Differences between revisions 3 and 35 (spanning 32 versions)
⇤ ← Revision 3 as of 2010-06-29 00:32:33 →
Size: 978
Editor: JohnCremona
Comment:
← Revision 35 as of 2011-03-23 23:32:23 → ⇥
Size: 4659
Editor: Jamie Weigandt
Comment:
Deletions are marked like this. Additions are marked like this.
Line 5: Line 5:
   -- Note: Items 1,2, and 4 are all *broken* until we resolve the bug with simon_two_descent. That is my project now since computing ranks (and descending to $\mathbb{Q}$: thanks Erin!) will be important in the Elkies-Watkins search over $\mathbb{Q}[i]$. (Jeremy)
Line 6: Line 7:
   1. Regulators over number fields
Line 8: Line 8:
   2. Linear dependencies over number fields    1. Regulators over number fields. NB over $\mathbf{Q}$ there are two functions, regulator() and regulator_of_points(), and it is the second of these which can be trivially implemented. Just copy the code from ell_rational_field.py into ell_number_field.py; also copy height_pairing_matrix(). In each case I mean "move" rather than "copy". Then in each case add doctests with a couple of examples over number fields. The other function (regulator) is something which depends on having an actual MW basis, so is not for now. See ticket #9372: JEC has uploaded a patch which is ready for review. (Aly and Jeremy -- reviewed, assuming just the doctest needed changed, positive review)
Line 10: Line 10:
   3. Finish ticket #8829 on saturation over number fields.    2. Linear dependencies of points over number fields (and over Q), modulo torsion. The attached file [[attachment:mwnf.m]] has Magma code for this (written by me) which may be useful, though you will have to find the Sage equivalent of LLLgram(). That function takes a gram matrix (of not-necessarily full rank) and returns a reduced gram matrix and a matrix in GL(n,Z) and the rank r. (Aly, Jeremy, Jim) $\mathbf{Note:}$ the Sage equivalent of LLLgram() is LLL_gram() which is a method for integer matrices. Second Note: Pari supports LLL_gram for real matrices as well, which we apparently need. There is already a wrapper, but we should add a member method to real matrices (I don't know exactly where this should go yet, haven't looked -- Jeremy).
Line 12: Line 12:
   4. Kilford's rank function patch #9342: complete and review.
Line 14: Line 13:
   5. Integral points over number fields    3. Finish ticket #8829 on saturation over number fields. [This depends on item 1.] Probably a job for JEC since he reviewed robertwb's patch and suggested changes, which he will now implement.
Line 16: Line 15:
   6. S-integral points over number fields    4. Kilford's rank function patch #9342: completed needs review. (Aly and Jeremy)(Done, has a positive review)
Line 18: Line 17:
   7. S-class groups    5. Integral points over number fields: ongoing project work re-implementing Nook's Magma code. - This code is done, and output matches the Magma code (and, over Q, it matches what is already implemented in sage). (Jackie and Rado)[[attachment:intpts.sage]]
Line 20: Line 19:
   8. K-Selmer groups esp. $K(S,4)$ and $K(S,2)_4$    6. $S$-integral points over number fields: ongoing project work, based on Smart-Stephens and taking as a model the code over $\mathbb{Q}$ for $S$-integral points.
Line 22: Line 21:
   9. All curves with e.g.r.    7. $S$-class groups: on going project work.
Line 24: Line 23:
   10. All curves with e.g.r. outside S (using most of the above)    8. K-Selmer groups esp. $K(S,4)$ and $K(S,2)_4$: ongoing project work.
Line 26: Line 25:
   11. Iterating through number field elements by height; special case for integers; simpler version not by height.    9. All curves with e.g.r.: putting together a lot of the above! When the time comes, I have some complete lists over imaginary quadratic fields of class number 1 which will be helpful.
Line 28: Line 27:
   12. Heights on projective space over number fields (Jackie's code)    10. All curves with e.g.r. outside $S$ (using most of the above again)
Line 30: Line 29:
   13. Elkies-Watkins over number fields -- post what function you have by the end of the week.    11. Iterating through number field elements by height; special case for integers; simpler version not by height. Ongoing project work.
Line 32: Line 31:
     12. Heights on projective space over number fields (Jackie's code: make into a ticket+patch)

   13. Elkies-Watkins over number fields -- post what code you have by the end of the week.
        -- This works very well now. I am working to get access to sage.math to run it on that server (rather than on my computer at home). I need to find a good place, if there is one, to add this into the sage library. Maybe in a miscellaneous file somewhere. (Jeremy)

   14. Implement a $2$-torsion rank function over number fields. Jamie is working on this: #9371. Patch is up, needs review. Generalized to any field at Sage Days 29. Positive Review.

