Profile

Cover photo
Gershom B
Works at S&P Capital IQ
Attended University of California, Berkeley
Lived in New York
786 followers|36,015 views
AboutPostsPhotosVideos

Stream

 
 
Haskell.org Survey on Crowdfunding

The haskell.org committee (http://www.haskell.org/haskellwiki/Haskell.org_committee) oversees how haskell.org funds are spent. Up until now, this has consisted simply of paying for hosting services, funded mostly by proceeds from our involvement in the Google Summer of Code program. 
Now that we are set up to accept donations (http://www.haskell.org/haskellwiki/Donate_to_Haskell.org), we are exploring additional ways to more proactively benefit the open source Haskell community.  In particular, if there is work that the community is willing to fund, haskell.org can serve as a funding conduit and organizing force to make that work happen.

By filling out this survey, you can help us gauge community interest and learn where we can make the biggest impact.
Drive
haskell.org funding surveyThe haskell.org committee (http://www.haskell.org/haskellwiki/Haskell.org_committee) oversees how haskell.org funds are spent. Up until now, this has consisted simply of paying for hosting services, funded mostly by proceeds from our involvement in the Google Summer of Code program. Now that we are set up to accept donations (http://www.haskell.org/haskellwiki/Donate_to_Haskell.org), we are exploring additional ways to more proactively benefit ...
1
Add a comment...

Gershom B

Shared publicly  - 
 
 
So this is pretty neat: Some people decided to study the structure of the primes by treating the natural numbers as a stochastic process.

Here's the idea: You can think of the natural numbers as being a giant network, with composite numbers connected to their prime factors. You can use the ordering of the naturals to think of this as a graph that's growing over time -- each number gets added in, one by one, connecting to all its prime factors if it has any, and getting labeled as "prime" if it doesn't.

What this paper does, is it gives a very simple and intuitive randomized algorithm for generating structures that look kind of like this graph. It turns out that the structures that are generated share a lot of the large-scale structure of the primes! For example, they obey the Prime Number Theorem (http://en.wikipedia.org/wiki/Prime_number_theorem), as well as a few other theorems about the distribution of primes.

From the article it seems like this is a pretty major step forward as far as probabilistic models of the primes go; it mentions a couple of issues (notably in modeling small-scale structure; the model assigns nonzero probability to consecutive primes, oops) but overall it sounds pretty awesome.
6
Add a comment...

Gershom B

Shared publicly  - 
1
Allan Erskine's profile photo
 
rip -- that track / riddim is a mighty legacy
Add a comment...

Gershom B

Shared publicly  - 
 
 
How many ways can a tie be tied?

What if I told you: there are 177147 distinct ways in which a normal necktie can be tied, including the exotic knots worn by the character Merovingian in the Matrix films? That's according to a recent paper (http://www.arxiv.org/abs/1401.8242) by Hirsch, Patterson, Sandberg and Vejdemo-Johansson.

The paper builds on work from 2000 by Fink and Mao, who identified 85 distinct tie knots. However, Fink and Mao were only interested in tie knots with trivial façade, meaning that the visible parts of the tie look like a standard tie knot. Since then, interest in more exotic tie knots has increased. This is in part because of the knots worn by the character Merovingian (also known as “The Frenchman”) in the films The Matrix Reloaded and Matrix Revolutions. The knot in the picture is an example of one of the knots Merovingian wears; note that it does not resemble a standard tie knot, even superficially. This means that the Merovingian knot does not have trivial façade, and does not appear in the Fink-Mao classification.

What Hirsch et al do is to classify the tie knots that take up to 11 moves to tie and that are anchored by a single depth tuck. A single depth tuck means that the tuck only passes under the most recent bow made over the knot. The possible moves used to tie the knot are denoted by T (clockwise), W (anticlockwise/counterclockwise) and U (tuck); the letters themselves stand for “Turnwise”, “Widdershins”, and “Under”. The number 11 is somewhat arbitrary, and was chosen because the well-known Eldredge tie knot uses 11 moves.

Hirsch et al develop a tie-knot language, which they use to classify their possible necktie knots. Each knot has to end in a tuck move, or the knot would fall apart under gravity. They prove that the tie-knot language is regular, which means that it can be recognised by a finite-state automaton. The automaton is shown in the top right of the picture. The basic idea is to follow a path that starts and finishes at the double circle in the middle of the picture, following the arrows and taking note of the sequence of letters so obtained. 

An example of an acceptable necktie knot by the authors' definition is the sequence TWWuWWuTTuTTuTTU, which appears in the paper as sequence L-447. This is an acceptable sequence because it is possible to trace a path in the diagram that (a) starts and ends at the middle and (b) follows this sequence of letters, and furthermore, there are at most eleven occurrences of T and/or W in the sequence.

