On Monday, February 9, 2004, at 01:45 AM, [EMAIL PROTECTED] wrote:
Of course with Transcript, the = operator means the same thing that the ==
operator means in other languages. That probably can't be changed anymore or
everything wouldn't be backwards compatible. So some other kind of
assignment operator would be needed like
A parser works by looking for "tokens", such as "put" or "set" or "if". When the parser sees the "if" token, it could decide that the "=" between the "if" and the "then" is a comparison operator (as opposed to an assignment operator). Similarly, when the parser sees "x = 1 + 2" at the beginning of a line, it could recognize this = as meaning assignment (as opposed to comparison).
In other words, there's no reason these two couldn't be defined to be identical in xTalk:
if a = b then put 1 into x end if
if a = b then x = 1 end if
(this latter is the same in VB, for example)
> Some languages overload = to mean 'defined as' in some contexts and >'equals' in others. Mathematics is in the same situation and some >folks use a different symbol to mean 'defined as'. > >Dar Scott
See the note above -- the context can be used to tell the parser whether = means assignment or comparison.
>Yes, but *only* if it allows both syntaxes. Over the years that I've > watched languages evolve, they too often leave a syntax behind as they > evolve to a new one. Dot notation isn't hard to learn, but it, again,
The first versions of Lingo (Director), which was quite similar to HyperTalk, supported both "put" and assignment = syntax, so you could write either "put 1 into x" or "x = 1". There's no reason xTalk couldn't do the same.
> x = y + z (instead of put y + z into x)
> x += 1 (instead of add 1 to x)
> x.myProperty = foo (instead of set myProperty of x to foo)"
>
> Sorry, Frank, but I just flat disagree. Those syntaxes -- in particular
> the far-too-cryptic and unreadably annoying "x += 1" -- are off-putting
> to all but professional programmers with backgrounds in C/C++/Java. And
Disagreeing right back at you. If you want professionals to use Rev, then you need standard assignment statement syntax -- without them Revolution looks like a hobbyist language rather than a real working language ("oh, it's just HyperTalk, and we all know that wasn't a 'real' language/development environment") And since it's perfectly possible to support "x = 1" without affecting the current "put" and "set" statements, I would argue that they should consider adding it.
[changing topics]
>Removing exceptions can simplify xTalk and enhance its power.
You mean try/catch/end try? It might simplify things, but it sure won't enhance anybody's power. There are numerous places that common functions can fail in xTalk (e.g. set the fileName of an image to an alias file -- oops) Since xTalk has no consistent failure reporting mechanism, exceptions are really the only reasonable way to handle exceptional conditions without having tons of if statements littered throughout your code. (What, you mean you don't handle error conditions? Shame on you :-)
>For example, if 'f()' is a built-in function then we can apply it as >'f()' or 'the f'. By why limit this to built-in? Why not allow this > for custom functions, too?
I'd second this idea as well.
-- Frank
_______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
