Re: singular library resjung.lib
Burcin Erocal
On Sun, 27 Mar 2011 13:04:39 +0200
philipp...@business.uzh.ch wrote:
> Dear Burcin, I have made my library compatible with the current
> Singular version and followed the suggestions in your mail.
I submitted the new version of the library to the Singular repository
just now.
> However I
> have discovered that the normalization routine of Singular apparently
> does no longer work for a certain problem I managed to solve before.
> I did the following LIB"resjung.lib"; ring r = 0,(x,y,z),dp; ideal I
> = (x2-y3)*(z3-x3); list l = jungfib(I); def r3 = l[3]; setring r3;
> list nor = normal(QIdeal); The normalization of this ideal (which is
> just one polynomial!) runs forever. If I use the option "prim" or
> the previous version of normal.lib (I tested it just now) the
> computation is almost instantaneous.
Do you mean normal(QIdeal) returned instantly in an older version?
I tried both on version 3.1.3 and 3.1.1. Both didn't return a result
after 8 minutes. The option "prim" gave instant results for both.
To make it easier to copy paste, here is QIdeal:
ring r = 0,(x,x(1),y(0)),dp;
ideal I = x(1)^15*y(0)^6-x^2*x(1)^9*y(0)^3-x^3*x(1)^6*y(0)^3+x^5;
Burcin
> The reference in BibTex:
> @misc{Popescu,
>
> author = "Popescu-Pampu, Patric",
>
> title = "Introduction to Jung's Method of Resolution of
> Singularities",
>
> note = "arXiv:math.CV/0703353v1",
>
> year = 2007}
> Concerning the last item: Sadly I do not have sufficient time to
> provide a test file. Best regards, Philipp
>
> ------------------------------------------------------------------------
>
> Philipp Renner
> Institut für Betriebswirtschaftslehre
> Lehrstuhl für Quantitative Betriebswirtschaftslehre
> 8044 Zuerich
>
> Tel.: +41 44 634 37 82
> Raum: 107
>
> -----Burcin Erocal <bur...@mathematik.uni-kl.de> schrieb: ----- An:
> philipp...@web.de, an...@math.uni-hannover.de Von: Burcin Erocal
> <bur...@mathematik.uni-kl.de> Datum: 18.03.2011 18:02
> Kopie: sing...@mathematik.uni-kl.de
> Betreff: singular library resjung.lib
>
> Dear Singular experimental library author,
>
> we are in the process of preparing for a new stable Singular release.
> In order to include a more recent version of your library in the next
> release, please send us the new files by March 31st. If you want your
> experimental library to be promoted to a standard library, please
> check again that it complies with the instructions in the manual [1].
> In particular, your library should have
>
> 1- documentation and examples for each function exposed to users [2],
> 2- references to literature for the algorithms implemented if possible
> 3- tests, apart from the examples provided in the documentation,
> to check correctness of all code paths. These tests will be
> included in the Singular test suite to verify your library works as
> intended on all platforms supported by Singular in future releases.
>
> Instruction on how to provide tests for libraries can be found
> here:
>
> http://singular-team.blogspot.com/2011/03/testing-contributed-libraries.html
>
> [1] http://www.singular.uni-kl.de/Manual/latest/sing_48.htm#SEC85
> [2] http://www.singular.uni-kl.de/Manual/latest/sing_61.htm#SEC98
>
> The library checker [3] can be used to check some of these issues.
>
> [3]
> http://www.singular.uni-kl.de/index.php/new-libraries/formal-library-checker.html
>
>
> For resjung.lib, we have the following suggestions:
> * function names should be more explicit:
> clocus -> branchlocus
> embR -> embresolve
> * jungresolve: Keyword arguments can be used instead of the integer
> argument to specify if the input ideal is in noether position.
> perhaps "is_npos"=1 or 0
> The default should be 0, i.e., not in noether position.
>
>
> Best regards,
> Burcin
Grischa Studzinski
Burcin Erocal
On Fri, 8 Apr 2011 20:03:47 +0200
philipp...@business.uzh.ch wrote:
> the older version of the normal.lib was
> version="$Id: normal.lib,v 1.49 2006/07/25 17:54:28 Singular Exp $";
> I think it was in Singular version 3.0.4.
That's exactly the information we needed. Thanks!
> The reason I mentioned it was that I think this might be a bug,
> especially considering that this is a radical principal ideal in 3
> variables. However I do not know how the new version of normal.lib
> works so this might be simply due to the particulars of the problem.
I don't have access to Singular 3.0.4 right now to check this. Perhaps
someone more familiar with normal.lib can do this and open a ticket if
necessary. (The example is at the end of the quoted message below.)
