FAQ and information for interested amateurs

[davidfarmer]

Occasionally people want to contribute their data; often it is amateurs.

An FAQ, and/or a page which explains why various types of data are not helpful,
would make it easier to deal with such requests.

If we could indicate ways that an interested amateur could contribute (by computing
something, for example), that could be good.

An FAQ, independent of the issue about contributing data, might also be a nice
addition to the website.

[davidfarmer]

Here is what I wrote to someone who had non-rigorously computed a million zeta zeros beyond what is already in the database:

There are two reasons that the zeros you computed would not be
suitable for the LMFDB:

1. Those zeros are not rigorous, so we would be mixing rigorous and
   non-rigorous data.

There are non-rigorous data in the LMFDB, but we prefer rigorous data
when it can be made available. Having a mixed set would be a logistical
nightmare to keep track of.

2. Your data is, proportionally, a very small increase in the number
   of zeta zeros. As it is unlikely there is a scenario where it
   is helpful to have that tiny fraction of extra zeros, it would not be
   worth the effort of the maintainers to add that data (even if it were
   rigorous).

[setiastro]

Speaking as the amateur David referenced above: this FAQ would be genuinely valuable, and I'm happy to help think through what would have been useful from the outside-the-field perspective.

Two things that would have saved me time (and saved David a version of my exact email):

1. A clear statement of the rigor threshold, per data type. I understood before submitting that Platt-Trudgian data is interval-arithmetic-certified and my computation was not. What I didn't know was that "non-rigorous computation of the same data type as existing rigorous entries" is a hard boundary rather than a case-by-case discussion. A one-liner in the FAQ (something like: "For zeta zeros, we currently only accept interval-arithmetic-verified data. Non-rigorous computational catalogs are welcome to publish independently and coordinate with the community, but are not integrated into the LMFDB.") would have been decisive.

2. Concrete redirection for contributions that don't fit. For amateurs whose work is legitimate but out of scope, a paragraph naming appropriate alternative destinations would be very helpful. Options I've seen or been pointed toward: arXiv preprints, direct contact with specific researchers working in the area, community mailing lists like lmfdb-support itself, GitHub as a public-artifact record. The point isn't that the LMFDB should host everything, it's that an amateur who's genuinely put work in benefits from knowing where else it might go.

On the "ways amateurs could contribute" angle from the issue description: the model that works in astrophotography (which is where I come from) is roughly (a) small, well-scoped computational tasks that can be distributed to volunteers with home hardware, (b) a coordinator role that partitions the work and validates returns, and (c) a clear public artifact at the end that participants can point to. Adapting this to computational number theory would depend on whether there are specific bounded computations the LMFDB would find useful (extending some existing table, spot-checking a specific range, computing a specific invariant for a well-defined set of objects) that could be split across amateur contributors with rigorous verification built in. If any such tasks exist or could be defined, publishing them somewhere findable would probably attract more contributor interest than you'd expect. Amateurs generally want to do real work, not novel work, when the field's standards are unfamiliar.

Happy to help draft either of the FAQ items above if that would be useful. Also happy to be a case study in "what an amateur got wrong on their first contribution attempt" if that's a useful framing for the FAQ, since I made most of the mistakes it would presumably warn against.