Cart
[Home] [By Thread] [By Date] [Recent Entries]
[Jonathan Robie] > At 10:57 AM 6/19/2001 -0700, Murali Mani wrote: > > >a) semantics is more important than syntax, but probably they are also > >linked together Semantics, syntax, and there's a third dimension: human comprehension. The syntax must promote human understanding. Even with SQL, for example, it's so easy to have a query return something other than what you wanted, and often it's very hard to be sure if the results are correct or not. To me, this dimension rules out an xml syntax as the only or primary syntax, because the xml has too much visual noise. An xml syntax would have to be considered an adjunct for machine processing, as it seems the current plan has it. Even when queries will be mainly machine-generated, a human is going to have to develop and debug the query generator, again emphasizing the importance of human comprehensibility. This entails certain consequences, like these: 1) minimal context-dependence 2) a judiciously small amount of operator overloading (e.g., "+" for string concatenation) - too much overloading gets hard to remember. 3) minimal exceptions, preferably none, so use of syntactical features is consistent. 4) recursion - e.g., expressions may contain expressions with minimal limitations, queries may make use of other queries. I don't know if the XQuery effort has a set of guidelines like this, but it looks as if they have been thinking somewhat along these lines. Perhaps Jonathan would care to comment on this? Cheers, Tom P
|
