[Rockhounds] Fw: SOFTWARE FOR INDENTIFYING MINERALS
Dennis Buffenmyer
buff1 at ptd.net
Thu Dec 13 07:35:04 PST 2007
T
>OK, NOW I'm going to climb far out on a limb:
>
>Now think about taking Mindat to the next level. That is; adding an expert
>system. Why Mindat? Well, to me it's a great mineral database that is open
>to the expert and amateur world input. Mindat brought mineral ID peer review
>to the common rockhound and expert alike.
>
>Perhaps, initially, we could add a decision tree section into every mineral
>family with subsections under every mineral. In there we can, individually
>and collectively, build the decision tree so necessary for an expert system.
>In these sections, fresh from identifying a new sample mineral, people can
>enter the decision pathway they followed. Also, there might be base decision
>trees already in the Mindat message boards.
>
>Just my mouth and mind running on, but there it is.
>
>Ted Kowalski
>Fredericksburg, VA, USA
>
>
>
>
This is obviously just thinking through typing, but, I agree that mindat
might be the best starting place to build onto. Perhaps the best thing
would be to build into it a matter of probability. I.E. you know you
have a specimen that is from
locale xyz and it has a hexagonal crystal structure. Perhaps entering
these two criteria alone would be enough to search the database to put a
probability that the mineral is such; or with additional input criteria
that would increase the probablity of being such. I suspect though, that
what perhaps the concern is, the hope of finding a really rare mineral
or new mineral. I.E. lets say that you have a hexagonal crystal from xyz
local but xyz local is not known to have any hexagonal cystal system
minerals, then, additional criteria would have to be entered to help
increase the probability identification of said mineral.
I like the idea of starting with mindat in that pictures are worth a
thousand words, BUT, can be misleading. The case in point of course
would have to be Mont St. Hilaire, where from a picture, it might be
easy to mis-identify a gmelinite for a mal-formed catapleite or visa
versa, in which the identification of the matrix rock would also be in
question. I need to make sure that no one takes me for an experts
opinion, as the experts have the tools at their disposal to make the
determinative ID, but rather as a hobbyist who is among the group
looking for such a tool that would be useful in our quest for greater
knowledge.
In thinking of this, it might be best when a question arises about the
ID of a mineral when attempting to make your own ID, that on the label,
( here again, how many times has it been emphasised here, how important
is it for the archiver ) that the specimen was purchased as such BUT
you have reason to believe it "MIGHT" be such, based on the criteria
that you have documented with the specimen. Or that you found such a
specimen and believe it to be, abc based on the criteria you have
listed. could only add to the provenance and future value of the specimen.
Certainly it would be hoped that those who are classifying themselves as
"professional" mineral dealers or mineral specialists are taking the
measures needed to positively ID the specimens before charging exuberant
prices for a guess.
Perhaps this whole thing might best be summarized to leave it as... the
specimen is labled as "abc" because we "believe" it to be "abc" based on
xyz criteria......
humble offerings....
Dennis Buffenmyer
More information about the Rockhounds
mailing list