In 2005, the Commonwealth of Massachusetts enacted rules prohibiting state agencies from using proprietary software, reasoning that citizens shouldn't have to pay to view documents created by their own government. But the open source software tested by Massachusetts couldn't effectively open documents created using Word and Excel (which it would have had to do to enable the proper functioning of state bureaucracy), dooming the effort.
Sam Hiser, former executive director of open source document consultants Open Document Foundation, lays this failure at the feet of the vendors, including Sun, IBM and Novell, whom he says have co-opted and politicized the standards-making process into a "Kafkaesque joke exemplifying vendor ambition, inexperience and stupidity."
But there are a couple of competing arguments for why the Massachusetts trials failed:
- Open source maven Glyn Moody blamed the Massachusetts failure on perfectionists "obsessing" over a level of interoperability he maintains can never be achieved.
there will *never* be anything like a perfectly-compatible solution with Microsoft's OOXML, given the 6000 pages of documentation, and the presence of opaque binary blobs.
- Computer scientists Rajiv Shah and Jay Kesan argue that Massachusetts failed to adopt an open source document application because the state was "effectively locked into OpenOffice.org," (which is sponsored by Sun).
[The] OpenOffice.org code commit process reveals it is not in practice an open source project... Sun engineers (IBM has forked and do not pretend to offer access) reserve commit access for themselves. Michael Meeks at Novell had to fork OpenOffice to contribute to it; that is another reason why files from Novell's version of OpenOffice.org are not interoperable with files from IBM's Lotus Symphony, with files from Sun's OpenOffice.org, with files from KOffice, with files from Ubuntu's "deskto" version of OpenOffice.I reached out to Meeks, who basically confirmed Sam's version of events, writing "I don't know that we fork it, but we do distribute a heavily patched version."
More importantly, Meeks validated Sam's description of the so-called community at work within the OASIS standards body:
Sun abuses its ownership of the OO.o [open office] code-base in it's commercial interest. Sometimes this means (incredibly) excluding other players (to keep this a "Sun" project).Meeks also agreed with Sam's view that Sun representatives have gummed up the works in order to support its corporate agenda,
"Open"Office.org is perverted away from being a community project that exists to pursue excellence... and instead towards something that Sun is happy to stifle if it helps Sun's agenda... Sun screw around all the time with OO.o, they treat it as their own, and badly want to keep it that way -- other large corporations are hardly welcome in OO.o land.Meeks also noted that Sun has "totally failed to build a volunteer developer community, or even a corporate developer community around OO.o."
The lack of interoperability doomed Open Office in Massachusetts, but it's unlikely that public officials in other states, not to mention developing nations, will cave in to Microsoft. Their alternative? Perhaps they'll live with imperfect interoperability, or perhaps with one of the "forked" versions of Open Office from IBM or Novell, as Moody suggests.
The other, more appealing alternative, is that Sun becomes more accomodating after the Oracle acquisition closes. Meeks wrote, "there is always hope that Oracle may adopt a different and more positive approach to the [open office] code-base," and that if IBM and other corporations were to allowed to contribute code, "we would have a really good chance of achieving feature parity with Microsoft Office, if not exceeding it in many areas."
[Image source: Wikimedia commons]