Too bad The Joy of Sets is already taken …
]]>There's a MathWeb Org and a MathWeb Wiki where some people were developing and discussing projects along these lines, but I haven't been back there in a while.
]]>So far I've noticed only the above two links.
]]>I just noticed that propositions as types goes to Toby's edition of the article while propositions as types done right still goes to my old version.
Is this a glitch? — or just something that feathers through after a few propagation cycles?
]]>I have followed Toby's suggestion and put my excursus (foiled again) at propositions as types in combinatory algebra. I put Toby's intro at propositions as types in type theory as the software kept sending propositions as types to the original article even after I deleted the redirects. I don't know how folks here like to handle disambiguation pages. I sort of liked Toby's use of "in {your sub*subject here}" instead of parenthetical subcategorizers, so long as it works grammatically.
I left the redirect from propositions as types done right in order to preserve several links from the cafe and forum.
By the way, how do you put invisible comments in an article here?
]]>By popular request, I've changed the name to propositions as types done right.
]]>I'm thinking that propositions as types done right might be just the thing.
I started doing this ambidextrous thing back in the days when I had group theory profs with opposing prefs — later on I was encouraged to see that a small but valiant band of catgorevores like Ernie Manes made of point of applying operators on the right.
So now you know who to blame …
]]>It seems like the easiest thing to do about plurals and other spelling variants is simply for the entry initiator to anticipate the most likely variations and to include the appropriate redirects in the article. I always put mine at the top because I normally paste in from a text file and that keeps the redirects from getting lost in the ebb and flow of editing.
]]>I think you should call it n-telechy …
]]>That page crashes Firefox. Will try with IE.
Can read the page with IE. Xcept for images.
Can download the graph and scope with Inkscape.
Not too pretty though. Labels overrun their ellipses.
]]>Myself, I usually take the All Discussions default view.
But try to think about how a casual first time viewer sees the place and whether they will ever be a second time viewer.
When I do click on the General view, I get a mix of (1) content-oriented stuff that a random visitor or link-follower from the Cafe or Lab might find provides additional detail or discussion, plus (2) a bunch of self-referential procedural-technical nit-picking stuff that looks a lot like this thread but is running outside this n-Forum subforum.
So I thought the easiest thing to do would simply be to broaden the purview of this staff-meeting sort of subforum to include all non-content-oriented stuff.
]]>I think it's just a Gestalt figure-ground attention thing — my eyes tend to glaze over when I see a lot of titles like:
It's some kind of cognitive-visual work to sort the content-oriented titles from the procedural-technical ones, and if I'm driven to the extremity of having to call on motor functions, to reach out for the scroll bar, much less click on the next page button, well …
At any rate, I'm guessing that casual visitors looking for content-ment in our n-virons are most likely even more n-dolent than I am …
]]>I just became aware of how many substantive discussions I'm overlooking because of all the technical posts in the "General" category that are pushing them beneath my threshold of attention.
I wouldn't want to create another category, so maybe the category currently called "n-Forum" could be expanded to include all discussion of technical issues, if you cache my drift …
]]>McCulloch thought that abductive reasoning occurred in the Reticular Formation (RF), if I recall — see the papers collected in Embodied Mind. Early work on the link between Context-Free Languages and Push-Down Automata made explicit use of language about hypothesis testing, where the available hypotheses were represented by the non-terminal symbols.
The learner I wrote never got that smart — I got a Master's out of it but never got around to documenting the program in full. I made another stab at documentation and exposition a few years back, and what I've done so far is currently in this web neighborhood:
]]>You are certainly playing out a number of different themes there.
As far as primitive models of learning go, I did spend one of my parallel lives getting a Master's in Psych, and part of what I did for that is to write a program for 2-level formal language learning that operated on ideas from Thorndike (law of effect + law of practice) and Guthrie (1-shot learning). Not that I was especially behaviorist or anything, but simply because that was the first step in trying program anything more complex.
Learning sequences of features in a Markov way is more or less the same thing as learning regular languages. Getting up to context-free languages is quite a bit harder — it requires a capacity for abductive reasoning, that is, making hypotheses and testing them against the data stream.
]]>Thanks, Andrew, that will do nicely until I can get back to the drawing board.
By the way, where would all those templates be documented?
Jon
]]>Newton? Hecuba, no, I was talking about Russell.
I see you're a bit more punctilious about your puns than the pun of the mill I grind out.
I guess it all depends on how much wine the boys've been cippin' …
Goodbye, Mr.
]]>Re: Jon, I think this is one of your less successful puns, since that's supposed to be a hard c in "principia".
Not to an old altar boy like me …
Anyway, it was supposed to be spin-off from our discussion of shells in the Cafe, since I sense a certain FOL-IE in trying to stuff any more math into the nutshell of FOL.
]]>Less Categories, More Functors.
]]>Good night, sweet principia …
]]>