The paper lists 2046 sequences like the one above. The lower case occurrences of U denote places where it would be possible to make a more complicated tuck than a single depth tuck. This means the number of possible knots is much bigger than 2046, which is where the figure of 177147 comes from. However, the final tuck is always a single depth tuck, and is therefore denoted by a capital U.

If you'd like to know how to tie the Merovingian knot in the picture, there is a nice stop-motion animated video on YouTube showing how to do it (Animated How to Tie a Necktie Merovingian Knot for your Necktie aka Ediety Knot - How to Tie a Tie).

#mathematics #scienceeveryday #spnetwork arXiv:1401.8242
14
3
P Tufts's profile photoZac Slade's profile photo
Add a comment...
 
It may be the case that you have seen Inside Llewyn Davis or intend to, and yet do not know much of the music of Dave Van Ronk. In which case you may enjoy this.
2
Add a comment...
Have him in circles
786 people
Don Stewart's profile photo
Dan Christensen's profile photo
Edward Kmett's profile photo
Ashish Agarwal's profile photo
 
Haskell.org Survey on Crowdfunding

The haskell.org committee (http://www.haskell.org/haskellwiki/Haskell.org_committee) oversees how haskell.org funds are spent. Up until now, this has consisted simply of paying for hosting services, funded mostly by proceeds from our involvement in the Google Summer of Code program. 
Now that we are set up to accept donations (http://www.haskell.org/haskellwiki/Donate_to_Haskell.org), we are exploring additional ways to more proactively benefit the open source Haskell community.  In particular, if there is work that the community is willing to fund, haskell.org can serve as a funding conduit and organizing force to make that work happen.

By filling out this survey, you can help us gauge community interest and learn where we can make the biggest impact.
Drive
haskell.org funding surveyThe haskell.org committee (http://www.haskell.org/haskellwiki/Haskell.org_committee) oversees how haskell.org funds are spent. Up until now, this has consisted simply of paying for hosting services, funded mostly by proceeds from our involvement in the Google Summer of Code program. Now that we are set up to accept donations (http://www.haskell.org/haskellwiki/Donate_to_Haskell.org), we are exploring additional ways to more proactively benefit ...
4
4
Darren Grant's profile photoGershom B's profile photo
Add a comment...
 
College Junior? Like Functional Programming? Want to be an intern in New York this summer? Maybe this will be of interest.
1
Add a comment...

Gershom B

Shared publicly  - 
 
 
Registration is open for the 1st ever New York City Haskell Hackathon!
Info: www.haskell.org/haskellwiki/Hac_NYC
#haskell   #hackathon   #nyc 
1
1
Don Stewart's profile photo
Add a comment...

Gershom B

Shared publicly  - 
 
Good news, everybody!
 
Playing around with Haskell in https://cloud.sagemath.com

Auto-indent and syntax highlighting thanks to CodeMirror; haskell thanks to Linux :-). 

(I just added this -- aggressive browser refreshing may be required.)
2
Jim Stuttard's profile photoGershom B's profile photo
3 comments
 
Tnx. I just found "new->terminal" and it works fine on salvemundo.hs. It would be great if fpcomplete and this site managed to  synchronise continuous build library sets.
Add a comment...

Gershom B

Shared publicly  - 
 
These are Howard's original notes (well, a later published version of them) that established the well known Curry-Howard correspondence and the "propositions as types" principle in 1969.

Now a great deal of work was done on realizing proofs as mathematical objects prior to Howard's work (and the notes themselves cite some). But not only had the connection to types not been made, but it seems that type theory itself had fallen by the wayside between Church's work through the early 1940s and Howard's notes in 1969. 

Is that the case? What happened to types  during their "years in the wilderness"? Anything?
12
3
mukesh tiwari's profile photoUrs Schreiber's profile photoGershom B's profile photoHarley D. Eades III's profile photo
4 comments
 
I believe what is linked is in fact the version as published in 1980. That volume ("To H.B. Curry") is quite hard to come by, but I believe contains foundational work by e.g. Scott, Lambek, etc. as well. I would love to see it digitized and made available in its entirety on the web.
Add a comment...
People
Have him in circles
786 people
Don Stewart's profile photo
Dan Christensen's profile photo
Edward Kmett's profile photo
Ashish Agarwal's profile photo
Education
  • University of California, Berkeley
    Electrical Engineering and Computer Science, 1997 - 2001
  • Boston University
    African American Studies, 2004 - 2005
Basic Information
Gender
Male
Work
Occupation
Coder
Employment
  • S&P Capital IQ
    present
Places
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Previously
New York
Links
YouTube