Re: [Gnu-arch-users] is there demand for itla?

[Doran]

Tom,

your description of ITLA sounds like just the icing I'm after: the declarative
syntax describing how commands are used would make building GUIs and overarchy
functionality fun (rather than tediously replicating syntax templates), and I
know that faced with an interface with the capability to interrogate the user,
a bunch of my collaborators would stop resisting and come on over to arch.

You have obviously a good grasp of the subtle design of such a system, which
makes me think "go for it!" -- hopefully saving people a lot of wasted time as
they independently engineer unmaintainable hacks in a similar direction.

Synchronising the command set with tla is obviously a core issue, and given
the fragmented approaches to shortcutting commands and embellishing (sorry -
similifying) syntax that the list has seen, it would bebeneficial for this to
be started by someone we all trust (ie yourself).

As for Guile I'm not a scheme hacker but must agree with Miles that it poses
a much smaller entry barrier for developers who want to quickly hack something
up than do a lot of other scripting languages.

One question though, would you see the itla syntax definitions being used at
any point in the tla source? There seems to be commonality of purpose with
tla's option-parsing code, but a quick meditate in this direction just gets me
thinking of big fat grey fuzzy lines. Guess that's why I'm not an UI engineer
:).

Oh, and I think building itla will make overarch emerge much faster and in a
more coherent fashion: it seems that (most) people have been resistant to
"take the first step" in creating oarch due to the lack of a common framework
to build on. A repository of itla command definitions would be far more
attractive to browse and submit to than a repository of random shell, ruby &
python scripts each written in the contributor's own style and with little or
no code reuse ...


Anyway, that's my 2c.