#10063: Some determinants can not be computed
--------------------------------------------+-------------------------------
Reporter: tmonteil | Owner: tmonteil
Type: defect | Status: needs_work
Priority: critical | Milestone:
Component: commutative algebra | Keywords: determinant, ring,
ideal
Author: Thierry Monteil | Upstream: N/A
Reviewer: Mike Hansen, Sébastien Labbé | Merged:
Work_issues: |
--------------------------------------------+-------------------------------
Comment(by tmonteil):
Hi Sebastien,
thanks for your comments. As you know, i am not a computer scientist nor a
programmer, so i do not know much about design principles, but if i had to
code a determinant function from scratch, knowing that:
* the determinant is an fundamental function
* there exist a division-free algorithm that works in any case
* there are some fast/fancy/pretty/lounge/fieldic... optimizations that
are very good in some particular cases but not completely finished, i will
have written a code like (even ''before'' meeting any bug):
{{{
try:
all the kind of optimized algorithms
except StandardError:
pass #or send an automatic bug report if the error seems new.
the division-free algorithm that always works
}}}
I do not feel it is a dirty code, since its structure shows that one
algorithm works everywhere and the other are fragile optimizations.
Making patches that only repair the bugs found will make a code whose
architecture depend on the history of the bugs discovery, not necessarily
a readable code (why is there a test for the `is_field` method and not for
`is_exact`,...).
The aim of the Python convention is to avoid masking bugs, but if the
exception is `NotImplementedError`, then the "bug" is already known, since
it means that someone wrote an empty method which raises this
`NotImplementedError`.
Note also that the Python convention is a bit slower since i have to test
the `is_field` method even if the algorithm is not `None`. Should i
replace:
{{{
try:
R_is_field_attempt = R.is_field()
except NotImplementedError:
R_is_field_attempt = False
}}}
by:
{{{
if algorithm is None:
try:
R_is_field_attempt = R.is_field()
except NotImplementedError:
R_is_field_attempt = False
}}}
in order to skip the test when another algorithm is is called?
Anyway, here is a patch that implements your recommendations (i rename it
to keep the previous version on the trac server and follow the
discussion).
I fixed the problem of the documentation as well (i took the opportunity
to fix the existing test, which i copied).
Should we think to clean the whole `determinant` code?
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/10063#comment:10>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en.
