On Fri, 10 Sep 2004 11:20:28 +0100, alex wrote
> On Fri, 2004-09-10 at 10:22, Danny Ayers wrote:
> > Not as it stands, nope - Java probably isn't a good language for that
> > kind of change, though changing the value of the feedback constant
> > would have a marked (and fairly unpredictable) effect.
> > What language(s) have you been using?
>
> I've been mostly using Perl, live coding is not a huge problem as
> it's an interpreted language. I interpret code changes into a
> separate area, and if that parses ok, I interpret it again, over the
> running code. The new code continues running with all variables intact.
I think interpreted languages are the natural choice for live programming.
Despite all the jokes on the subject, compiling stuff live would not be fun! :)
Something I'd like to research is changing code on the machine level live.
There may even be precedents for this. I'm thinking of a system where you have
a virtual machine with a very simple instruction set (as in evolving machine
code systems like tierra). You'd be able to grab processes, and move the
program counter around and hack their instructions *while they are running* -
no write/interpret cycle at all.
For the ultra livecoding geek - it may be possible to get your head round a
real system without a virtual machine in simpler 8bit computers, or
microprocessors that allow code to be modified.
dave
Received on Fri Sep 10 2004 - 11:43:44 BST