discussion.
(1) Further detail work in my code for the Elephant semantic-net
implementation. These are nitty-gritty details that I will presumably
either find in the documentation or discuss with Elephant developers,
however, it may be helpful to state the issue first for you (and if
you want to follow up you can). The main point is that I want to key
Things by Object Identifiers (OID). And yet at the same time, I want
to be able to scroll over all Things in a given Theory. Graphically,
the implementation I have in mind looks like this:
Things --
[key: OID | slots: name, data]
a Theory (is also a Thing) --
a b-tree containing elements [key: name | slot: OID of some Thing]
a Triple (is also a Thing) --
[slots: beginning, middle, end]
The current implementation (produced by Ian Eslick in response to my
in-person inquiries when I was last out east, around May Day) enables
the user to find all Triples which match on any particular data in the
three slots. This is how the bi-directed network "works".
The current difficulty comes in now that I want to be able to "import"
Things (including Triples) into a Theory. I do not simply want the
Things to *reside* in a given Theory, or I could just tack on an
additional slot naming the home theory. Rather, I want Things to live
in a Universe and Theories only to designate arbitrary subsets of that
Universe.
But at the same time, I want the data-matching routines to work within
a given Theory, so that a Theory can also be viewed as a network. I
am pretty sure this can be done using the UPCOMING tools in Elephant
(a "query system" that would enable me to focus on all of the Things
that have an OID named in a given Theory, and subsequently tool on
this subset with the data-matching code).
Epilogue 1: Now, these are, like I said, nitty-gritty details. It may
be advantageous for you to think instead about more theoretical (or, I
suppose it is "meta-theoretical" aspects of this model). Are these
types of systems "enough", assuming they can be made to work? Etc.
(2) In response to a possibility for rewriting and submitting a
variant of the Metacommons Manifesto that Aaron mentioned to me, I
prepared these notes:
http://planetmath.org/~jcorneli/metacommons-survey.pdf
I'm not sure that the possibility he mentioned is still viable,
however, I do think that this rewrite project is -- even if it must
become a big project. As the title of the paper indicates, I am
looking at it as a "survey article" which might be, say, 30 pages
long. The current outline (at this link) is only about 10 pages long,
and might be viewed as a rough-draft of a prospectus for this larger
paper.
Epilogue 2: I have not even printed this draft off to read it and mark
it up, although I did look over it yesterday, and at that point in
time it still looked pretty good. However, it lacks many important
details. My guess is that finishing it might require additional
person-hours equal to twice the final page-length. Does anyone have
any statistical data that could back this claim up (or challenge it)?
In any event, if anyone is going to work on this, it will need time
and a decent work schedule and realistic publication goals and
information-gathering strategies (none of which have we had yet for
this particular paper).
(3) My notes from Mathfest might also be worth turning into a paper.
In addition to the ~15 minute recording of the actual talk (which I
already indicated and which I know Aaron has heard, but once again it
is
http://planetmath.org/~jcorneli/joe-mathfest2007.wma
Please let me know if you need another format, Ray), I have about 40
notecards with my thoughts, pseudo-definitional snippets, and
recordings of events from Mathfest written down on them. I imagine
that these could be used as the raw ingredients for another paper,
this time about PlanetMath and HDM (instead of about the Metacommons).
Epilogue 3: I could relatively easily make photocopies of these
notecards into a "zine" for you guys. Do you want me to do that? Do
you like the idea of writing a paper like this to submit to some
popular math journal? Can we get the approval of any editor in
advance (e.g. since I was just at an MAA conference, the Math Monthly
or Journal of Online Mathematics)? IMO, preparing such a document
ties right in to item 4...
(4) Promoting PlanetMath by following up with people I was introduced
to at MathFest.
In addition to several potential sponsors -- I should send you the
list at least of companies; I have email addresses and could at least
follow up with these, although I'd expect as a prerequisite to have
some COPY to work from, and I think this is something you guys have
been talking about preparing! -- it looks to me that MAA itself may
likely to want to enter into some relationship with PlanetMath
assuming we can hash out the RIGHT relationship. Aaron said something
along the lines of: working with PlanetMath might really "revitalize"
MAA for this century.
But there are some serious problems to be worked out with that, for
sure. I think given the state that PlanetMath is in currently, it too
needs some revitalization. Perhaps this is a real case of a possible
"synergy". Just for example: many of the "governance" things we have
talked about have not been resolved (from licensing concerns, which
was what I first brought to the table here, to the latest editorial
and etiquette concerns embodied in the content committee proposals);
and of course, as another example, none of our "capacity" is supported
(yet) in an ongoing or stable way, and this includes the whole
technology side of things. This isn't just a matter of a single
stalemate; I see it more as a whole quagmire of "foggy" disagreements
found within a long and still growing series of discussions. I hope
that the particular installments in this series that I am proposing do
not further muddy the waters.
Epilogue 4: I have specific people in MAA I want to talk to about
these matters -- again, I think it would be good to have some copy
fleshing out the points I have brought up here (and other relevant
points), even if just (or mainly) to serve as a starting point for
conversations.
Meta-epilogue: I think I have started four conversations going in one
thread. That is probably OK for the time being since the things
inherit from each other in various ways (sometimes subtly). I also
think it would be good to discuss them in (at least) one phone
conversation, or further by email. Can we schedule a chat for
Wednesday or Thursday of this week?
0 comments:
Post a Comment