[
Lists Home |
Date Index |
Thread Index
]
At 4:48 PM +0100 4/15/04, Michael Kay wrote:
>I personally don't think that XML parsing is that time-critical to most
>applications, but I can imagine that there are applications where avoiding
>the cost of repeatedly checking (for example) that every name is made up of
>valid characters would be beneficial.
A lot of parser use naive algorithms for this task that are far from
the most efficient possible. For one thing, the cost of checking a
single character can be made O(1) with a very low coefficient, at the
cost of about 32K of memory. Furthermore, you can cache previously
verified names and not recheck them when they're encountered 1000s of
times apiece.
I'm not convinced even the best parsers have hit fundamental limits
yet. It's way too early in the game to say we a binary format for
speed purposes. Time would more productively be spent by exploring
better algorithms for standard parsers.
--
Elliotte Rusty Harold
elharo@metalab.unc.edu
Effective XML (Addison-Wesley, 2003)
http://www.cafeconleche.org/books/effectivexml
http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/cafeaulaitA
|
