LMFDB Work

If you choose to tensor product two elliptic curves, then select a random elliptic curve in each box and click on “tensor these”, you get a server error

http://beta.lmfdb.org/TensorProducts/show/?obj1=EllipticCurve%2FQ%2F272080%2Fl%2F2&obj2=EllipticCurve%2FQ%2F132800%2Fc%2F1

Elliptic Curves- (assigned to John Cremona)

Picture looks weird; smaller; numbers just off the edge http://beta.lmfdb.org/EllipticCurve/Q/102/b/

I (JEC) do not understand the problem. Which numbers are off which edge? If you want to the isogeny class to be smaller I have been working on that.

Elliptic Curves- (assigned to John Cremona)

If I want to look for a curve whose Jacobian has torsion Z/29Z, I'm supposed to write [29] in the "torsion" field, and this works; however, if I forget teh brackets and just write "29", it produces a server error. I think this should be changed either into an explicit "syntax error" message, or (better) make the search more clever so that it understands 29 as well as [29]. http://beta.lmfdb.org/Genus2Curve/Q/

The same issue for elliptic curves has been fixed (for curves over Q): check it out at beta. It would be easy to adapt that here. More seriously: the search page seems to suggest that the torsion can have at most 2 cyclic factors, but it it should be at most 4! Pull request #383 deals with this -- but error handling of the input on this page is almost nonexistant still. FIXED

David will email Pascal about issue with Hecke characters

Classical Modular Forms:

Documenting the structure of the code: http://scripts.mit.edu/~kedlaya/wiki/index.php?title=Classical-mf-structure

When you put something inappropriate into a search box, instead of letting you know as in the elliptic curves page, it gives a server error. http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/ranges?level=r&weight=&character=0&Submit=Search gives a TypeError: coercing to Unicode: need string or buffer, int found

General: wrong fields of coefficients, and too large display. Choose better generator for field of coefficients.

On http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/9/12/1/ , the coefficents of form c are defined in terms of a2, which is not defines clearly - all that is clear is that it lies in Qsqrt70. Besdies, clicking on form c reveals that a2 is actually 6sqrt70, and not sqrt70 as one might think. This is extremely confusing. Maybe it would be good to have a script looking at the modular forms in the database, searching for the field of coefficients in the database of number fields and replacing the coefficients with a new expression in terms of the generator that is in the database?

http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/23/12/1/

Page does not load (it only does once in a while): http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/23/10/1/

Crash trying to use truncate_powerseries: http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/24/8/1/ http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/24/10/1/

Error message: "Problem with Hecke orbits in the datbase!": http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/24/12/1/ http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/18/6/1/

General (wishlist): Need more that 100 coefficients for the data to be useful

Search for a weight >=26 crashes: http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/ranges?level=&weight=26&character=0&Submit=Search produces AttributeError: 'str' object has no attribute 'get'

Range searches don't work (not implemented?): neither level nor weight

Search for Gamma_0 redirect to search for Gamma_1

dimension not clickable: http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/9/10/

This incorrect url should give an error instead of redirecting (confusing): http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/11/2/0/a/

Wrong root of unity is defined: http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/13/4/4/

"Show more coefficients" is maxed at 87 even though 100 coefficients are stored. Crash when trying to download more than 100 coefficients. http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/1/12/1/a/

==

Also, a wish list! Please also email Holly with things you would like to see here.

In general, the database/website should be explicit about what it knows and doesn't know, for example, the ranges that are available for downloading coefficients of modular forms. It shouldn't give you a server error every time a page doesn't exist or information is not in the database.

http://beta.lmfdb.org/ModularForm/GL2/Q/holomorphic/51/2/0/

Classical Modular Forms-

Add Eisenstein forms, with toggle for viewing all forms, only cuspforms, or only newforms.

Elliptic Curves-

Add higher rank elliptic curves? By itself this makes little sense. But there are many other collections of elliptic curves which could be added, of course. The work this would involve includes: preparing data produced by others into a standard format for upload; having a way to attach a code indicating which collection any one curve belongs to, for proper attribution; a page explaining which collections are included; and for individual curves it will not be practical to include all the data which we currently have for every curve, so all the code needs to be adjusted to allow for incomplete data. [Example: I started to add generators to the Stein-Watkins curves of prime conductors up to 10^10; for ranks>1 these had been found 10 years ago by myself (JEC), Watkins and Fisher, but I was using Heegner points to find the rank 1 generators. However after 3 months running on 64 cores I was still only 3% though the list, at which point I moved on.]

Galois Groups-

A knowl or would be nice to explain the notation or origin of name for the uglier groups.

@@ Line 111: / Line 111: @@
 *Elliptic Curves-
-Add higher rank elliptic curves?  By itself this makes little sense.  But there are many other collections of elliptic curves which could be added, of course.  The work this would involve includes:  preparing data produced by others into a tsandard format for upload; having a way to attach a code indicating which collection any one curve belongs to, for proper attribution; a page explaining which collections are included;  and for individual curves it will not be practical to include all the data which we currently have for every curve, so all the codes to be adjusted to allow for incomplete data.  [Example: I started to add generators to the Stein-Watkins curves of prime conductors up to 10^10; for ranks>1 these had been found 10 years ago by myself (JEC), Watkins and Fisher, but I was using Heegner points to find the rank 1 generators.  However after 3 months running on 64 cores I was still only 3% though the list, at which point I moved on.]
+Add higher rank elliptic curves?  By itself this makes little sense.  But there are many other collections of elliptic curves which could be added, of course.  The work this would involve includes:  preparing data produced by others into a standard format for upload; having a way to attach a code indicating which collection any one curve belongs to, for proper attribution; a page explaining which collections are included;  and for individual curves it will not be practical to include all the data which we currently have for every curve, so all the code needs to be adjusted to allow for incomplete data.  [Example: I started to add generators to the Stein-Watkins curves of prime conductors up to 10^10; for ranks>1 these had been found 10 years ago by myself (JEC), Watkins and Fisher, but I was using Heegner points to find the rank 1 generators.  However after 3 months running on 64 cores I was still only 3% though the list, at which point I moved on.]
 *Galois Groups-
 A knowl or would be nice to explain the notation or origin of name for the uglier groups.

LMFDB Work

Revision as of 00:57, 24 October 2015

==

Views

Personal tools

Navigation

Search

Toolbox