Another step forward, down the long path of standardized meta-formats.
Being able to describe and manipulate information at ever-higher levels of abstraction is essential to increasing productivity and unlocking the true potential of computers.
One way to look at XML is that it does the same thing for standardizing the structure of information that ASCII and Unicode did for standardizing the character encoding. In the same way that standard character encodings allowed a whole range of generic text processing tools to be written and used with a whole world of data, XML enables the creation of generic tools that deal with the structure of a whole world of data.
IMO, its biggest problems are in the area of schemas. DTDs lack sufficient expressiveness, but are tied to the standard at a very fundamental level. Alternate schema languages exist, but none enjoys anywhere close to the level of support that exists for DTDs, and the officially endorsed altenative (W3C Schema Language) also seems to raise the most complaints.
