quote:
Since the big XML push of several years back, the first time I found a practical use for it was this year as way of writing out some meta-data that needed to be re-parsed. For this purpose, it saved me countless hours of writing a custom parser.
Repeatedly, I see "parseability" as one of the stated benefits of XML; people use XML because it's easy to support, and because there is a large base of written code that does much of the "legwork" of interpreting and manipulating XML files. But these are not benefits of XML per se; these are benefits of vendors' support for XML, and had vendors invested the amount of time, money, and effort that they've spent on XML on other, more-efficient file formats, then we could derive the same benefits from those file formats. We wouldn't have to write parsers, as they'd already have been written; we would have easy ways to recreate tables from these other file formats, easy ways of translating from these formats into other formats, we would have formats that are just as readable, just as useable, and just as easy to develop with, but formats that are far more efficient as a means of data transfer and storage.
I use XML, I know how easy it is to use, and I also know how it easy it is to understand; I enjoy the ability to interface with SQL Server via XML. But I also realize that these things have nothing specifically to do with XML. It's easy to use because vendors have made it easy to use; its readability and understandability are not unique to XML. All of these benefits, I believe, result from its base of support, and really, that is the only benefit that XML has: its broad base of support.
The question is, "why does it have that broad base of support?" As Mr. Peterson has pointed out, it isn't founded on mathematical or logical models, and has yet to be proven to be a sound data model. As a language for data transfer, it is bloated, as even its proponents will agree. Readability is nice, but again, that's not unique to XML; in fact, one might point out that it still lacks somewhat in the readability department or there would be no reason to transform it into other formats for end users. And again, any other benefits I see result from its existing support; e.g. it's faster to develop an XML based solution, not because XML is a better language, but because parsers such as MSXML exist, web services exist that are built on XML, etc. Time and again, I fail to find any foundation for XML that is based on anything other than a sort of "bandwagon": vendor X has support to it, and that makes it easy for us, so let's use it too.
If that is its only foundation, then we have no way of proving that it will indeed be here for a long time to come, unless we assume that the opinions, fads, and interests of the computer industry are fairly stable and seldom change; I've yet to see the computer industry prove that to be the case.