He does have a point though, it is possible to reduce
instruction sets down to a small subset but those commands
(verbs) tend to end up having a gadzillion options and parameters
on them to cope with all the different semantic situations
they have to address.
If you assumed that programming was all about doing
incredibly high level things ( move that data from here
to here ) then you might end up with the VIC. But it is
a naive model - not least because the amount of underlying
abstraction layer coding to achieve it would rely on so
many assumptions it would be near impossible to apply
in all circumstances.
The biggest problem with programming languages is
achieving one that is easy to use, simple to understand
and difficult to phuck up on. Seems to me VIC does not
address any of those.
Meta-programming languages ( those that describe how a program
should work and code-gen the implementation ) have been
around donkeys years but whenever you point this out to
Tim he goes around accusing them of plaigarism ( IBM, Sun,
OMG ), piracy, breach of IP and the person who points it out
of trolling.
But seeings it took nearly a decade to get this far I doubt
that the next upgrade will come any time soon.
Tim, I *think* has good ideas but he is totally unable to
articulate them or even flesh them out.
