December 28, 2004

Bear with me on this.... read the following paragraph, memorize the general meaning, close your eyes, and...

Think.

This is the paragraph:

"I think the way forward is to focus on what it does that nothing else does well -- querying XML stores -- and not promoting it as an alternative to existing, more mature technologies."

Do you have it memorized? If not, here are the key phrases:

- Focus on what it does that nothing else does well -- querying XML stores.
- ...not promoting it as an alternative to existing, more mature technologies

Do you have it memorized now? How about changing the format to questions:

- Does this technology do something that another technology can not?
- What about this technology makes it better than pre-existing technologies that are already in use, mature, and understood by enough developers to make it useful in the real-world?

Answers:

- Ummm.... arguably yes. But then again... I guess it really depends on how dedicated you are to solving the problems with existing technologies or the next version of existing XML transformation technologies. Can the same result be returned by using another more mainstream, and as such proven language such as XSLT? And what about XSLT could we make better so that it could more easily solve the problems we are trying to solve with XQuery? Could libraries like FXSL and extensions like EXSLT do just as much with existing processors as we are trying to accomplish with XQuery (and the same goes for XSLT for that matter)?

So then what about XSLT 2.0? Could we use an XSLT 2.0 processor like Saxon, build a dynamic XML query language to our specific liking, compile it, using the subsequent compilation to query our XML data stores, and further use the language we created as our personal base language for querying the same XML data stores? Could XSL Scripting languages such as XSLScript, NiceXSL, PXSL, and XMLStarlet
be representative of such solutions, ultimately giving us the easier to implement and understand syntax we are looking for and yet not have to invest one single dollar in "new" technologies that ultimatelly give us nothing more that what we already have minus a few little nifty gadgets to play with here and there? Can we live without these little gadgets, finding ways to solve them with existing languages and processors? And wait, while were thinking about it, can't we use the same underlying XML processor to implement both XSLT 2.0 and XQuery? And if yes, why don't we just finalize the spec, build the damn processor, and work out the details of the language dynamically at a later date and in a syntax that we inherently like better just because we do?

Damn it I think we might be on to something here!

Question:

WHAT DO WE ULTIMATELLY GET WITH XQUERY THAT WE CAN NOT DYNAMICALLY CREATE AS WE NEED IT USING THE SAME PROCESSOR AS XSLT 2.0 IS BUILT ON?

OR

WHAT CAN WE SOLVE RIGHT HERE AND NOW WITH EASE USING SOME OTHER TECHNOLOGY THAT WE ALREADY HAVE OR WITH A TECHNOLOGY THAT BUILDS UPON AN EXISTING BASE OF BOTH PROCESSORS AND DEVELOPER UNDERSTANDING?

By the way, the above quote is from Mike Champion, somebody who kind of knows what hes talking about.

But wait, before you XSLT champions walk off in belief some battle has been won, you'd best think again. These are the questions that companies like Microsoft, and Sun, and Apple, and IBM, and Oracle, and BEA, and Google, and shutup M. we get your point, are having to ask themselves right this very moment. Why should we invest millions of dollars into these technologies when they don't give our customers anything more than what they have already and are simply not willing to accept this as such? Can we put these same millions into other technologies that further advance computing in the direction we believe it should be headed and ultimatelly give our customer base exactly what they are telling us they want, even though they can't quite find the words necessary to tell us in easy to understand terms (kind of like XQuery does? err, wait.. ummm... hmmmm... well... uhh...)? And ultimately...

Can't we just start over at the beginning, using customer experience instead of bureaucratic opinion as the basis for building the standards in which the future of the WorldWideWhatTheHellWillTheyThinkOfNext will be based?

I sure as hell hope so 'cause it seems that what we've got ourselves here is a mess... Anybody feel like writing some code and solving some real-world problems instead? Just a thought....

Oh, I got an idea... lets create an easy way to query XML stores using a familiar SQL-like syntax so that we can then further use that data within our applications to do cool stuff with. We could call it XQuery! And then we could... err, maybe we could just stop there and get on with it and stop trying to be the bigger and better alternative to everything and just kind of do what SQL ultimately does, passing the data on to some other process, written quite possibly in some other language it knows nothing about nor does it care as once its job is done it can kick back and sip on Pina Coladas until the next request to access or do something to some data arrives from some other process written in some other language by some other developer who simply wants to verify that the same 16 digits offered on a web form match that in which is stored on file, allowing it to continue accumulating cash flow like a good little server process is supposed to do.

I like the idea. It has merit.

Moving on...

---

In further developments here at <XSLT:Blog />, I have another Mike Champion quote to use for the XSLT Quote of the Day (that Mike Champion, he's always giving me such great content to document and write about... Thanks Mike! ;)... Sorry XQuery, but the giving holidays are now officially over... now all thats left is the celebration of the old year passing and the new year beginning.... kind of like.... no, I better leave it alone for now ;)

Have a happy day! :D