On 12 January 2013 19:39, David Barbour <dmbarbour_at_gmail.com> wrote:
> I would ask a question in return: who should be a programmer? Are user
> interfaces not, in many formal senses, live programming in a problem/domain
> specific languages?
Yes and its not so unusual for user interfaces to be Turing complete
by accident.
I got talked out of considering Turing completeness as necessarily in
live programming after bringing it up on this list. The strong family
resemblances between use of languages on either side of this
theoretical divide makes it unimportant here.
> My interest in live programming has never been about performance (which
> seems the primary interest of this list). Rather, to me it is about
> understanding PL as a form of HCI. There is undeniably a gap in practice
> between PL and HCI, but I think it is not essential. It's more like a
> canyon... a consequence of where we chose to settle.
The Psychology of Programming is a pleasant valley between them, and
generally a good place to shelter from the aggression on either side.
:)
> There are regions in the design space where the distinction between PL and
> HCI is weak, I.e. where intuitions, patterns, principles, extension and
> composition learned in one are directly applicable to the other.
Yes I think there is huge value in breaking down this distinction,
especially when we look at how people work with command lines,
spreadsheets, web search..
> The ultimate audience for live programming should be everyone.
I agree but there are two things being conflated:
[programming] -> [music] -> [audience]
[programming] -> [audience]
In the first case we might argue say that for the audience, the
programming isn't important, the music is. But the problem is that
the programming is the music. This isn't a problem if you realise that
programming and music aren't things or end-products, but activities.
So the audience enjoys the activity of programming as music, and
vice-versa, and they don't have to understand the code to do it.
In the second case the audience is supposed to understand and enjoy
the code itself. This is often a design consideration of live
programming languages, for example I seem to remember Ge Wang saying
that this was a main reason why he made ChucK. The obvious counter is
that you don't have to be able to play the guitar to enjoy listening
to someone play one, and aren't upset if you see someone playing a
guitar faster than you can follow what they're doing.
Of course an audience is by no means necessary, but unfortunately it's
not possible to watch some one do live programming who doesn't one!
This also presupposes that the programmer themselves knows what
they're doing -- for me the point of live programming is that you
don't know what you're doing, otherwise it wouldn't be a live
interaction. Live programming should be one long surprise.
alex
Received on Sun Jan 13 2013 - 00:27:42 GMT