Waiting for Godot No More: XQuery's Done!

After an almost interminable finalization and approval process, XQuery 1.0 is now out, done, a real standard, or -- in W3C parlance -- an official "recommendation." Here is the official press release, entitled "W3C XQuery 1.0 and XSLT 2.0 Become Standards: Tools to Query, Transform, and Access XML and Relational Data."

This is great news for MarkLogic because it formalizes the language around which our server is built. Roughly speaking, I feel like Oracle in 1982 when IBM had de facto standardized SQL (over Quel and other rivals) as the standard relational query language, yet no major vendor had a commercial implementation. The SQL standard de-risked purchasing Oracle and gave them about a 5-year headstart over the big guys. All they had to do was wipe out Ingres, Informix, and a few others to (in American Idol parlance) "go through" to the next round, in which they compete today, fighting against IBM and Microsoft.

But, you might wonder, don't major vendors (e.g., IBM) have various degrees of XQuery support in their products today? How does this provide Mark Logic with the same time-to-market advantage that Oracle had in 1982?

The answer lies in the focus of the XQuery implementation. Unsurprisingly, RDBMS vendors are focusing their XQuery implementations on XML data. As I often say, when you have a multi-billion dollar database business and you look at XML, you see data -- not content. One need only look in a few places for verification. Consider their:

• Poor or non-existing support for XQuery with full-text search
• Indexing strategies, such as xmlpattern indices, which create b-trees on paths within XML documents
• Benchmark focus, such as the IBM-backed transaction processing over XML (TPoX) project which is clearly data- and not content-focused
  

By the way, there's nothing wrong with the major RDBMS vendors focusing on XML data. Were I running product management at Oracle or Microsoft, I'd do the same thing. It's simply that none of them are focused on XQuery and XML content, which creates the opportunity for MarkLogic to do so with our special-purpose XML content server. So, when looking precisely at XQuery implementations focused on solving content problems (which is what we do at MarkLogic), indeed I feel like I'm Oracle and it's 1982.

Here are vendor testamonials and quotes, including one from me, about the news, and here is a link to the official XQuery 1.0 standard.