> > I know this is very much to ask for, but i think it would make it a
> > hell lot easier to parse xml et.al.
>I think that's a whole 'nother can of worms. Current thinking here is
>that "parsing" html using scripts is probably the wrong way to go.
>Instead we should consider a way to convert it directly into an object
>architecture, and access (or edit) it using that, and then convert it
>back to a text string if necessary.
How would that look like? I don't think we should go to much for
HTML; XML has a lot more to give. And there may be other things
people may want to parse, be it RTF, TeX, real SGML, whatever.
The main point I was trying to make was that the discussion if
regular expressions should eat the minimum or maximum number of
characters is not enough. Balancing is a most important issue.
> > P.S.: Two things i'd like to see in xTalk that are not worth
>their own thread:
>> 1. "the ms" as an abbreviation of "the milliseconds"
>
>Opinions on this? How many times do you think "ms" has been used as a
>property or variable or user function out there?
I would guess not so much. It is not very speaking but for
milliseconds and not as short as "m". And I can use "a" as a variable
name, too >;-)
> > 2. MetaCard has call-by-reference, but you can not see that in the
>> call. One of the great things about C is that you see in the call
>> that this variable is to be changed. How about adding the "@" to the
>> call as well?
>
>Actually passing things as pointers is pretty much obsolete in C++
>where one usually uses references for this. And in that case, you
>*can't* tell from the call that it's going to be changed or not
>without looking at the function (or at least for a "const" parameter
>in the prototype, which makes the problem somewhat easier than in
>xTalk).
That's why I said C, not C++. I have allways wondered why they did
not force you to pass references to calls-by-reference as well as in
the called function. But that is a different story that has to be
told at a different time :-)
>Would it be sufficient to just add this as "syntactic sugar" and
>have it stripped off, or must it actually generate an error if they
>don't match?
A stack property? I would prefer to have it a must, but others may
disagree. How many people are using call-by-reference by now?
Regards
R�diger
--------------------------------------------------------------------
| Ruediger zu Dohna GINIT GmbH 0721-96681-63 [EMAIL PROTECTED] |
| PGP-Fingerprint: F1 BF 9D 95 57 26 48 42 FE F8 E8 02 41 1A EE 3E |
--------------------------------------------------------------------
