ODF Add-in for Microsoft Word

ODF Converter Team Blog

Thursday 30 November 2006

Launching of 0.3-M1 release

Last week we released version 0.3-M1 of the converter. What do those numbers mean?

  • 0.3 means that we are now working mainly on the reverse conversion (from DOCX to ODT); the direct conversion will still continue to be improved, but it will be far less visible than during the previous months (we fixed a lot of bugs since the last release, though - the number of open bugs on SourceForge dropped from more than 100 to less than 50 at the time of the release)
  • M1 stands for "Milestone 1" and corresponds to a set of features that were implemented according to the roadmap of the project.

For simple documents, the reverse conversion works quite fine, allowing users to manipulate OpenDocument text files directly in Word. Our main concern is now to make the process of opening an ODT file and saving it back to ODT as accurate as possible. That means that if we have to implement workarounds to convert features that are not directly available in one format or the other, those workarounds will have to be preserved during the reverse conversion. To ensure that this process works fine, we iterate it several times on one file, and see the final result as something we could call the "fix point" of the converter (refering to a famous mathematical theorem - but I'm not sure of the english name).

Once we have an acceptable result for direct / reverse conversions, we will enhance our transformations so that they can also work correctly on legacy doc files produced by previous versions of Word (there are tons of features that are marked as deprecated in the OpenXML specification).

And to finish, a good news: Google team finally fixed the problem of ODT export in Google Docs. You can now import those documents in Word without problem!

Friday 3 November 2006

What's the problem with Google Docs?

You may have learnt that we've just released a new version of the ODF Translator yesterday, called 0.2-Final (even if it's not really a final version, as I've already explained on this blog), providing a lot of new features for the direct conversion (from ODT to DOCX) and a prototype of the reverse conversion (from DOCX to ODT). While the reverse conversion is only at its begining, the direct conversion now looks quite good and provides most of the features that we expect from a converter. A complete list of features is available on SourceForge. The main functionalities that are still to develop are:

  • digital signature
  • encryption
  • section protection
  • drop caps

along with several other small features. The roadmap for both direct and reverse conversion is also available for download on SourceForge.

This release has been intensively tested by our test teams from Dialogika and AztecSoft. Test scenarios are described here and here. But this time, we also wanted to have an overlook of how the converter was behaving with real-life files. To this purpose, Dialogika gathered more than 450 ODT files on the internet and here are the results:

  • 411 documents were converted, validated and opened successfully in Word
  • 10 documents were valid, but could not be opened in Word
  • 28 documents were invalid but could be opened in Word
  • 7 documents were invalid and could not be opened in Word

Of course, this does not show anything about the way the documents were actually rendered in Word! But anyway, the results were not as good as we might have expected, so we made some quick investigations to identify the issues that lead to those results. After a one-day exploration, most of them could be fixed, and we therefore decided to publish a "Hot Fix" for the 0.2-Final next week. It won't add any new feature, but will fix those file crashes.

During this "real life" tests, we noticed that all the files created with the online application "Google Docs" were not converted successfully. This was strange enough for us to look in detail at what was wrong. And we found out that Google Docs was simply not able to export to ODF. Actually, the file menu says "Save as OpenOffice" and not "Save as OpenDocument". The output file is an SXW file (the legacy format from previous versions of Star Office and OpenOffice.org)... with an ODT extension! I don't know if by doing this way the guys from Google wanted to make people think that they had implemented the ODF format, but that was a nice try! ;-) I guess that they are working hard to achieve the compatibility, but in the mean time our converter won't be able to open documents made with Google Docs - no need to complain, we have commited to handle OASIS OpenDocument format, not all the formats of the earth!

Powered by DotClear

Project page on SourceForge

SourceForge.net Logo