[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Accumulator



--- Sundar Narasimhan <sundar@ascent.com> wrote:
> Yikes. This discussion is dangerously close to becoming a Python
> vs. Lisp thread. As a longtime Lisp hacker (we still use it in all
> our
> products -- PaulP you may want to check out our web site)
> and a Zope/python developer -- it would be a real shame were that to
> happen. 

You are right about that.  To avoid spamming this list I've added
a page adressing some of the issues people have raised with 
RotN, at http://www.paulgraham.com/icadmore.html

> 
> Personally, I find both to be very good and very useful for the kinds
> of things they are targeted at. I think the real shame is 'language
> designers' who've never built real applications (let me modify
> that.. language designers who've built ONE or TWO applications)
> extrapolate from their experience to decide what should be good for

Hmm.  Wonder if this is supposed to be me.  In case it is, let
me mention I am now 37 and have been writing apps for a little
over 20 years.  My main goal in designing Arc is to create
a language that is designed to be convenient for application
writers rather than interesting for compiler writers. See
http://www.paulgraham.com/langdes.html

As for basing a language on your own experience, whatever faults
it has, it is better than the two alternatives, which are to design
based on a priori principles (Prolog) or by committee (Ada).  All
good languages I can think of were shaped by some app the language
designers wanted to write.  Lisp, C, Smalltalk...

> the rest of us. I also sorely wish we had a better way of mapping our
> mental constructs and understanding languages than the somewhat
> adversarial -- 'how would you do code-piece X in language Y', and a
> better way to do multilingual programming than CLR -- which is
> largely
> still driven by 'compilation units' and file concepts that date from
> forty years ago. Why is it that if you find features in Python or
> Perl
> that you sorely like and wish were added to Lisp (or vice-versa) you
> you are going to be stuck like PaulG trying to invent yet another
> language.. a lot of work, a lot of risk, and likely not to get
> anywhere fast.

Are you saying it's a waste of time to design new languages?  If so
then you believe that the language design problem is now solved, and
that there are no new ideas out there to be discovered?  Hmm.  Not
so sure about that.  

In any case it is an amusing idea to express here.  This mailing 
list was established for people interested in designing new
languages.

--pg


__________________________________________________
Do You Yahoo!?
LAUNCH - Your Yahoo! Music Experience
http://launch.yahoo.com