Re: CMake and Lua

["Brandon Van Every" <bvanevery@...>]

On Thu, Feb 14, 2008 at 4:53 PM, Eric Tetz <erictetz@gmail.com> wrote:
> On Thu, Feb 14, 2008 at 1:16 PM, Ben <thebassplayer@gmail.com> wrote:
>  > Yeah, but didn't most languages like Lua start out as "roll your
>  > own" languages?
>
>  Sure, but the Lua guys are writing a *language*.  The CMake guys are
>  writing a *build tool*, of which a language is a component.
>
>  A huge amount of time and effort has gone into making Lua the
>  well-oiled-machine it is today. It's had 15 years to work out it's
>  kinks. Is it reasonable to expect <tool builder X> to invest the same
>  level of resources into a subcomponent of their tool?

Why would they need to?  They don't have to solve general purpose
programming problems, only their tool's problems.  CMake has been
around for 7 years, spent most of its energy on build system rather
than language issues, and still has a passable if not exemplary
language.

>  More to the point: is that a constructive use of their time? The whole
>  existence of Lua is predicated on the notion that it's not.

Well, if Lua had better advertizing in 2001, maybe they would have
used it.  But even today, advertizing in the Lua community is
substandard.  Marketing is part of the burden of getting people to use
things.  I don't subscribe to the "build it and they will come" school
of thought.  Microsoft has clearly taught us that marketing matters.

Moving to Lua is a techie way to deal with syntax issues.  Writing
decent docs on the syntax corner cases, and making it clear to people
why command(arg KEYWORD arg arg) is actually beneficial, is a
marketing way to deal with syntax issues.  That's why I keep searching
for "the better programming paradigm."  I know that documentation and
marketing can deal with the syntax squabbling.  Either there's a
better programming paradigm for build systems, or there isn't.


Cheers,
Brandon Van Every