Hi Mr. Campbell!
On Monday 05 May 2008, Gary Campbell wrote:
> >> I seem to be in the opposite mode from Mr. Fish.
> >
> > How so, exactly?
>
> I thought that was clear. I'm still having fun with coding my solver.
> You are on to writing about what is high quality software. We've
> each been in the other's place and "done that."
Ah. Well, while I write many essays now I'm still having fun working on free
software. Most of my recent contributions have been to
http://cpan.uwinnipeg.ca/~shlomif and some of the stuff there is more
important than a Solitaire solver can ever be. (And everything there is
open-source.).
And like I said, I'm still interested in Freecell Solver, and do not rule out
that I'll revive it in the future.
>
> >> I'm still having fun
> >> writing the "last" version of my solver, and taking my time doing so, I
> >> have to add.
> >
> > That's "latest" not "last". I should note that I had fun writing my
> > solver too, but now lost interest to other projects. I may return to it
> > in the future.
>
> No. I intentionally put "last" in quotes rather than say the obvious.
>
Ack.
> > In any case, there are many more parameters for being
> > "industrial-strength" than speed as I've shown in the essay. Personally
> > as a hard-core UNIX developer, I have a hard time calling a standalone
> > (non-library) Freecell solver written in 8086 and DOS Assembly
> > "industrial strength". But maybe that's just me.[1]
>
> My belief is that "industrial strength" can come in many flavors. The
> design, coding, and robustness of the deliverable is what counts,
> regardless of the hardware, software environment, or implementation tools.
> But maybe that's just me.
You are right in a sense. However, in the Unix world, no one will take a look
at such a library (which is composed almost fully algorithms and text
processing) if it's not written in C or in any other portable language. So in
the context of DOS or possibly also x86 Windows, your solver may be adequate,
but the world does not end at DOS and Wintel. Furthermore, it's a standalone
command line application and not a library, which complicates things further.
Freecell Solver OTOH, can be compiled as a static or shared library (= DLL)
and is fully capable of instantiation.
So your solver, while an impressive piece of engineering, is not adequate for
many needs.
In any case, part of the motivation for my essay, was to show you why your
argument that Danny Jones's solver was better was in a sense false. While it
possibly outperforms my own, the fact that it is not available for download
makes it a complete no-starter as far as "shrinkwrap" software is concerned:
http://www.joelonsoftware.com/articles/FiveWorlds.html
Again, I'm sure Danny's solver is an impressive piece of engineering.
Regards,
Shlomi Fish
-----------------------------------------------------------------
Shlomi Fish
http://www.shlomifish.org/
Interview with Ben Collins-Sussman -
http://xrl.us/bjn8s
The bad thing about hardware is that it sometimes work and sometimes doesn't.
The good thing about software is that it's consistent: it always does not
work, and it always does not work in exactly the same way.
Received on Tue May 06 2008 - 01:57:55 IDT