Are the recent advancements for GIMPS (VDF, Gerbicz check, Jacobi check) worth publishing?

[chalsall]

Originally Posted by pinhodecarlos

Sometimes papers are on hold for months for peer review, goes back and forth for corrections (peer reviewers and editors)...

Yeah. This is in-and-of-itself is an artificial economy I don't have (much) time to rant about.

As the Author(s), you sometimes have to pay for the reviews.

Even then, YOUR paper will not be freely available to read by anyone who's not already a subscriber to the service. (Unless, of course, their institution already has paid the ransom for the knowledge).

Anyone else (the independent researcher; the student, et al) is charged between $10 to $100 USD to read the paper YOU wrote.

Can anyone quickly link to the "Pre-print" methodologies being used nowadays, to get around this profoundly stupid bovine excrement?

[pinhodecarlos]

Chris, moderate your language. Publishing a paper is a deal and final owner is the journal.

[chalsall]

Originally Posted by pinhodecarlos

I can provide any academic paper if you make me a list of DOI’s.

Just out of interest...

I don't have the DOI (it has one), but could you find "HIV Gateway: enhancing access to Caribbean HIV research via a public website" by Allen, C.F., Halsall, C., Hambleton, I.R. , Landis, R.C.

[chalsall]

Originally Posted by pinhodecarlos

Chris, moderate your language.

Bovine Excrement is considered offensive nowadays?

You should try living in my world.

I thought snowflakes weren't allowed in serious work.

Seriously.

[kruoli]

For every paper that's availible freely, I'm really grateful. In Corona days, Springer is even giving some free access to a lot of information I would have bet bet to be publically availible.

[kriesel]

Nominated so far, upthread, as notable recent advances are:
Jacobi symbol check as an occasional LL test error check;
Gerbicz Error Check on PRP test;
VDF and proof generation, Cert check as a sufficient and efficient validity check, reducing double check effort drastically while improving its reliability.

The whole development of gpuowl as a multiOS OpenCL based primality tester on gpus from multiple lines (AMD, NVIDIA, even some Intel IGPs) is interesting to me. Mihai and George have done a lot of work refining its fft performance. That's more in the software engineering direction.

If an article was constructed as a review of recent advances in the GIMPS algorithms, a section on future directions might be good to include.
Remember that there are multiple methods proposed for verification of TF NF work done, to ensure system and user reliability.
The remaining area that needs more error detection and verification attention than it's getting is P-1 factoring. There do seem to be some possibilities there, both in error detection, and in sharing some small fraction of the initial computations among multiple close-spaced exponents of similar bounds.

[chalsall]

Originally Posted by kriesel

If an article was constructed as a review of recent advances in the GIMPS algorithms, a section on future directions might be good to include.

A sincere question: Would you be willing to lead on the paper?

In the Social Sciences, all the authors are named alphabetically (I won't bore you on all the very enjoyable arguments that leads to).

In this case, clearly others did the "heavy lifting". You would simply be documenting their work.

But, sometimes, it takes someone skilled in the documentation arts to get something published.

Are you willing to take on this job?

Separately (and very importantly), are others comfortable with this idea?

[Uncwilly]

I think that we should wait a bit for George, Mihai, et al of those that were most deeply involved to consider this. Then, there should be a separate thread for this.

[chalsall]

Originally Posted by Uncwilly

I think that we should wait a bit for George, Mihai, et al of those that were most deeply involved to consider this. Then, there should be a separate thread for this.

Sound guidance.

[xilman]

Originally Posted by kriesel

Nominated so far, upthread, as notable recent advances are:
Jacobi symbol check as an occasional LL test error check;
Gerbicz Error Check on PRP test;
VDF and proof generation, Cert check as a sufficient and efficient validity check, reducing double check effort drastically while improving its reliability.

Not sure whether such a paper would be best placed in a mathematics journal or one specializing in computer science. Math. Comp. might be a possibility for the former category; Mathematics in Computer Science for the latter.

The latter 's author guidelines seem to imply that an institutional affiliation is necessary. That might be a (to me needless) hindrance to publication.

Bob: do you have any comments on the proposal and suggestions for where to submit, if anywhere?

[Batalov]

Nominating is a fine thing, and asking questions, too. But once the nomination is done - I'd suggest to leave the decision making to the scientific contributors. Robert, Mihai, George.

Otherwise it will start looking quite a bit like Stephen King's "Misery". "I will make you write this new novel, I will break your legs if I will have to <make you write it>."

P.S. Every couple of years there are some threads here about "publishing" opinions written by folks who wouldn't know publishing if it kicked them in the Atlas. (borrowing from the latest bons mots du jour)