I think it is foolish to assign a metric to compatibility between OOO and LibreOffice, and particularly between either and ODF. If the 1% matters to me, it can be a show-stopper for my interoperability needs. The larger the take-up of *Office.org, the greater the number of folks impacted by such things, including deviations attributable to platform, configuration, and version of the software as well as the origin of documents that already exist.
The variety of contingent factors is rather extensive. Adding to the degree of contingency is the extent of discretionary provisions in whatever ODF specification a program's support is based upon. There are places where provisions are loose if not altogether underspecified, where provisions are explicitly implementation-dependent in various ways, and where the conformance conditions are extremely flexible. The sharing of a reasonably-common *Office.org code base promises alignment of discretionary elements and even of extensions having nothing to do with requirements for ODF. There is a barrier of entry, so-to-speak, if the ways discretionary matters are handled are not made explicit so other producers of ODF-supporting software can choose to align or at least to deviate knowingly. I am not talking about the code being available for inspection, I am talking about explicit statements that can be understood without having to examine an implementation. That is what provides for independently-derived interoperably-usable implementations based on an open standard. The work invested in arriving at such arrangements can also lead to valuable feedback to the perfection of the evolving specifications for ODF. One way to accomplish this is to provide implementation notes that explicitly account for the conditions of support for ODF provisions and any deviations that also exist. I have only seen this attempted by one producer. Although that particular provision could be done much better, I find it remarkable that this is the only case where it appears to be done at all. The idea is to be accountable for discretionary matters in a way that adopters of software can make informed choices about which products are suitable in a particular interoperability situation, including support for already-existing documents for which continual conversion is not an option. This is also a way for a development team to be mindful of what their discretionary choices are and how evolution of their code base needs to account for what those choices have been. There is an opportunity for Apache OpenOffice.org to raise the bar in this respect. And how can one provide a reference implementation without such an account of all the discretionary and contingent factors? I have heard it said that no desired quality of software is achievable in the absence of a concrete measure for it. Implementation notes are a way of assuring whatever the desired quality of ODF support is to be, along with the discretionary provisions that are part of that achievement. - Dennis -----Original Message----- From: Andrea Pescetti [mailto:[email protected]] Sent: Saturday, July 02, 2011 07:38 To: [email protected] Subject: Re: Differences between OOO and LibreOffice. On 30/06/2011 Ian Lynch wrote: > If I save an odf file from OOo it will open exactly the same in LibO. > If that isn't true than I would be interested to know where things > break. Talking about 100% compatibility is probably exaggerated, since there are portions of the ODF standard (e.g. table styles, if I recall correctly) that are not implemented yet in either suite, and that could be implemented with different accuracy/priority in future. There are also cases of data loss originating from features that are currently implemented in LibreOffice only, see http://openoffice.org/bugzilla/show_bug.cgi?id=118037 But of course these are corner cases and, if not 100% compatibility, OpenOffice.org and LibreOffice do indeed offer >99% compatibility. Regards, Andrea.
