Re: JPush: Lush on Python on Java

> If I am reading you correctly then you are favoring staying with the
> "standard" Lush side of things. On that side, I agree with your points.
> My idea though is to have two different stacks, so to speak, of Lush (there
> could be more). Each one takes its own path yet at the same time they have
> clear connections to each other to take advantage of different
> functionality or other specifics.

Sure, the more the merrier. Most "standard" languages have many
independent implementations, why not Lush.

> From what you say Lush is primary just a niche application as it
> stands, but you have also mentioned in the Lush documentation that
> some people use it for very general tasks, like system
> administration and so forth. Part of my own use of Lush involves
> very general applications.

I use Lush for pretty much everything too (not just my research work),
and so do many of my students and several of my friends and
collaborators. I use it for such things as text processing scripts,
grading scripts for my class, database engine and mailing list
maintenance for the conferences I organize, pedestrian image
processing tasks (like grabbing hand-drawn figures from a webcam and
converting them to EPS so I can insert them in slides for my
lectures), pseudo "make" scripts for various tasks, HTML pre-processor
for websites...... I also use it for things most other people would
use a spreadsheet for (like computing NSF grant budgets).

That said, the real added value of Lush, compared to say Python, Perl,
Common Lisp, Scheme and OCaml, are the ones I mentioned in my previous
message (the same ones Ralph pointed out).

> I don't know how much you know about Jython, but it actually makes it quite
> easy to implement domain specific interpreters, a Lush implementation could
> be one.

I don't know much about Jython.

> What is needed here is a proof of concept. I'll have to come back to this
> topic when I have something that at least does a rudimentary REPL.
>
> As an aside, in my view open source software, perhaps along with ubiquitous
> networking, parallels biological processes. For instance, imagine any
> particular application as a type of plant. Some plants grow only in
> specific areas under a limited set of environmental conditions that must be
> met. Others spread far and wide and can tolerate a wider range of
> conditions.
>
> What is the main cause of the difference between these types of plants,
> some very adaptable and others not so?
>
> The difference seems to be that the adaptable one is able to take root
> under conditions that it did not originate from. In terms of evolution, the
> more adaptable plant has a higher probability of long term success, having
> at least some semblance of the original plant existing well into the
> unknown future. There are definitely parallels in computer applications
> that we all know of: Linux, Perl, etc.
>
> I think Lush should be included with those, the more adaptable
> applications, therefore having more chance of staying around long into the
> future.

I'm certainly not against such an effort (I'm all for it in fact).

But again, since people's time and energy is limited, it might be a
good idea to think about fixing some of the weaknesses that Ralph is
pointing out, (visualization, fewer more powerful functions, better
compiler...), and strengthen what make Lush unique (numerical stuff,
easy C/C++ interface).

  -- Yann



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php