IBM's Sutor responds to ODF SDK analysis

In a blog post entitled Regarding an [OpenDocument Format] SDK, IBM's director of standards and open source Bob Sutor responded to the questions I raised yesterday:David Berlind asks over in his ZDNet blog where the SDK (Software Development Kit) is for OpenDocument.
Written by David Berlind, Inactive

In a blog post entitled Regarding an [OpenDocument Format] SDK, IBM's director of standards and open source Bob Sutor responded to the questions I raised yesterday:

David Berlind asks over in his ZDNet blog where the SDK (Software Development Kit) is for OpenDocument. It's a good question

Sutor covers a lot of ground in his post, talking about how it should be open source (a point I made yesterday), comparing it to what's available from Microsoft, what it should be capable of, what languages it should support, and how an ODF SDK should come to market.  I'll take his quotes in these categories one by one.

...it is a good idea and it should be open source. We want everyone, open source and proprietary, to be able to use it everywhere.

Making sure it is equally available to developers of proprietary software, proprietary services (eg: online services where the code behind the service isn't visible), and open source software is critical.  ODF's biggest challenge may not be winning battles like the one it had against Microsoft recently in Massachusetts.  In the end, it's all about adoption.  What can you do to empower developers and make their lives easier.  I have spoken to several developers in a position to support ODF.  All they have to says is, "when there's an SDK." 

Maybe others know of one, but I know of no such SDK for the older Microsoft binary formats that is freely and openly available for developers to use on any platform without actually purchasing Office or paying for a development subscription.

Comparing to Microsoft's existing situation for its older formats is probably a moot point.  It probably remains to be seen what sort of SDK Microsoft will make available to the aforementioned set of developers for its newer more open formats.  Rather than compare to what the competition has done, anticipate what the competition will do.  Whether it's ODF or any other technology vs. Microsoft or any other competitor, that's what I'd be thinking about if I were running a strategy for the company I worked for.  In this case, Microsoft has already proven its resolve.  Both sides have already dug deep -- very deep -- for whatever resources were necessary to win. 

With Microsoft's former top open source thinker Jason Matusow now having almost exactly the same title as you Bob, one has to assume that Microsoft is prepared to release an open source SDK for its newer formats if that's what Matusow thinks it will take to emerge victorious.  In fact, that's what I'd recommend to Matusow if Microsoft hasn't planned to do that already.  For Microsoft to release an open source SDK for its newer Office file formats wouldn't be much of leap from where it is now in terms of having pretty much unencumbered those formats (for use by developers of all types including open source developers).  Not only has Microsoft already open-sourced other technologies, wth its recent IronPython release, coming up with a Python-based SDK that could be used by just about anyone (especially Web-based productivity apps like Writely and INetWord) is't much of a stretch either.  And, if Microsoft doesn't release those SDKs on open source terms, then having done so for the ODF SDKs is an adoption lever.

Minimally it would provide some sort of code to make it easier to read an ODF document into a software application and then write it back out again to disk.....An ODF file is a zipped collection of XML files....Most people would probably want an API that abstracts away the Zip and the XML aspects of things. That is, there would be some model for what an ODF document and its associated metadata looks like and an API for putting information into and taking information out of the model....We would need pretty broad language support for the SDK. Minimally I think we would want Java, C++, C#, PHP, and Python.

Perl might not be a bad option to consider in well.   I don't know if there's any AJAXy-code that could be released as well.  But that might not be a bad thing to look into. For example, given the Cut and Paste Web that  Microsoft's Ray Ozzie introduced yesterday, is there any work that JavaScript could be doing to preserve any formatting of the text that's been cut, and then seasoning it for pasting into something (local or hosted) that's ODF compatible?  Also, bear in mind that, of the languages listed, Microsoft is involved with three of the five (and there are third party Perl tools for Microsoft's .NET such as those from ActiveState).

The first SDK might not be the best. There should be some good market competition to come up with the best one.  Do I think this will all happen? Definitely, given the nature of the ODF and open source communities. Let me know if you are aware of anything happening in this space and I'll pass it on. If we at IBM get involved in this sort of thing, I'll pass that along too.

I still feel as though the lack of any such SDK so far has been the biggest miscalculation on behalf of the two biggest proponents of ODF; IBM and Sun.  Yes Simon, I'm aware that it's not just IBM and Sun that are behind ODF (as you so succinctly pointed out in your blog a couple of days ago).  But over time, Sun and IBM have been the power behind ODF.  No two companies have remained more committed or sunk more resources into getting ODF to where it is today (some critics would argue "not very far"). So, when I read these words from IBM's Bob Sutor, it sounds to me like the existence of an SDK is being left for the market to decide (and compete over).  It certainly doesn't sound like IBM has anything in the works right now and I can't speak for Sun (although it has, in the past, acknowledged that such an SDK would be a good idea). 

My guess is that the cost to produce such an SDK, release it as open source, and make it the standard, is probably a fraction of that which has already been spent on ODF by the two companies combined already.  Leaving the fate of an ODF SDK -- a lynch-pin to that all important developer adoption --  up to the market?  Unless Sutor is simply being coy about a secretive project, this news comes as quite a surprise to me. IBM and Sun usually have all the bases covered with something as important as this.  The fate of ODF might as well be left to Microsoft. 

Editorial standards