Slashdot Mirror
XML Turns 5
GiMP writes "According to the World Wide Web Consortium, XML turns 5 years old today. XML is used by many programs as a generic container for data. Applications range from websites, to word processor documents, to video games. It seems like only yesterday it was only a working-draft."
In computing terms 5 years is quite a long time. I wonder what replacements are in development, if any?
It will probably be another 5, or 50, years before we know to what extent XML was the answer to the problem of data obsolescence and the degradation of old formats (like "bit rot", this is a handy but misleading way of framing the problem, which is not that the formats themselves degrade but that the supporting software infrastructure fragments, evolves or falls into disuse. The question with XML is then whether XML-encoded data will prove recoverable and intelligible after SAX, DOM, SOAP and all the rest have fallen into obscurity).
XML's self-description is one layer deep: data and metadata are packaged together. This layer can be seen as one layer of insulation against obsolescence: so long as the metadata remains meaningful, the meaning of the data can be ascertained and recovered. But the metadata is itself data, and if it too loses its meaning then it will be of no help at all.
For any data at all to have a semantic value it must have a context, and contexts change over time. XML is meant to ease the translation of data between contexts, but it cannot preserve meaning for all time.
Experience is a hard school, but fools will learn no other.
and how old is XML's grandpa, SGML?
the idea of structured text isn't THAT new!!
"bit rot" of supporting software is itself only a piece of the problem. even if we had XML decades ago, what good would it do you if you had perfectly formed XML data... sitting on your 8" floppy disk? IIRC google had to jump through hoops to get some older Usenet archives off magnetic tape, a feat that may not be possible at all in 5 years.
Quite so; a recent attempt at preserving old media is noted here.
With this in mind, may I direct the attention of budding geek archivists and antiquarians to Bruce Sterling's (and others') Dead Media Project, which seeks to document and analyse the conditions surrounding the life and death of media?
Experience is a hard school, but fools will learn no other.