Activity

From 15 Feb 2018 to 16 Mar 2018

08:57 Slug #1165: MinPoly over QQ: verification may be very slow: After many many experiments, code is settled as this:
- *MinPolyQuot(f, I, x)* makes the complete verification (over... Anna Maria Bigatti

14:12 Design #1168 (In Progress): ideal: does const ourGetPtr really need to be private?: If many functions have to be @friend@ that is not a good sign.
I'm too busy to think about it now, so suggest making... John Abbott
10:10 Design #1168 (In Progress): ideal: does const ourGetPtr really need to be private?: To be able to call this... Anna Maria Bigatti

18:48 Feature #1167 (In Progress): New class VerificationLevel: JAA has checked in a first impl (using KISS philosophy).
Anna will test it shortly.
The idea is that fns which ... John Abbott
18:43 Feature #1167 (Closed): New class VerificationLevel: To help offer the user the chance to choose between guaranteed algorithms and fast, probably-correct ones,
create a ... John Abbott

16:45 Feature #1131 (Resolved): Sturm sequence: The implementation seems to be mostly complete (for coeffs in QQ).
The elements of the sequence are scaled so that t... John Abbott
11:51 Design #1166 (In Progress): C++11: allows large integer literals (in NumTheory-prime.C): I'd like the code to compile whenever a 64-bit (or wider) integer type is available. This is guaranteed if the C++ s... John Abbott
11:47 Design #1166 (Closed): C++11: allows large integer literals (in NumTheory-prime.C): I would like to include some "large" (greater than 2^32) integer literals in @NumTheory-prime.C@.
This cannot be d... John Abbott
08:57 Slug #1165: MinPoly over QQ: verification may be very slow: Tried with TwinFloat: good for the 87 points (10 times faster) bad for example QQ-rand (10 times slower)
Anna Maria Bigatti
08:24 Slug #1165: MinPoly over QQ: verification may be very slow: An impressive example is given by 87 points in QQ^3.
@MinPoly@ takes ~4s and the verification take 44s on my Mac.
... Anna Maria Bigatti
08:05 Slug #1165 (Closed): MinPoly over QQ: verification may be very slow: The verification step after the modular algorithm may be very slow.
On the other side it is easy to construct exampl... Anna Maria Bigatti

18:18 Design #950: factor and SmoothFactor for integers --> FactorINT, FactorINT_TrialDiv, FactorINT_PollardRho: Maybe there could be a fn called *@IsPrime3@* which is guaranteed (reasonably) fast, and returns a @bool3@?
John Abbott
18:14 Design #1145 (In Progress): File names, coding conventions: Fns, Ops, Operations? part 2: JAA has a preference for *@Ops@* rather than *@Operations@*. I think they are equally clear, and see no advantage in... John Abbott
18:08 Feature #1155 (Resolved): Create a new "prime source" iterator: I have added 4 "iterators: *@PrimeSeq@*, *@PrimeSeqForCRT@*, *@FastMostlyPrimeSeq@*, *@NoSmallFactorsSeq@*.
First ... John Abbott
18:05 Design #1162: Check semantics of NoSmallFactorSeq: Currently the choice of biggest prime which is "filtered" is somewhat arbitrary.
Is it worth doing anything more ref... John Abbott

09:40 Feature #1158: New function: MinPolyQuotHeuristic --> MinPolyQuot with VerificationLevel: Added @error@ if CoeffRing is not QQ. Anna Maria Bigatti
09:14 Feature #1158 (In Progress): New function: MinPolyQuotHeuristic --> MinPolyQuot with VerificationLevel: Anna Maria Bigatti

18:22 Design #1163 (In Progress): Split SparsePolyRing.C in smaller files: Anna Maria Bigatti
18:20 Design #1163 (Closed): Split SparsePolyRing.C in smaller files: The file is now too long.
Split off code for ideals and for involutive operations,
and whatever can be usefully be ... Anna Maria Bigatti
16:40 Support #1161: Split NumTheory.C into several smaller files: I get a strange infinite loop on a 32-bit machine. Seems to be related to @FastMostlyPrimeSeq@.
Must investigate.... John Abbott
16:37 Design #1162 (Resolved): Check semantics of NoSmallFactorSeq: I have introduced a new class *@FastMostlyPrimeSeq@* which produces an increasing sequence of positive integers which... John Abbott

16:32 Design #1159 (In Progress): Add global enum "verify/DontVerify": After speaking to Anna...
* a "boolean" interface is easy to understand (and simple to use)
* a more sophisticated ... John Abbott
15:48 Design #1162 (Closed): Check semantics of NoSmallFactorSeq: I have just added the new class *@NoSmallFactorSeq@* but it is awkward to use.
Reconsider its public interface. John Abbott
15:41 Support #1161 (In Progress): Split NumTheory.C into several smaller files: I have split off the code to do with prime testing and generation. It is in *@NumTheory-prime@*
*Note* it is better... John Abbott

17:36 Support #1161: Split NumTheory.C into several smaller files: Some suggestions for the subfiles:
primes, RatReconstruct, CRT
Note that there is no need to split the docume... John Abbott
17:35 Support #1161 (Closed): Split NumTheory.C into several smaller files: The file *@NumTheory.C@* has become rather large. Consider splitting it into several smaller files.
John Abbott
17:33 Support #1160 (New): Clean up revised RatReconstructByContFrac code: The @RatReconstructByContFrac@ code has been revised (made better?), but there is lots of commented out cruft.
Cle... John Abbott
16:55 Design #1159: Add global enum "verify/DontVerify": John says it might be better to have a "reliability flag", which takes "0" as "certain".
I am thinking of the functi... Anna Maria Bigatti
16:32 Design #1159 (Closed): Add global enum "verify/DontVerify": should we have a global enum for "verify/DontVerify"?
this could be a pervasive feature in cocoalib.
Anna Maria Bigatti
16:27 Feature #1158 (Closed): New function: MinPolyQuotHeuristic --> MinPolyQuot with VerificationLevel: In our applications a wrong Minimal polinomial can be detected (we proved that if wrong, the degree is too low).
So ... Anna Maria Bigatti

Also available in: Atom

CoCoA » CoCoALib