Re: [glob2-devel] scripting language

[Bo Lorentsen]

Andrew Sayers wrote:

> First, I rather liked Stephane's list of requirements idea, so here
> goes...
>
> * Security.  It should be possible to put the program in quite a small
>  sandbox while it's running.  It would be bad for the program to, say,
>  upload /etc/passwd to a website (which would be pretty easy in Perl)
>  
I agree, we don't want to create a new worm/vira platform like outlook :-)

> * Size.  I think everyone's already said this, but it's worth
>  remembering that Perl is Big.
>  
Size as in language complexity, or as in binary footprint ?

> * Documentation.  To be useable, a programming language needs lots of
>  it, with examples.
>  
I am aware of that and are working on it btw. Language documentation are 
an iterative process, I just can't sit down writing a manual over a week 
or two, I don't have the ability :-)

> * Sustainability.  With no offence meant to Bo, what happens if you fall
>  under a bus?  Is there anyone else that could maintain the code?
>  
God point, but I don't plan to :-) The project is free, and I work at 
making TEA as easy to use and read as possible, but bad things could 
happened.

But, we all have to start sometimes, and if we don't do things 
different, how can we be better ?

> * License.  TEA currently has no license information on it.  Does that
>  make it public domain?  I'm not sure.  Any language will need to be
>  GPL-compatible anyway.
>  
Sorry about that. I plan for either LGPL or BSD, or whatever license 
that don't take up 40 lines of every source file ... I just hate that. 
The top of the source file is for nice comments about the functionality 
! It is an absolutely free project, and will remain that way, as I don't 
see the point in doing otherwise for a core language, like TEA. I just 
like any code changes to stay in the project, but otherwise I will not 
put restrictions on its usage (I use it professionally in a non GPL 
project, myself)

Let me know if You have some insight into this license subject, as I am 
not very knowledgeable in this regard. I too have a tendency to ignore 
things that is not LGPL or BSD (or apache), but I have not studied the 
subject closely and have no plan doing so in the future.

> * A programmer.  Ultimately, the only thing that really counts is that
>  there's someone offering to do the job with TEA.
>  
Well, I was hopping to be able to do this myself, to make the things I 
learn, doing the processes, mature the TEA language.

> First, I've played with LISP before, and I'm not a gigantic fan.  But
> it is much easier to do GA-type stuff in LISP than in C.  Basically,
> LISP programs can be represented as trees, which makes it significantly
> easier to genetically modify them - say, by pruning a branch off one
> tree and grafting it onto another.  Lua is a LISP-like language, which
> makes me happy but isn't an argument on its own for using it.
>
>  
Ok, I am not that strong in the GA part, but how about Lua's type 
checking system ? As far as I remember most of its checking is done at 
runtime, and not at compile time (the nature of the language). Tea is 
strong typed (like C++ and ML), and the goal is to be able to check 
everything at compile time (except NULL values, sorry).

> However, I would
> prefer to see a general-purpose language in future versions, that can be
> easily expanded with hooks into the program.
>  
I agree, the amount of things that would be possible using scrips will 
give nice future possibilities.

Now, back to VM coding :-)

/BL