- Henry Cohn:
- - I hate the E.conductor(). I would like conductor(E). - Make it so SAGE can open any format file. MATLAB can do.
- Load a jpeg; fancy demos.

- Want SAGE to be able to text message you when a - Databases of math objects: - - should have 50 times as many - easy for random people to submit - e.g., I use Magma just for sloane's lattice tables, - refereed: people sign off on a database. - personal databases

- Specific math that should be more developed: - - Optimization:
- conj grad
- newton
- grad descent

- Numerical computation of f'(x) in an *intelligent* way. - Give easy black boxes for clueless people but well - labeled boxes for other people.

- Linear programing: - the best free ones are all mediocre,
- so I have to use CPLEX (?). $10-20K/cpu/year.

- But we could do exact, arbitrary precision, and double
- precision LP solver that could do tons of comptuations with doubles without too much overhead.

- Semidefinite programming: - nothing arbitrary precision available
- implement a descent but not so super-good arbitrary precision

- Automorphism groups: - nauty.
- But I want edge-labeled graphs!! It would be tremendously useful to deal with edge-labeled (hyper) graphs.

- Cellular automata: - make it easy for someone with no clue to get beautiful
- visualizations of cellular automata.

- fractals; julia sets

- Analytic Riemann Surfaces: - explicit covering maps in terms of hypergeometric functions
- Bernard Deconinck (from UW applied math) -- maple code

- Lesson plans: - Publish class-room tested lecture plans that use
- Get undergrads to make lesson plans.
- NSF GK-12: UW grad students involved in local community.

- Publish idealogical articles in off-beat places - You don't have to rely on other people; you get to
- Students can makes things arbitrarily more complicated.

days4schedule/cohn (last edited 2008-11-14 13:42:07 by localhost)