Google Docs OpenDocument Fail

Would you believe Google Docs doesn’t actually support ODF, at least for spreadsheets? According to some reports on the web, it only supports the older StarOffice format (despite having an option labelled “OpenDocument” on the export menu). Every time I try and upload an ODF document generated by OpenOffice 3, it fails with an unspecified error. So to get documents for local working before re-uploading them, I have to use XLS :-((

6 thoughts on “Google Docs OpenDocument Fail

  1. Yes, love you Google! This is exactly the right direction to go, get rid of those od* formats!

  2. Pretty sure that’s a regression in the past few months. I had been using a .ods to import dates (because Spreadsheets doesn’t support adding date ranges, but finally added the “cloud clipboard” for document-to-document pasting) for a while, but recently when I added new dates to the spreadsheet and went to re-import, it failed as you described.

    That said, it’s quite an annoying regression/bug.

  3. Note that OO.o 3 (and I think also some of the more recent 2.x versions) generates ODF 1.2, which is not entirely compatible with software that supports older versions of ODF (used in, say, OO.o 2.0 — which is different yet again from the .sxw and so on used by even earlier versions). Try taking a document from OO.o 3 and opening it in OO.o 2.0, and it won’t work. This is true for Writer documents, as well as the spreadsheets. (I’m not sure if the other formats changed or not, since I haven’t yet discovered a use for any of the other ones.) I don’t happen to know what format Google Docs uses, but it’s possible that it supports the older version of ODF and not the newer one.

    For a while, there was also a new version of OO.o that couldn’t correctly handle spreadsheets created with older versions. (It could open them, but it where computation results should have been it said ERROR most of the time. Something about empty cells no longer evaluating as zero in numeric context.) I think that’s fixed in recent versions now, though.

    Personally, I think a format change that introduces incompatibility should always mean a new Content-type and a new filename extension, but apparently the OO.o people feel otherwise. (They wouldn’t be alone. Microsoft Word has gone through at least half a dozen different versions of its .doc format that I’m aware of, quite possibly twice that many.)

  4. ODF has nothing to do with OpenOffice.org, so format changes are completely outside of their control.

    What exactly is it that Google is supporting? If it’s ODF, then perhaps Google should re-name their “OpenOffice” format to “Open Document Format” because OpenOffice’s original format SXC (pre-ODF) is completely different.

    I just exported a spreadsheet from Google Docs and according to file, it is:
    Zip archive data, at least v2.0 to extract

    Alternatively, saving the same spreadsheet from OOo reveals:
    OpenDocument Spreadsheet

    Definately something’s not right with what Google are doing. They should make sure they are supporting ODF and SXC, but make the differences clear!

    -c

  5. > according to file, it is: Zip archive data

    That’s true as far as it goes, but it’s kind of like peeking inside the wrapping paper on a birthday gift and saying excitedly, “It’s a box!”

    All ODF variants, as well as the old StarOffice formats, Java archives (JAR), and various other things use the old PKWare Zip format as a generic container. This is the same format that info-zip handles, among other software, and recent versions of Windows call it a “compressed folder”.

    It’s what’s inside the archive that really matters. If you take any OpenOffice document, rename it with a .zip extension, and double-click it in a file manager, you’ll see a directory hierarchy inside populated mostly with XML files, plus any images you’ve embedded and a manifest and a couple of other bits and pieces. The differences between the old StarOffice format, the old ODF that OpenOffice 1.1 used, and the new ODF that current versions use, all have to do with various details of the markup in those XML files inside the archive.