Story: Google: OOXML 'insufficient and unnecessary'

Posted by: Marbux (Friday 29 February 2008, 6:28 PM)

Extensions are bugs, not features

Albert said: "By using the namespace prefixes it also improves things like XML validation and extensibility compared to ODF."

Let's begin by taking extensibility off the table. A standard that allows vendor-specific extensions to be conformant is a standard in name only. A lawful international standard or technical regulation identifies a product, specifies all product characteristics only in mandatory terms, and requires that only products that match those characteristics shall be deemed conformant. See e.g., World Trade Organization Appellate Body decision in WTDS 135 EC - Asbestos (March 12, 2001), para. 66-70.

Extensions fall outside the definition of a standard. They may be great for folks who want to play feature war games with their own formats, but they are the antithesis of a standard. Standards are about taking the feature war out of the competition game by specifying a standard product, allowing competition based on efficiencies in production expense and on customer service. The standard product we are concerned with here are competing sets of file formats.

Extensions make software goods non-substitutable and create interop nightmares. Short story, if you want embrace and extend a standard, forget about claiming conformance. Both ODF and OOXML violate the law governing international standards because they allow vendors to claim conformance despite vendor-specific extensions.

Second, validation is next to useless in both ODF and OOXML because of the extension problem. In ODF, validation occurs after stripping all extensions from the file. In OOXML, validation is a two-stage process in which the markup specified in the schema is validated and then a second step that "validates" the extension's compatibility markup and checks to see that they use defined extension points. But that stage is virtually worthless because there is by definition no schema against which extensions' functionality could be validated.

In both ODF and OOXML, what is being validated is only part of the real-world file; the validation process itself is therefore invalid. See e.g., Validation - Wikipedia. ("Validation against an incomplete or insufficient set of criteria can lead to a state of "validated" where "validated" does not confer the confidence that the term intends. Thus validation of the validation criteria is an important aspect that is often overlooked.")

Third, I've already discussed the fact that terseness in markup is a negligible factor at best. You can argue a contrary position until the cows come home and it will not make it true. As I've already pointed out, a million-row ODF spreadsheet loads faster in Excel equipped with the da Vinci plug-in than an equivalent OOXML spreadsheet is loaded by Excel equipped with the OOXML plug-in. There are other factors involved that have far more dramatic effect on loading speed.

I intend to stop arguing this processing speed issue with you. You are obviously more concerned with defending an indefensible position than you are with having a frank discussion. I've got better things to do, like actually getting something done about interoperability issues.

--Buck ("Marbux") Martin
Universal Interoperability Council

View complete profile

Private message disabled

View all posts