While he's right about a lot of things (especially that the ODF people could have spent more time on their spec rather than fighting against OOXML), it doesn't change the fact that even the most recent versions of MS's own products will generate files that:
1. Don't validate against the OOXML schema.
2. Contain large blobs of encoded binary data from previous versions of MS Office in the XML files.
For this reason alone (that the spec isn't followed correctly), I still prefer ODF.
Honestly, I'd love to have all devel builds of any software that writes XML as output have to validate that XML against the schema before it's written to a file, and have passing that test be part of the the acceptance criteria for the build.
This isn't hard to automate - it's very basic test driven design stuff. I do the above in my own code, and all it takes is a single extra line...
I was one of the first people to reverse-engineer the binary Word file format and was one of the early contributors to POI.I've spent a career writing code against various file formats.
Unfortunately, most of the comments here are not based on first-hand technical knowledge. How many of you have tried to code against the OOXML spec? I have. It's pretty good.
Most people in the community of "people actually doing stuff in this area", which Miguel is a part of, feel like Microsoft is acting in good faith. See http://www.opensource.org/node/351.
He can slam Groklaw as much as he wants, but I've still never seen them offer anything other than rational and backed interpretations of the presented facts.
I particularly love how he thinks ISO approval of OOXML is some sort of silver bullet against all allegations against him. Everybody in the know is very aware of the nastiness that went on with that. And as zdw in this discussion notes, it's ISO approval doesn't mean jack in reality.
> He can slam Groklaw as much as he wants, but I've still never seen them offer anything other than rational and backed interpretations of the presented facts.
How hard have you looked? They republished a lot of IBM's anti-OOXML FUD, and if anyone tried to challenge that in comments (well-researched comments, with cites to the specs for OOXML and ODF proving IBM lied), those comments would quickly get IP restricted. (Instead of deleting comments, Groklaw flags them so that they are only shown to people from the same IP that the comment was left from).
If Java was an ISO standard, Harmony would have been a legal Java implementation, Apache's vote would have meant something, supersets of the standard would have been allowed. This and the fact that the standard would have been kept small to accommodate all versions of Java, and the lawsuit against Google's Android wouldn't have happened.
As it is right now, Microsoft cannot sue alternative implementations of their ECMA standards for .NET. They can only do that for parts that aren't in those standards.
Also, in reality many governments like ISO standards.
Groklaw ... I've still never seen them offer anything other
than rational and backed interpretations of the presented facts
I would take unbiased over apparently_rational_but_filled_with_emotions any day of the week.
ISO approval doesn't mean jack because even if you follow the standard, you won't be able to load supposedly standardized documents that were generated by Microsoft's software 99/100 times.
This is just FUD ... I never had problems opening MS Office documents in OpenOffice, except a few problems here and there with formatting: but it's not like browsers do any better.
And I'm talking about reverse engineered formats here, for which there was no documentation, ISO or otherwise.
Also, have you ever implemented an ISO standard to be able to speak about it? I have.
The argument against OOXML is that it cannot be a standard since it says things akin to "must work like Word XXX here" - ie, it's just a bundling of a proprietary format inside XML.
Moreover, this is just not the way to "clear the air". I know relatively little about all the events but this kind of ad hominem rant sure sounds like something someone in a conspiracy would write ...
There is nothing whatsoever like that in OOXML, nor was there ever anything like that. That was just FUD (largely from IBM).
Here's what there actually was. There are older documents, from a variety of older programs (not all Microsoft, BTW), that do things that cannot be represented in OOXML (or ODF, for that matter).
There are many organizations that have documents in those formats, and have tools that understand those formats. E.g., a law office might have a lot of Word Perfect documents, and might have tools that understand WP format and can automate various things involving those documents.
Such organizations might want to update their document store to use a more modern format, like OOXML or ODF. However, they might need to preserve some of those things in the old documents that are not representable in the new format.
Suppose the standard says nothing about this. What should they do? Simple--they will find some way to add that extra information to the document. Both OOXML and ODF provide a couple of ways that one could include non-standrd. So what would likely happen is that everyone who deals with old documents would come up with their own in-house way of embedding this extra information.
Eventually, people would get smart. We'd realize that if you are embedding information about WP5 line spacing options in your OOXML or ODF conversions from WP5, and I'm embedding information about WP5 line spacing options in my OOXML or ODF conversions from WP5, then we should get together and agree to do it the same way--and we should get everyone else involved. We should come up with some kind of industry standard for storing WP5 line spacing options in documents converted from WP5 to OOXML or ODF.
Same goes for a bunch of other things from old word processors and spreadsheets. We'd eventually end up with a bunch of industry standards for all of these for OOXML and ODF. Hopefully, we'd combine them all and have one standard.
OK, now back to what they actually did in OOXML. They simply realized people would be doing this, and so reserved some names for them. That's all. Nothing in the OOXML spec said that you have to support WP5 line spacing, or know what it means, or have your OOXML programs be able to do it. All the spec said was IF you are going to extend the document to record WP5 line spacing (or the other specific features of old formats they identified), THEN use this as the name of your tag or attribute or whatever that marks it.
They also said that new documents should not use these things. It was strictly for people who were converting old documents and needed to keep that information.
I've yet to see anyone offer a rational objection to this.
The prevalence of ad-hominem attacks against Icaza is unfortunate. You can agree or not with his more-pragmatic-than-some stance towards open source software, Microsoft, etc., but the fact remains that the guy is a first-rate builder and hacker.
There have been a seemingly endless stream of incidences of him shilling for MS, going on for several years now. The two most common themes of shilling are trying to push Mono as a viable linux technology, trying to push Mono as a dependancy on common linux systems, and shilling for OOXML like you see here.
Another (slightly less common) theme is his ad hominem attacks on [F]OSS figures/organizations/supporters such as the FSF and Groklaw, like you see here.
He's been at this for years, it's no stretch of an imagination to call it a "long history".
1. Don't validate against the OOXML schema.
2. Contain large blobs of encoded binary data from previous versions of MS Office in the XML files.
For this reason alone (that the spec isn't followed correctly), I still prefer ODF.