[REBOL] Re: Parse limitation ?
From: g:santilli:tiscalinet:it at: 8-Oct-2003 22:04
Hi Petr,
On Wednesday, October 8, 2003, 6:57:38 PM, you wrote:
PK> Well - I am not sure my example will be any slower, except the penalty
PK> of extra function call. First, I pass it string at certain position and
First of all, FIND searches char by char too. It's just way faster
because it's native; but, if you end up searching the string n
times, you get n*m complexity (where m is the size of the string),
and this scales up so badly that in the end it gets slower than
using a PARSE loop.
Probably FIND is still faster for two or three alternatives. We'd
have to test it. When the alternatives are just strings, you could
speed up the PARSE loop using a charset, and I have the feeling
that PARSE is as fast as FIND in such a case, so the PARSE
solution would be n times faster for n alternatives.
PK> Yes, I can imagine it, really. The problem is (at least for me), that I
PK> am able to understand such grammar once someone creates it, but am not
PK> able to come up with it to solve problem at hand. Will you blame us
PK> little bit underskilled rebol programmers now? :-)
Not at all, but you are underestimating yourself. ;-)
PK> Sounds interesting. I am just curious, if e.g. html only (not trying to
PK> complicate things with java-script for now :-) browser would be possible
PK> with Rebol? IIRC Python has web browser. Just curious.
The problem for a web browser is not HTML parsing, it's rendering.
In my dream-future, I will finish the PDF Maker 2 and then write a
HTML2PDF translator. Rendering in View would be possible too, but
I'd like RT to offer us some kind of native rich text handling
first... you see, I'm too lazy to do all of that myself. ;-)
Who needs a REBOL web browser? I'd like an email client much
better.
Regards,
Gabriele.
--
Gabriele Santilli <[g--santilli--tiscalinet--it]> -- REBOL Programmer
Amiga Group Italia sez. L'Aquila --- SOON: http://www.rebol.it/