Uli said
>> I'd also be happy to discourage anyone who'd like to speak up
>> for using PERL as a language to be used for
>> implementing the HyperCard clone.
On the HyperCard list when discussing QTML, many people noted that what
we want is not HyperCard, but hypertalk, the ability to speak English
about objects.
Of course the cards model is enticing, but equally it is restrictive. In
the web age, i wonder if the web browser is not a better metaphor than
the rolodex?
If that view is accepted, in many ways opencard 2000 involves making
hypertalk into an extensible browser control language. Two quick ways to
achieve this, would be to:
a. build an opentalk parser which spits out javascript.
b. write the opentalk device in Java
Java on Mac is already connected to OSA.
Another way to be considered might be to model opentalk on php3
Tangentially, have you guys seen php3? This is very applescript-like,
right down to a highlighting parser and no variable declarations.
<http://www.htmlwizard.net/interview.html>
<http://www.php.net/>
Points in favour:
1. Nobody needs to install a runtime for opentalk: Just fire up your
browser and point it at the php3-like opentalk server.
2. This server can be local, of course, blurring the client-server model
beyond relevance.
3. Decoupled server more readily ported to other open platforms (Linux).
4. Hundreds of millions of $ are being spent building our client for us
(by AOL and Micro$oft). We are assured of having an extremely high
quality layout engine maintained indefinitely.
PS regarding using Perl to develop HyperCard, I say that OpenCard should
be written in C, just like Perl ;-).
