Re: Scripting


Subject: Re: Scripting
From: Mike Meyer (mwm@mired.org)
Date: Sat Jun 24 2000 - 23:31:18 CDT


Tomasz Wegrzanowski writes:
> On Sun, Jun 25, 2000 at 01:58:20AM +0200, Aaron Lehmann wrote:
> > I don't see anything wrong with pleasing the geeks too. Selfishly, I work
> > on AbiWord for fun and I don't really have a user focus. I think it's
> > important to focus on users but I don't like doing it myself in an open
> > source context. I definately don't see anything wrong with user focus.
> What about first making list of things that script-language would be doing,
> (!= what VB does in MSOffice)
> and then deciding which one to choose, instead of choosing
> language on base of personal feelings.

Because choosing a language is the wrong choice :-). You need to
choose a technology first. The contenders are the embedded scripting
languages (that list is very long), DCOM (which may not exist in an
implementation we can use) and CORBA (multiple usable implementations
exist).

Any of those provide the scripting. What you can do with a script then
depends on the implementation, independent of the technology. Nuts,
adding CORBA means adding CORBA wrappers to the program objects that
export the methods you want to make available along with a mechanism
for invoking user-specified external programs ("scripts") and linking
against a CORBA library. Adding Python means adding Python wrappers to
the program objects that export the methods you want to make available
along with a mechanism for invoking user-specified Python strings or
files and linking against the Python library. If the objects and
methods are the same in both cases, you could create two systems that
had scripts that differed only in an initial import statement.

If you really want to work on a list (I think it's a bit early), then
be sure and read <URL:
http://www.mired.org/home/mwm/rexx-interface-design/Contents.html
>. But in my mind, that list can be simply described as "anything the
user can do sitting at the keyboard". Otherwise, you've failed the
"completeness" criteria of that paper.

        <mike



This archive was generated by hypermail 2b25 : Sat Jun 24 2000 - 23:31:20 CDT