Friday, September 9, 2005 - Fri, 9 Sep 2005

Just about finished the 3 things I wanted to do on Tuesday. My new goal is to get this version out on Monday. I’ll see what I can do.

In the meantime, I had two very interesting questions come my way by e-mail. I just can’t help myself from printing these and my responses in this blog.

Q1: My family wants to make photos (thousands) available on the web and enable people to access the photos, download full photo files (typically 0.3 MB each), and write comments tied to each photo. Is there a service that will provide these capabilities and the storage (gigabytes’ worth) necessary? Family tree capability is not important. Just photos, in all sorts of digital formats.

Yes, have considered:

- familysearch.org, but gigabytes’ worth of files doesn’t seem feasible, given the organization’s thrust.

- mygreatbigfamily.com, AND THIS MAY BE THE BEST AVAILABLE, but gigabytes’ worth of files would be prohibitively expensive, and photo files aren’t downloadable, and comments on photos don’t appear to be searchable.

Hoping you can share some of your wisdom . . .

A1: If you want to share lots of photos with other people, I would recommend you
get a web host (i.e. your own website) from a service that has a good photo
gallery system included as part of its service.

One image gallery system I have seen and think is excellent and will probably
do everything you want is a free one for Unix called 4images. Their website
is: http://www.4homepages.de/ and they have an online demo at:
http://demo.4homepages.de/?l=english

There is a Fantastico package that many good webhosts provide, and it includes 4images as well as many other great tools. Some of the best sites that support this package can be found at:
http://www.hosting-comparison.com/info/fantastico-hosting. These sites will
typically give you 3 GB or more of webspace for $7 to $8 a month.

For what it’s worth, that’s what I would do.

The second one relates back to my past life of Computer Chess programming, which has been laid to rest in favor of Genealogy programming. But my interest is still there when something related comes up, as was this question:

Q2: Do you think that it is possible for a perfect gambit to be determined for chess, that is an unavoidable draw, or win for white or black? Matiyasevich demonstrated its impossibility in his book. What is your opinion and why?

A2: What an interesting question!

It is my understanding that Matiyasevich’s theorem can be stated as: “Every
recursively enumerable set is unsolveable”.

A “recursively enumerable” set would be a set S of integers that has an
algorithm that behaves as follows: When given as input an integer n, if n is a
member of S, then the algorithm eventually halts; otherwise it runs forever.

In chess terms, S might be all positions for which an evaluation can be made
(draw, white win or black win). For this to be a recursively enumerable set,
there would need to be some positions for which the evaluation algorithm would
run forever. Since chess has rules that limit the maximum length of a game
(because of the 50-move limit, and forced draw positions), a minimax algorithm
to evaluate any position might take a very very long time, but it would not
run forever. So chess is NOT a recursively enumerable set, and Matiyasevich’s
theorem does not apply.

A case in point is that many other opponent vs opponent games can be solved:
Tic Tac Toe is easily solvable. The game of Othello has been solved. Checkers
is on the verge of being solved - see: http://www.cs.ualberta.ca/~chinook/.
Chess is a couple of orders of magnitude more complex than Checkers, but it is
still finite, and therefore is technically solveable.

For more of my thoughts on this, read Section 6: Unfinished Work and
Challenges to Chess Programmers on my Brute Force page at
http://www.lkessler.com/brutefor.shtml#bf6. In there I give many of my
thoughts about how to solve the game of chess. I even stick my neck out and
say “Maybe, 100 years from now the game might be solved.”