> Wouldn't the effects of recursion will be lost in the static,
> compared to the effects of loading the doc into memory to facilitate
> tree processing?
That produces slow processing, not a hard failure (unless indeed there
is simply too much document for even virtual memory). Java, and
all other HLLs I know of, provide no way to recover from
stack overflow, short of starting the app all over again with
a command-line switch for a bigger stack.
A general-purpose routine ought not to generate a preventable
hard failure no matter what the document looks like, IMHO.
> BTW, what languages can be relied on to do tail recursion?
Scheme and ML and their descendants. The Scheme version of
Stephen's algorithm will detect the tail recursion, and will
be recursive down the tree and iterative across it.
Indeed, Scheme *has* no (primitive) way to do iteration except
with tail recursion (there are macros that syntactically sugar
this, if you want). As a result, Scheme compilers can concentrate
on making the very few constructs they have to understand
(function call, function closure, assignment, IF) very very
efficient.
> Also, shorter algorithms are better. -Tim
But constant-space algorithms are better too.
-- John Cowan http://www.ccil.org/~cowan cowan@ccil.org You tollerday donsk? N. You tolkatiff scowegian? Nn. You spigotty anglease? Nnn. You phonio saxo? Nnnn. Clear all so! 'Tis a Jute.... (Finnegans Wake 16.5)