At 2:07 PM -0800 on 11/28/99, Alain Farmer wrote:
>> Alain: This new OpenTalk syntax was inspired from
>> AppleScript's favorites feature, but my spin has
>> nothing to do with MacOS favorites or with Web
>> bookmarks.
>
>Anthony: The problem is that your syntax will only
>allow paths ...
>
>Alain: What makes you say this ??
Because you're defining a property which stores search paths, correct?
>Alain: How much limited script space (chars) are going
>to be occupied by this search-feature?
First off, THERE IS NO 30K LIMIT. There is not a problem with limited
script space, period.
Second, these script should be but a few lines, I'd guess. The simplest one is:
function NuCard:FindStack name
for each line l in the stack search paths
if there is a file (l & name) then if there is a stack (l &
name) then return l & name
end for
answer stack "Please find "" & name & ".""
if it�empty then
put it&return before the stack search paths
return it
else return empty -- could not locate
end NuCard:FindStack
(This would, btw, be the built-in algorithm. And there I go informally
proposing syntax again.)
>How likely is
>it that users will toy around with such a complex
>feature?
Complex?! It's pretty simple, really. See above.
>How reliable will their changes be?
Very, I'd imagine.
>What if
>users inadvertently changes (a search and replace all,
>for example) or deletes these critical handlers?
If the messages reach NuCard, they'd be handled internally. So it won't be
catastrophic if you delete them.
>Alain: Could we objectivize handlers? In other words,
>treat individual handlers inside a script as an
>instance of the handler-class. Each handler could thus
>have properties: like dontSearch, dontChange,
>dontRemove ... for examples.
This would be a major change from HyperCard, and would be better discussed
on a different thread (translation: let me think about it)
