On Mon, 2006-08-14 at 12:24 +0100, alex wrote:
> On Mon, 2006-08-14 at 10:22 +0100, dave wrote:
> > A lot of large scale software development is done wholly in UML, which
> > is then converted to C++ and built.
>
> Then I stand corrected... My understanding was that a lot of UML
> modelling was done in English, which was then manually converted to
> code. I imagine that this stuff is highly specific to the domain of
> business logic though.
I think that it's initially used as a planning tool, but the idea is
that it generates code, and in some instances can be compiled directly.
It definately comes from the business end of the IT industry, with
prices to match, but there are free versions availible:
http://uml.sourceforge.net/index.php
> > > However there is something special about languages rich enough to allow
> > > such things as introspection, self-description and so on.
> >
> > There is no reason that this can't be done with a diagramatic language
> > as far as I am aware.
>
> Oh I agree. I'm just trying to address the problem of defining what
> livecoding is. If we can say that it's possible to livecode with
> 'visual' interfaces (although text is a visual medium, or at least
> piggybacks on one), we need to work out what it is about languages that
> makes them good for (certain kinds of) livecoding. I think it would be
> useful to be able to say what livecoding is and isn't. Thinking about
> diagrammatic languages might actually help us do that.
>
> > For example, I've been playing around with a
> > script that converts lispy languages into dot graphs, it's not working
> > properly yet, but here is the graph representation of the program
> > itself:
> >
> > http://www.pawfal.org/dave/images/dotty.png
>
> Very nice! There is still some text there, it might be fun to replace
> each token with a unicode character. Otherwise I suppose the difference
> is that relationships between tokens are shown explicitly with lines,
> rather than just their position on the page.
Or the brackets, and this is the point I feel - that for some people
diagrams are more useful than text for programming. For some tasks
(designing large programs) text is a hindrance. Really programming has
nothing to do with text - it's just the culture which it has grown up
with.
I think the use of diagrammatic programming languages are interesting
for livecoding as it can make programs easier to read at a glance,
without as much need to understand the syntax - which might make for
more audience friendly performances.
StarLogo is an interesting language in this context:
http://education.mit.edu/starlogo-tng/
If I had more time I'd write a starlogo style interface for fluxus...
cheers,
dave
Received on Tue Aug 15 2006 - 09:57:32 BST