   15. Given an elliptic curves defined over $K$, determine whether it is isomorphic over $K$ to a curves defined over $\mathbb{Q}$. [Get the j_invariant, decide if it is rational, if so construct a curve over Q with that j-invariant, base change itback to K, and test for the quadratic twist. All these steps are implemented already. See attached example [[attachment:eg.sage]] ] (Erin - patch posted #9384)
          *I've fixed this function, descend_to. It now returns the correct twist for the general case and works when j=0,1728.
     However, I haven't posted a replacement patch yet because I want the output to be an elliptic curve that sage recognizes is defined over the subfield K of L.
     This will hopefully be fixed by the end of the day.
          *Update: Fixed!

To-do list for elliptic curves in Sage

This is a list of things (small or large) which came up during SD22 at MSRI as deserving to be fixed or implemented in Sage.

  • -- Note: Items 1,2, and 4 are all *broken* until we resolve the bug with simon_two_descent. That is my project now since computing ranks (and descending to \mathbb{Q}: thanks Erin!) will be important in the Elkies-Watkins search over \mathbb{Q}[i]. (Jeremy)

  • Regulators over number fields. NB over \mathbf{Q} there are two functions, regulator() and regulator_of_points(), and it is the second of these which can be trivially implemented. Just copy the code from ell_rational_field.py into ell_number_field.py; also copy height_pairing_matrix(). In each case I mean "move" rather than "copy". Then in each case add doctests with a couple of examples over number fields. The other function (regulator) is something which depends on having an actual MW basis, so is not for now. See ticket #9372: JEC has uploaded a patch which is ready for review. (Aly and Jeremy -- reviewed, assuming just the doctest needed changed, positive review)

  • Linear dependencies of points over number fields (and over Q), modulo torsion. The attached file mwnf.m has Magma code for this (written by me) which may be useful, though you will have to find the Sage equivalent of LLLgram(). That function takes a gram matrix (of not-necessarily full rank) and returns a reduced gram matrix and a matrix in GL(n,Z) and the rank r. (Aly, Jeremy, Jim) \mathbf{Note:} the Sage equivalent of LLLgram() is LLL_gram() which is a method for integer matrices. Second Note: Pari supports LLL_gram for real matrices as well, which we apparently need. There is already a wrapper, but we should add a member method to real matrices (I don't know exactly where this should go yet, haven't looked -- Jeremy).

  • Finish ticket #8829 on saturation over number fields. [This depends on item 1.] Probably a job for JEC since he reviewed robertwb's patch and suggested changes, which he will now implement.
  • Kilford's rank function patch #9342: completed needs review. (Aly and Jeremy)(Done, has a positive review)

  • Integral points over number fields: ongoing project work re-implementing Nook's Magma code. - This code is done, and output matches the Magma code (and, over Q, it matches what is already implemented in sage). (Jackie and Rado)intpts.sage

  • S-integral points over number fields: ongoing project work, based on Smart-Stephens and taking as a model the code over \mathbb{Q} for S-integral points.

  • S-class groups: on going project work.

  • K-Selmer groups esp. K(S,4) and K(S,2)_4: ongoing project work.

  • All curves with e.g.r.: putting together a lot of the above! When the time comes, I have some complete lists over imaginary quadratic fields of class number 1 which will be helpful.

  • All curves with e.g.r. outside S (using most of the above again)

  • Iterating through number field elements by height; special case for integers; simpler version not by height. Ongoing project work.
  • Heights on projective space over number fields (Jackie's code: make into a ticket+patch)
  • Elkies-Watkins over number fields -- post what code you have by the end of the week.
    • -- This works very well now. I am working to get access to sage.math to run it on that server (rather than on my computer at home). I need to find a good place, if there is one, to add this into the sage library. Maybe in a miscellaneous file somewhere. (Jeremy)

  • Implement a 2-torsion rank function over number fields. Jamie is working on this: #9371. Patch is up, needs review. Generalized to any field at Sage Days 29. Positive Review.

  • Given an elliptic curves defined over K, determine whether it is isomorphic over K to a curves defined over \mathbb{Q}. [Get the j_invariant, decide if it is rational, if so construct a curve over Q with that j-invariant, base change itback to K, and test for the quadratic twist. All these steps are implemented already. See attached example eg.sage ] (Erin - patch posted #9384)

    • I've fixed this function, descend_to. It now returns the correct twist for the general case and works when j=0,1728.
    • However, I haven't posted a replacement patch yet because I want the output to be an elliptic curve that sage recognizes is defined over the subfield K of L. This will hopefully be fixed by the end of the day.
      • Update: Fixed!

days22/cremona/to-do-list (last edited 2011-03-23 23:32:23 by Jamie Weigandt)