Proposal for partially known labels

[davidfarmer]

We have the possibility of classical modular forms computed numerically, outside the range where we can obtain algebraic expressions for the coefficients.

If I understand correctly, we may not know the Galois orbits, so we would only know part of the label of the form.

This is comparable to high-rank elliptic curves, or the curves in the Stein-Watkins database: we have not put them in the LMFDB because we don't know their labels.

Here is a proposal, partially based on a discussion Stephan Ehlen and I had at AIM:

Use an underscore, or a label starting with an underscore, for the unknown parts of the URL.

An elliptic curve with a large conductor might have this URL:

/EllipticCurve/Q/1234567890000/_a/_1

or maybe you only need to use an underscore for the first unknown quantity:

/EllipticCurve/Q/1234567890000/_a/1

or maybe this is better, because (in the case of Galois orbits of modular forms) you might not know if two different objects have the same unknown part of the label:

/EllipticCurve/Q/1234567890000/_/1

Read the underscore as "fill in the blank"

As things are added to the database, arbitrary names can be given to the underscore and post-underscore parts.

Labels would have the same length as "real" labels, and labels can be permanent in the sense that once the complete label of an object is determined, the database entry with information about the underscore labeled object can be updated to point to the correctly labeled version.

[jwj61]

This sounds reasonable to me. It looks like the final details just need to be hammered out.