begin  quoting David Brown as of Sat, Jan 12, 2008 at 02:04:46PM -0800:
> On Sat, Jan 12, 2008 at 01:07:23PM -0800, SJS wrote:
> 
> >Equivalent features would be nice as well. Something that takes
> >two strings from the command line and prints out the result in an
> >easy-for-a-human-to-verify way would be very desirable.
> >
> >Plus, we could all then time it the same way, using the same "time"
> >program, giving comparable results.
> 
> The results I gave were equivalent to that. 

I want to use the same tool to approximately measure time, so that we're
actually measuring the same things in the same way. 

>                                              In lisp, you generally use the
> lisp interpreter as your shell and give your expressions there to evaluate.

That's not very user-friendly.  I don't expect Java programs to be run
only from bsh, or sh scripts to require that I change my login shell just
to run a program.

> I haven't even figured out yet how to make a standalone program (I did a
> hello world), or parse arguments and such.

Important things for a language to be able to do, I should think.

> Lisp's time uses the same system call to get the time as time, so I would
> expect the same results.
 
I expect all sorts of things that the profiler and unit tests tell me
aren't true.  I don't have that warm fuzzy from the version of lisp I
have available to me. :)

> >A suduko solver seems like a good fit for a lispy language. :)
> 
> I've already moved on to directory parsing and tree walking...

Heh.

> >Also, it seems like there's going to be a potential problem when the board
> >size is > 3x3, as then you'll have items that are bigger than your power
> >increment.
> 
> Yes, that's a bug.  The 10 was just for debugging, and I forgot it.
 
Perhaps someone fluent in lisp could offer some constructive criticism.

> Although there are far bigger problems with larger boards, such as utterly
> rediculous amounts of memory needed.

Ah. Rouge-ish behavior, eh?

> For larger puzzles, a smarter algorithm is definitely needed.  There are
> actually fairly simple algorithms to always solve the puzzle, but they
> won't necessarily have the smallest number of moves.

I think they'd be harder to grok than the plain old search. What do they
do, have a series of "swap" operations that leave the other tiles unmoved?

-- 
I have come to distrust languages that are self-contained
To the point where the OS compatibility is at best feigned.
Stewart Stremler

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Reply via email to