Value, or otherwise, of running others' code

[R.D. Silverman]

Originally Posted by rogue

I attached a stand-alone program that will fully factor an input number with GMP-ECM.

So??? Writing such code is trivial. One justs puts together a wrapper
that calls GMP-ECM several times, using different methods, with
appropriate parameters.

A contribution to GMP-ECM *itself*, on the other hand, would be
something worthwhile.

[henryzz]

it might be really easy to make but a lot of people dont have programming skills and so cant make thier own

[R.D. Silverman]

Originally Posted by henryzz

it might be really easy to make but a lot of people dont have programming skills and so cant make thier own

So? Perhaps it might motivate people to reduce their ignorance by
*acquiring* such skills?

[rogue]

Originally Posted by R.D. Silverman

So??? Writing such code is trivial. One justs puts together a wrapper
that calls GMP-ECM several times, using different methods, with
appropriate parameters.

A contribution to GMP-ECM *itself*, on the other hand, would be
something worthwhile.

So what if it is trivial? For some projects it is useful (Home Prime Search), for others it isn't (Cunningham factorization). Some people focus more on finding factors than the methods used.

BTW, at what point does any piece of software become non-trivial? What might be trivial to you could be non-trivial to someone else. At the same time, something that is trivial to that other person is not trivial to you.

Agreed on your second point.

[R.D. Silverman]

Originally Posted by rogue

So what if it is trivial? For some projects it is useful (Home Prime Search), for others it isn't (Cunningham factorization). Some people focus more on finding factors than the methods used.

BTW, at what point does any piece of software become non-trivial? What might be trivial to you could be non-trivial to someone else. At the same time, something that is trivial to that other person is not trivial to you.

Agreed on your second point.

Focussing on "just finding factors" is a rather pointless activity. The
actual factorizations are not really useful unless they are being used for
something else; e.g. furthering the solution of some problem.

And, IMO, the "home prime" stuff is *totally* pointless. It is just an excuse
to avoid working on more difficult computations. It is another instance of
the IGG.

[xilman]

Originally Posted by R.D. Silverman

And, IMO, the "home prime" stuff is *totally* pointless. It is just an excuse to avoid working on more difficult computations. It is another instance of the IGG.

As you wish.

Alex Kruppa and I used HP(49) to gain real practical experience with non-trivial GNFS factorizations. Given that some of them took us weeks or months work, using a significant number of machines between us, I differ with your categorization as "instant gratification".

Some day when we want to learn how to factor 200+ digit hard numbers, there's a chance we may return to HP(49).


Paul

[R.D. Silverman]

Originally Posted by xilman

As you wish.

Alex Kruppa and I used HP(49) to gain real practical experience with non-trivial GNFS factorizations. Given that some of them took us weeks or months work, using a significant number of machines between us, I differ with your categorization as "instant gratification".

Some day when we want to learn how to factor 200+ digit hard numbers, there's a chance we may return to HP(49).


Paul

Why??? There are plenty of C200+ numbers left in the Cunningham project
and these factorizations *are* sometimes of interest to mathematicians.

[xilman]

Originally Posted by R.D. Silverman

Why??? There are plenty of C200+ numbers left in the Cunningham project
and these factorizations *are* sometimes of interest to mathematicians.

Why not?

In my opinion, the value of knowing the prime factorizations of the great majority of Cunningham numbers, and of all of the Home Prime numbers, lies somewhere in the range between nil and neglible.

Secondly, there is no shortage of people and machine resources willing to be devoted to Cunningham. I spend some time on them myself. If you regard the Cunningham project as so much more important than factoring any other integers, why do you spend so much effort on the a^n \pm b^n tables?

Finally, there's no great rush, as I see it, in "finishing" the Cunningham tables. For a start, it's an endless task because they are invariably extended when only a few composites remain in the tables du jour and, any way, are you looking for instant gratification yourself or are you prepared to wait a few months longer than strictly necessary to reach a particular level of factorization in the Cunninghams?


Paul

[R.D. Silverman]

Originally Posted by xilman

Why not?

In my opinion, the value of knowing the prime factorizations of the great majority of Cunningham numbers, and of all of the Home Prime numbers, lies somewhere in the range between nil and neglible.

Secondly, there is no shortage of people and machine resources willing to be devoted to Cunningham. I spend some time on them myself. If you regard the Cunningham project as so much more important than factoring any other integers, why do you spend so much effort on the a^n \pm b^n tables?

Paul

Some of the Cunningham numbers (those that represent the order of finite
fields. i.e. prime bases) can be useful. And they have a great deal of
historical interest.

AFAIK, noone outside this forum knows about the "home primes". They
have no real mathematical structure.

And you are correct in that the actual numerical value of the factors
really has little value. (except in the case I mentioned). They have a
lot of value as a measure of progress in algorithms however. And for this
purpose I see no need for so many different factoring projects.

"so much effort" on the homogeneous Cunninghams? I have a single,
6 year old Pentium 4 at 1.5Ghz working on them. For such small numbers,
the process is almost completely automated. Sometimes I do need to
intervene when the filtering process "over-compresses" the matrix, but
I spend very little time on them. As for why I do it, I started these
tables myself over 20 years ago. I used them as "small" numbers to
test new code, while production code was running on Cunninghams.

[wblipp]

Let me paraphrase this to see if understand it.

Originally Posted by (paraphrased)

Of all the pointless worthless factoring projects, I like only the Cunningham Project. I'm a smart guy, so everybody should like only the Cunningham project.

There is another pointless worthless factoring project for which I have recruited members. But it's a project of my own invention and gives me pleasure. I'm a smart guy, so people should get more pleasure from my pointless worthless invention than from their own pointless worthless inventions.

Do I have it right?

[bdodson]

Originally Posted by wblipp

Let me paraphrase this to see if understand it.
...Do I have it right?

Uhm. Seems a bit sharp, from a bright guy, whose 11-smooth project
seems to have given some enjoyment to the people participating. And
anyway, we're all jealous that so many of Paul's projects catch-on so
well (Cunningham extensions, for a recent one we've recently been
considering). Hard to say which topics retain interest over time. Can't
claim that 2-3 decades of pounding on a particular topic gives anyone
a pass on self-centered-ness; still, no reason to join-in on that!
-Bruce