Here is a list of changes to normal.lib since your version:
9662 *hannes: Michael C. (2006-07-25 by Singular)
11910 *hannes: new normal (2009-03-26 by Singular)
11954 removed some docu errors prior to release 3-1-0 (2009-04-07 by seelisch)
11987 *laplagne: new start. in normal (2009-04-10 by Singular)
12030 minor textual changes prior to 3-1-0 (24 months ago by seelisch)
12162 *santiago: new version (22 months ago by Singular)
12171 slaplagne: (22 months ago by Singular)
12395 *hannes: typo in doc. (19 months ago by Singular)
12410 Small bug fixed, reported by Leonard, when splitting the ideal by a zerodivisor. (19 months ago by laplagne)
12411 new version from Gert-Martin, as of Sept 6, 2009 (19 months ago by seelisch)
12414 *hannes: format (19 months ago by Singular)
12548 Adding Id property to all files (17 months ago by hannes)
12858 genus only for char 0 (14 months ago by hannes)
12864 syntax fix (14 months ago by hannes)
12888 genus for char p (13 months ago by hannes)
12918 minor typos in proc docs (13 months ago by seelisch)
12932 genus: typo fixed (13 months ago by hannes)
12933 genus: example (13 months ago by hannes)
12935 genus for larger char.: use normal instead of normalP (13 months ago by hannes)
13121 format (11 months ago by hannes)
13132 extended version for intBasis.lib (10 months ago by hannes)
13286 pfister: genus for char p (9 months ago by hannes)
13635 monomial.lib ->monomialideal.lib (6 months ago by hannes)
13670 new LIB versions and debugging in code for matrix minors (6 months ago by seelisch)
13695 recovered unix linefeeds (6 months ago by seelisch)
13696 locAtZero was removed by error from proc normalizationPrimes. It is set back. (6 months ago by laplagne)
13734 tabs and spaces (6 months ago by hannes)
13736 format (6 months ago by hannes)
13752 Fixed bug in positive characteristic. (6 months ago by laplagne)
13764 fix paraplanecurves.lib ex. 7 (6 months ago by hannes)
13814 format (5 months ago by hannes)
13818 fix counting of singularities (5 months ago by hannes)
13819 fix counting of singularities (5 months ago by hannes)
13924 fix hadling of options (5 months ago by hannes)
14376 duplicated w_deg in paraplanecurves.lib (need it there too but do
not want it to be non-static) (2 weeks ago by seelisch)
Can we identify which changeset caused the problem?
Burcin
> -----Burcin Erocal <bur...@mathematik.uni-kl.de> schrieb: -----
> An: philipp...@business.uzh.ch
> Von: Burcin Erocal <bur...@mathematik.uni-kl.de>
> Datum: 08.04.2011 19:07
> Kopie: singula...@googlegroups.com
> Betreff: Re: singular library resjung.lib
Burcin Erocal
philipp...@business.uzh.ch wrote:
> I checked the dates and two changes are a canditate for my problem
> 11910 *hannes: new normal (2009-03-26 by Singular)
> 11987 *laplagne: new start. in normal (2009-04-10 by Singular)
> The version from 2006-07-25 (1.49) is the one that still works and the one from 2009/04/15 (1.53) no longer works. I could not find the inbetween versions on the homepage to further narrow it down.
This is really helpful. Thank you.
Here is a part of the difference between those two versions:
@@ -43,133 +43,126 @@
"USAGE: normal(id [,choose]); id = radical ideal, choose = optional string. @*
Optional parameters in list choose (can be entered in any order):@*
Decomposition:@*
- - \"prim\" -> computes first the minimal associated primes, and then
- the normalization of each prime. (default)@*
- \"equidim\" -> computes first an equidimensional decomposition,
- and then the normalization of each component. @*
- - \"noDeco\" -> no preliminary decomposition is done. If the ideal is
+ and then the normalization of each component. (default)@*
+ - \"prim\" -> computes first the minimal associated primes, and then
+ the normalization of each prime. @*
+ - \"noDeco\" -> no preliminary decomposition is done. If the ideal is
not equidimensional radical, output might be wrong.@*
- - \"isPrim\" -> assumes that the ideal is prime. If assumption does
+ - \"isPrim\" -> assumes that the ideal is prime. If assumption does
not hold, output might be wrong.@*
- - \"noFac\" -> factorization is avoided in the computation of the
+ - \"noFac\" -> factorization is avoided in the computation of the
As you can see "prim" used to be the default. This changes it to
"equidim".
Burcin