2005-06-09

Are XML documents concrete or abstract?

A while back, Dan Connolly and I argued this point. He said that XML documents were like integers: just because you write a new one down doesn't mean you've created anything new, and likewise writing something to a file doesn't cause the set of XML documents to grow either.

But I think an XML document is first and foremost a document, and documents can be created and destroyed. The sentence "John Rumpelmayer Jones sold me a goat, a bridge, a boat, a wedge, a coat, a dredge, a groat, a hedge, a stoat, a pledge, a moat, a midge, and a vote." is a grammatical English sentence, and a fortiori a document, that never existed before this posting.

If there is a poem beginning "In Xanadu did Kubla Khan" (and there is), then is there a poem whose text is solely:

Singing of Mount Abora

and every other possible poem-fragment?

The translation of this poem into German exists and is a poem (it begins "Im Xanadu ließ Kubla Khan"). Does the translation into Zarphatic (the dead Jewish language which is to French as Yiddish is to German and Ladino to Spanish) also exist and also constitute a poem, regardless of the fact that no one has ever made such a translation, nor will (very likely) in future? I think not.

2 comments:

Anonymous said...

This is the kind of thing philosophers get interested in. One reason for saying integers exist even if we don't write down or utter a representation of them is that there's a formal system describing them that we subscribe. In said formal system, the existence of numbers we haven't named is (apparently) implied.

"An integer is anything that has the following characteristics: (list)"

With XML documents, it's a little more fuzzy; there's a formal system describing what constitutes XML documents: "An XML document is anything that has the following characteristics: (a different list).

So the cases are parallel as far as that goes; if there's a felt difference, perhaps it's that we think the formal system describing the integers gets at an independent reality, while we can actually trace the invention of the formal system describing XML documents, and that system exists in order to solve certain kinds of problems, not because it describes an independent reality.

Anonymous said...

Gar, that should be "to which we subscribe" in the first paragraph.