Hi!

I read some more on the url you gave me (http://catb.org/~esr/trove/)
especially under the "browse HTML" link. The FAQ seemed broken to me -
it was some kind of todo list. But the design doc is interesting,
nevertheless it seems very old (1998?). Ayway, I also ended up in some
PEPs over at the Python guys for example:

http://www.python.org/peps/pep-0241.html - This one describes what
should be in a package registration (fields). I think we have most of it
covered in SM (didn't look in detail)
http://www.python.org/peps/pep-0301.html - This one describes a network
protocol for doing what SM does. Somewhat interesting to read.
http://www.python.org/peps/pep-0262.html - This one talks about having a
local "database" containing info about installed packages etc. Well, in
SM we have that.

Anyway, all of this is more or less in many ways identical to SM. I
should in detail read through the PEPs in this area.

Btw, some serious (IMHO) shortcomings of Trove at first glance (I may be
wrong):

1. Doesn't support "package releases" (neither does SM1.0x of course).
2. The Trove "discriminators" (=categories) are simply Strings. They
can't be renamed, nor do they seem to have any one-line description nor
associated url. And they can't be made mandatory as in SM I think.

Anyway, we should definitely learn from all these other projects (and I
have extensively been trying to do that) while still building something
simple and Squeaky. I think I have done that too. And as always - ideas
regarding SM are much welcome.

regards, Göran