ODF VS OOXML: Library support shootout
28 November 2007
Web and Application development
One of the core reasons for wanting an XML based document format is that third-party applications and web services will be able to create, import and process them. For that, one needs language bindings. There are three levels of abstraction available:
- Firstly, you can interact with these documents (at least with ODF) at a pure XML level using just XSLT, XQuery and so on; the basement approach. This is, of course, somewhat painful and time-consuming to work at this low level.
- Secondly, you can have libraries that map ODF or OOXML documents to native datatypes (and back again) in the language that you use; to follow the analogy, this is the main floor.
- Lastly, you can have APIs that allow you to control and automate desktop applications such as OpenOffice, Word and Excel; we can can think of as the loft or rooftop. This is pretty-much useless for applications that run on a server of course.
It is the rooms on the main level, the libraries for programming languages, that are the most useful and widely applicable for application and Web development.
How well is ODF and OOXML supported in different languages?
I started off by choosing six common languages used by application and Web developers, then I went and looked for what libraries are available for each document format:
| Language | ODF | OOXML |
|---|---|---|
| Python | Yes 1 | No |
| Perl | Yes 2 | No |
| Ruby | Yes 3 | No |
| PHP | Yes 4 | Started 5 |
| Java | Yes 6 | Started 7 |
| Microsoft C# | Yes 8 | Yes 9 |
So a pretty clean sweep for ODF at the moment, though as you can see, there are a couple of projects just starting up to provide Java and PHP support for OOXML, how far they get will be interesting to see - is it really possible to create a complete and valid OOXML document without Microsoft software? There is a big difference between getting some text out of a OOXML document and making a library that fulfils the massive ECMA specification.
The Mono project are (or at least were) trying to make C# bindings work on non-Microsoft platforms. Hopefully, if OOXML takes off to any degree then some bright spark will write Python bindings too, even if just to help us all rescue data out of OOXML documents.
I'm sure if I took less well known languages, the same trend would be found - the ODF library is produced first and provides the most complete bindings. ODF has a number of advantages to make this the case:
- Firstly, ODF has a couple of years head start over OOXML. A real implementation of OOXML did not appear until 2007, while ODF has already become the default office format for the Free Software/Open Source world (if you don't include plain text).
- Secondly, because ODF is much purer XML than OOXML, making an ODF library for a language is not a lot of work if the language already has well developed XML support.
- Thirdly, ODF is not owned or controlled by a single company, other IT organisations are not in any rush to support OOXML or help Microsoft maintain it's office monopoly. There are only so many times that you can pee in the pool before no one wants to swim with you anymore.
Please let me know using the comments if I have missed a project out. I know that OpenDocument in some cases has language support from multiple vendors which is nice, the OpenDocument Fellowship has a more complete list of libraries, among other useful lists.
1 Tim says...
this is a test of the emergency broadcast system...
Posted at 3:39 p.m. on November 28, 2007
2 Zeth says...
Hi Tim,
"please state the nature of the medical emergency"
Posted at 3:41 p.m. on November 28, 2007
3 bug says...
"there are a couple of projects just starting up to provide Java and PHP support for OOXML"
Shouldn't it be vise versa? As PHP and Java are the ones who get the OOXML support. OOXML by itself doesn't support PHP nor Java.
Posted at 6 p.m. on November 28, 2007
4 Joe says...
> There are only so many times that you can pee in the > pool before no one wants to swim with you anymore.
Nice analogy ;-)
Apparently however, even though the MS pool is bright yellow, and has what look like candy bars floating in it, there are gold doubloons scattered around the bottom.
People are camping in line for a chance to dive in.
BTW, you might also include the actual command line in your list, because it's quite possible to do some useful operations on an ODF file using command line tools.
E.g., list all the titles in a presentation:
$ unzip -p ~/doc/ODF_What_Who.odp content.xml | sed -e 's!<draw:frame !n&!g' -e 's!</draw:frame>!&n!g' | grep 'presentation:class="title"' | grep -o '<text:p.*</text:p>' <text:p text:style-name="P1">ODF: Open Document Format</text:p> <text:p text:style-name="P1">ODF: What Is It?</text:p> <text:p text:style-name="P1">ODF: So what?</text:p> <text:p text:style-name="P1">ODF: What Is It?</text:p> <text:p text:style-name="P1">ODF: What Is It?</text:p> <text:p text:style-name="P1">ODF: What Is It?</text:p> ...
Granted, not as nice as a real library, but useful at need, and readily available.
Posted at 7:18 p.m. on November 28, 2007