffice with the also standardized file format OOXML. But also with focus on an improved and a standardized change tracking proposal. I attended as individual AOO member and IBM representative. I made clear that I am no official spoke person for AOO and explained once more that we don't really have a hierarchy or formal leaders. One part of the workshop was to review the first project [2] which was a success for open source but not directly for OpenOffice. We know all that the patches are not yet integrated in AOO. I reported that I have informed the AOO project/community about the availability [3] of the related patches but that nobody has worked on it so far. And that it is not easy without having access to the test documents. Svante Schubert gave a good overview presentation about the change tracking proposal that is currently discussed and proposed in the related OASIS sub committee. All attendees agreed more or less that it is important to have it more formalized and be part of the ODF standard. Funding to work on the ODF specification is one aspect ... During the workshop new problems were reported and feature requests communicated. This will be me worked out in detail and new use-case specifications will be prepared similar to the first project. When they are available I will share them with the community. Interested developers and companies can give an offer to work on the implementation later, similar to the first project. A further important point was the potential collaboration between OpenOffice and LibreOffice at least on source code level. Some of the sponsors of the first project were not 100% satisfied because they can't benefit from the work they have paid for which I can understand. The availability of patches under ALv2 is not enough to have them integrated. The integration work have to be done and ideally from the people who were paid for. Or at least in time and in collaboration with other volunteers. Anyway something that will be probably improved in the future. Jan Holosevsky from Collabora and a developer on LO and me as a developer from AOO were asked about a proposal/idea how such a collaboration can happen. We all know that it is not so easy to answer and that it comes quite fast to an ideological and political discussion. I simply tried to explain the situation we already have today. In detail I showed the code flow from AOO to LO and the dependency of LO to AOO since their rebase. They mirror our svn repos and merge fixes and features on a regular basis into their code. And most of their source code is under the ALV2 because you can't remove the license. You can only add additional licenses for significant changes you made in a source file. As one possible way for collaboration I proposed to work more directly on the same code base. And that the TDF could continue to provide binaries and could continue with their community work they are doing today (I like of course many things they doing). The only requirement would be to work together on the same code and contribute the changes upstream. I believe this would make most sense and the resources in both project would be used more efficient. And the most important point from my point of view it would reflect the main idea of open source and would benefit the open source spirit. Jan Holesovsky with backing from Simon Phipps proposed that we could use LO code which is under MPLv2. As a reminder the additional MPLv2 is the result of their rebasing work against the AOO code base after our first official release AOO 3.4.1. Well I found not very much information about the exact licensing on their webpage, mainly LGPLv3. And no reference that at least major parts of their code is under ALv2 today. At least to me it looks quite confusing and I am happy that we have it much clearer today. But back to the proposal I have to confess that I don't really understand how this should work in detail. MPL is category-b and we can link against it but we can't host any MPL code in our repo. And it would work on completely new code only that is quite well encapsulated and modularized. It can be potentially an option for some of their new filters but that have to be checked in detail and is only one aspect. We talk about million lines of code mainly. It was also mentioned that mixing of ALv2 and MPL code is possible in general and that it is more a problem of the ASF and the processes applied to projects here at Aapche. I was thinking what it should mean, confuse people even more or an indirect recommendation that OpenOffice should be hosted somewhere else? I stopped thinking about it because it's out of scope I think. If
Re: OSBA Office Interoperability Workshop last week
I agree, I think this should be commented upon a new post at the AOO blog. Specially because nevertheless you dont spoke on AOO behalf, is also important efforts are being put into interoperability and others. On 11/4/13, Kay Schenk kay.sch...@gmail.com wrote: On Mon, Nov 4, 2013 at 7:53 AM, Jürgen Schmidt jogischm...@gmail.comwrote: Hi, I would like to inform you about my attendance of a workshop from the OSBA [1] to discuss a potential future project to improve the interoperability of OpenOffice/LibreOffice with the also standardized file format OOXML. But also with focus on an improved and a standardized change tracking proposal. I attended as individual AOO member and IBM representative. I made clear that I am no official spoke person for AOO and explained once more that we don't really have a hierarchy or formal leaders. One part of the workshop was to review the first project [2] which was a success for open source but not directly for OpenOffice. We know all that the patches are not yet integrated in AOO. I reported that I have informed the AOO project/community about the availability [3] of the related patches but that nobody has worked on it so far. And that it is not easy without having access to the test documents. Svante Schubert gave a good overview presentation about the change tracking proposal that is currently discussed and proposed in the related OASIS sub committee. All attendees agreed more or less that it is important to have it more formalized and be part of the ODF standard. Funding to work on the ODF specification is one aspect ... During the workshop new problems were reported and feature requests communicated. This will be me worked out in detail and new use-case specifications will be prepared similar to the first project. When they are available I will share them with the community. Interested developers and companies can give an offer to work on the implementation later, similar to the first project. A further important point was the potential collaboration between OpenOffice and LibreOffice at least on source code level. Some of the sponsors of the first project were not 100% satisfied because they can't benefit from the work they have paid for which I can understand. The availability of patches under ALv2 is not enough to have them integrated. The integration work have to be done and ideally from the people who were paid for. Or at least in time and in collaboration with other volunteers. Anyway something that will be probably improved in the future. Jan Holosevsky from Collabora and a developer on LO and me as a developer from AOO were asked about a proposal/idea how such a collaboration can happen. We all know that it is not so easy to answer and that it comes quite fast to an ideological and political discussion. I simply tried to explain the situation we already have today. In detail I showed the code flow from AOO to LO and the dependency of LO to AOO since their rebase. They mirror our svn repos and merge fixes and features on a regular basis into their code. And most of their source code is under the ALV2 because you can't remove the license. You can only add additional licenses for significant changes you made in a source file. As one possible way for collaboration I proposed to work more directly on the same code base. And that the TDF could continue to provide binaries and could continue with their community work they are doing today (I like of course many things they doing). The only requirement would be to work together on the same code and contribute the changes upstream. I believe this would make most sense and the resources in both project would be used more efficient. And the most important point from my point of view it would reflect the main idea of open source and would benefit the open source spirit. Jan Holesovsky with backing from Simon Phipps proposed that we could use LO code which is under MPLv2. As a reminder the additional MPLv2 is the result of their rebasing work against the AOO code base after our first official release AOO 3.4.1. Well I found not very much information about the exact licensing on their webpage, mainly LGPLv3. And no reference that at least major parts of their code is under ALv2 today. At least to me it looks quite confusing and I am happy that we have it much clearer today. But back to the proposal I have to confess that I don't really understand how this should work in detail. MPL is category-b and we can link against it but we can't host any MPL code in our repo. And it would work on completely new code only that is quite well encapsulated and modularized. It can be potentially an option for some of their new filters but that have to be checked in detail and is only one aspect. We talk about million lines of code mainly. It was also mentioned that mixing of ALv2 and MPL code is possible in general and that it is more a problem of the ASF