Prime-Wiki

[Happy5214]

Originally Posted by kar_bon

I've updated the extension ExternalData to the newest version, but this did not eliminated the error.
I've inserted into the mime.types file of the MediaWiki source the line

application/csv csv

there was only an entry with "text/csv csv".
I've uploaded a new version of the data file for Proth primes of the form k*b^n+1, least n-values and no error occured.

Try again now if the issue is solved for now.

I've searched for an error like this, but found no solution, an update to a newer MediaWiki version is not needed (and could also create other issues then).

The issue is now fixed, thank you. I looked at the current version of the MediaWiki code while investigating the issue, and the newer versions appear to have this problem too (they only list text/csv as a valid CSV MIME type), so an upgrade would not have fixed this issue (instead causing many others). I don't know if something changed in the browser support for CSV files, but I'm probably going to file a bug report with MediaWiki about this.

[Happy5214]

I'll preface this by saying that I haven't discussed this with anyone, but since you've already thought of the idea of referencing FactorDB numbers as a Prime-Wiki namespace, I'm wondering if you'd be open to the idea of using PW as a centralized database of ECM progress for general factorization efforts? They would be added as either remarks or a new field to long numbers, using the standard wiki process.

[kar_bon]

I don't know exactly, what do you mean here? Can you give an example, please.

[Happy5214]

Originally Posted by kar_bon

I don't know exactly, what do you mean here? Can you give an example, please.

As an example, https://www.mersenneforum.org/showpo...postcount=1006 stated that ryanp performed 4200 ECM curves on http://factordb.com/index.php?id=1100000002997883099 (the cofactor of 1528152860898312226820507829734311038694803153043007^5-1) at B1=26e7. We could have a "long number" page for that linked FactorDB entry (or, alternatively, the entry for 1528152860898312226820507829734311038694803153043007^5-1, which would likely be more stable if additional factors are found but the number isn't completely factored) with a history entry in the remarks saying something like "2022-03-30: Performed 4200 curves at B1=26e7 by Ryan Propper", with a link to the forum post.

[kar_bon]

I think this is not the right place to store such information.

Questions:
Who should put those intormation in? ryanp or Batalov?
Those information are even more unverifiable than for say Riesel numbers.
The amount of such numbers are big, especially for the OPN roadblock files. The work to create a page for such number is more time-consuming than the benefit any person can have from.
How will those numbers be stored? How find them in the Wiki? Storing by type or length or project?

There was such information stored in the FactorDB but not available?
Studio Kamada does a good job for documenting ECM efforts, as example for (185*10^272-2)/3 (the same number in FactorDB):
- you can reserve such numbers
- automated generated ecm-call available
- fill in your amount of ecm done

This system would be best for documenting such work, but not the Wiki I think.

[Happy5214]

Originally Posted by kar_bon

I think this is not the right place to store such information.

Questions:
Who should put those intormation in? ryanp or Batalov?
Those information are even more unverifiable than for say Riesel numbers.
The amount of such numbers are big, especially for the OPN roadblock files. The work to create a page for such number is more time-consuming than the benefit any person can have from.
How will those numbers be stored? How find them in the Wiki? Storing by type or length or project?

There was such information stored in the FactorDB but not available?
Studio Kamada does a good job for documenting ECM efforts, as example for (185*10^272-2)/3 (the same number in FactorDB):
- you can reserve such numbers
- automated generated ecm-call available
- fill in your amount of ecm done

This system would be best for documenting such work, but not the Wiki I think.

AFAIK FactorDB never had in-progress ECM progress data for yet-to-be-factored composites (at least I'd never seen it), just the data for the curve that found the factor. While I agree a purpose-built site like Studio Kamada would be better suited, that particular site appears to only cover the near-repdigit effort, so that doesn't appear to be useful in its current form to other efforts (e.g. OPN, high-importance aliquot, Cunningham).