At 8:42 PM -0800 on 12/3/99, Alain Farmer wrote:
>Anthony: I don't have a problem bypassing the message
>hierarchy; that can be done easily.
>
>Alain: Good. What do you propose?
>
>Anthony: I do have a problem with building commands at
>runtime.
>
>Alain: This and the above are two different issues.
>SEND does not necessarily involve "building cmds at
>runtime" We thus keep SEND, right?
Send builds commands at runtime by its very nature... However, I can work
around this. A send with a single, quoted argument will be fine.
>Alain: There was a heated discussion on the pros and
>cons of HyperCard's DO command on the HC-List. Most
>contributors to this discussion were insisting that
>the DO command is one of HyperCard's hottest features.
>I am among them.
I missed the discussion, and I hardly ever use the 'do' command [though it
did come in quite handy in StackSearcher, but only as an optimization
because HC is slow].
Is there an archive of the HyperCard list? I'd appreciate a URL to the thread.
>
>Anthony: Most that use send only use it to send a
>mouseup to a button; quick fix will work there.
>
>Alain: I also use SEND a lot. One cannot mix several
>scripting languages inside the same script, so I put a
>lot of my AppleScript and MacPerl scripts inside
>hidden buttons that are remotely-controlled by the
>stack script. I also use SEND to extend HC's
>stacksInUse to more than 16 stacks. And I am sure I
>could think of many other good reasons for DO and SEND
>if it becomes necessary.
No stacksinuse limit, sending of simple messages is fine.
>
>Anthony: As for runtime variable naming, hashes can be
>used instead. A fairly quick fix.
>
>Alain: A fairly quick fix for runtime variable naming
>that doesn't affect the Interpreter. Sounds great.
>Let's do the same (or a similar trick) to keep DO and
>SEND in NuTalk/OpenTalk.
Not possible.
