search

LMFDB / lmfdb

L-Functions and Modular Forms Database
Other
253 stars 200 forks source link

fixes for issue #4943 #6268

Closed JohnCremona closed 5 days ago

JohnCremona commented 1 week ago

This addresses #4943 for ECQ. For the two lists of invariants to look consistent I added a symbol/name for all the quantities there, not just the real ones which may be approximate. All now do have names except the Sato-Tate group, and the same names appear in the appropriate knowls so a couple of those have had minor edits (Szpiro ratio, abc-quality). We could show something like ST(E) for the ST group of that is suffiently standard -- @AndrewVSutherland ?

I fixed but left the "(exact)"/"(rounded)" for analytic Sha since the knowl (originally written just for this purpose) does say something concrete, which is not obvious, namely that for both rank 0 and rank 1 the quantity shown is exact, having been computed as a rational, while for ranks 2 and above it is a rounded floating point number (not known theoretically to even be rational, let alone integral). I think this is markedly different from all the other approximate reals on the page. But of course I am willing to be persuaded!

Once this is done (with further changes if desired) I will do the same thing for ECNF.

AndrewVSutherland commented 1 week ago

I think ST(E) is fine.

roed314 commented 6 days ago

Should there also be symbols for conductor, discriminant and $$j$$-invariant? I agree that "(exact)" / "(rounded)" are good as is.

JohnCremona commented 6 days ago

Should there also be symbols for conductor, discriminant and j -invariant? I agree that "(exact)" / "(rounded)" are good as is.

I have put in "ST(E) = " with ST in mathrm and E in math mode.

About N, $\Delta$, j: yes for consistency, but consider this: these 3 have both values (integer or rational) and their factorization in a separate column. So after this extra change, either we put the factorizations into the same column as the value after " = ", with the equals not aligning; or we add a column. The latter would then require more work if you look at what happens after clicking on one of the code buttons: the code already appears in an extra, usually invisible column, and after the change (2nd option) all the code would be pushed further over.

I will try the first option and update the PR if it looks acceptable, otherwise I will try the second. I may even make a second branch so you can look at both.

Update: it was not as hard as I had thought to do both together using some "colspan=3" in the right places. Is this good now?

roed314 commented 5 days ago

Looks great!