Re: [glob2-devel] Beta 4 Continued Plan

[Bradley Arsenault]

First of all, there is no safe way to store the password on the client
end. I thought about it long and hard and there isn't. If a person
gets on the clients computer, his glob2 password is compromised. We
don't need to worry about this much at all though. And the server
already hashes the password in the file, that was done at day 1. The
only problem now is that we need to use public key encryption to
transmit the key from client to server.

Secondly, scores aren't going to matter that much. Its just a game,
and people aren't going to be comparing scores over time. What matters
more is how a player compares to another player, especially if we want
to do auto-matching.

Decoupling should not become a problem, the number of games needed to
do it would be huge, and if people do, their scores won't matter until
they play someone else anyhow, where their score will quickly be
proven wrong.

Switching alliances should not be allowed.

I very highly reccomend we stick to a simple and structured calculas.
The chance a player wins is his score + his allies scores against the
sum of his enemies scores, we do the expected value from there,
adjust, like given in the wiki page.  The more complex we make the
score formulas, the more room for error, however this
bottom-simplicity approach is based entirely on simple statistical
analysis and should be more than good enough.

Player scores should be given streamlined in the game or won't matter
to players at all. We can't have it in a website, very few people will
look at them, not to mention that it will seem irrelevent. If we want
a system that matches a player based on his score,it will have to be
in the game, which is a better place anyhow.

I'm working on a Map Upload/Download system streamlined into the game,
including ratings, filters, and perhaps catagories. I haven't yet
decided what categories there will be.

On 4/23/08, Leo Wandersleb <address@hidden> wrote:
> hi
>
> > After that, I will continue with some improvements given on our wishlist.
>
> this list is miles long. can you pick what you see is realistic and
> important
> enough to do first? with you beeing offline so much i can't get enough of a
> plan.
>
> Leo
>
>
> _______________________________________________
> glob2-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/glob2-devel
>


-- 
Extra cheese comes at a cost. Bradley Arsenault.