วันอังคารที่ 30 พฤศจิกายน พ.ศ. 2553

Re: [sage-devel] Re: Searching OEIS is broken

On Tue, Nov 30, 2010 at 7:10 PM, William Stein <wstein@gmail.com> wrote:
> On Thu, Nov 25, 2010 at 9:02 PM, William Stein <wstein@gmail.com> wrote:
>> I think the right thing to do in this case is to carefully understand
>> the intention of the license (possibly consult with a lawyer), and
>> also to possibly write a letter to the OEIS foundation with a petition
>> for them to reconsider their license. [...] if
>> we just take a breath, carefully understand the situation, and
>> possibly get a lot of momentum...
>
> (Something will appear on the number theory list soon about this.)
> I've just also got some offline emails from Neil Sloane, and we (the
> Sage project) definitely have their permission to distribute the OEIS
> file.   This is a something they are letting *us* do, but not
> necessarily any other projects.

Erratum: It has been pointed out to me that in fact I misinterpreted
Neil's email. English is ambiguous. In fact there is no permission
to redistribute a copy of the sequences from OEIS.

I'm very disturbed by this whole situation. It disturbs me that
fundamental data contributed for free by the community is locked down
like this. Projects like Wikipedia [3] (or the human genome project
[1]) are leaders that show that it is not necessary for
user-contributed content to be restricted like this.

This could be a good topic for a letter to the Notices of the American
Math Society (like [2]). If anybody reading this is also concerned,
please contact me and we can think this through.

-- William

[1] "It is therefore NHGRI s intent that human genomic DNA sequence
data, generated by the projects funded under RFA HG-95-005, should be
released as rapidly as possible and placed in the public domain where
it will be freely available." http://www.genome.gov/10000926

[2] Look here in the upper right: http://www.ams.org/notices/200710

[3] Wikipedia is Creative Commons licensed:
http://en.wikipedia.org/wiki/Wikipedia:About#Trademarks_and_copyrights

--
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Searching OEIS is broken

On Thu, Nov 25, 2010 at 9:02 PM, William Stein <wstein@gmail.com> wrote:
> I think the right thing to do in this case is to carefully understand
> the intention of the license (possibly consult with a lawyer), and
> also to possibly write a letter to the OEIS foundation with a petition
> for them to reconsider their license. [...] if
> we just take a breath, carefully understand the situation, and
> possibly get a lot of momentum...

(Something will appear on the number theory list soon about this.)
I've just also got some offline emails from Neil Sloane, and we (the
Sage project) definitely have their permission to distribute the OEIS
file. This is a something they are letting *us* do, but not
necessarily any other projects.

Their main concern now is figuring out how to rebuild the file
automatically, so they can make up to date versions of the file
available to us on a regular basis. It will be very important for
us to find a way to keep the optional Sloane table spkg up to date.
(We could possibly setup a cron job, and also something in the
sloane_sequence function, maybe, that detects if the downloaded
database is too old, and warns the user.)

If people have ideas about how to address the technically issues
mentioned above, do share.

-- William

William

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

Thanks to everyone for their suggestions! I'll try to implement (or
start implementing) them soon.

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On Tue, Nov 30, 2010 at 4:23 AM, David Kirkby <david.kirkby@onetel.net> wrote:
> On 30 November 2010 10:48, Jeroen Demeyer <jdemeyer@cage.ugent.be> wrote:
>> On 2010-11-30 00:55, Volker Braun wrote:
>>> nobody wants to give a positive review until it has been tried on all
>>> platforms
>>
>> I think "testing on every platform" is not a necessary condition for a
>> ticket to receive positive_review.  For reviewing a ticket, I think that
>> looking at the code is far more important than testing.  If the ticket
>> gets positive_review, it will certainly get tested and any problems will
>> be caught.

+1. I have the feeling that people are doing more testing than reading
of code, which is omitting the most important step, and in particular
the one that only a human can do. Testing should happen orthogonal to
someone reading the code and giving it a positive review.

>> Jeroen.
>
> IMHO, it should not be too hard for someone to test on one Linux
> system, one OS X system and one Solaris system. William has the
> hardware and will create accounts for people if they need to test
> Sage.

But it is too high of a burden for someone who doesn't enjoy patching
and building Sage on a variety of systems.

> Otherwise I fear too many positively reviewed tickets will end up
> getting bounced out for reasons that a bit of testing would have shown
> up much earlier. Of course, it also depends on the type of patch, and
> how large it is. Some are trivial. But certainly in the case of
> updating .spkg's, IMHO it's better if there's confirmation it wiorks
> on at least those three systems before giving a positive review.

Spkgs are a bit messier, because they have more platform independent
code, but for most tickets I don't think that's needed at all.

> Testing on every Linux system is a different matter altogether though,
> and unfortunately it does seem that some tickets are getting stuck.
> because they build on one distribution, but not on another.
> Unfortunately, to make matters worst, there are no buildbots for some
> of the systems where Sage is breaking on. To be honest, I feel the
> only answer to that is that if someone wants Sage to build on a
> specific version of an OS, they need to make a buildbot available.

+1

- Robert

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] easy review #7644: power series reversion

If you're looking for an interesting but easy review, check out

http://trac.sagemath.org/sage_trac/ticket/7644

The patch implements power series reversion (inverse under
composition) with under 20 lines of code (plus ~60 lines of
docstring), using Lagrange inversion. If you need to remind yourself
about Lagrange inversion, you can check the Lagrange–Bürmann section
of the Wikipedia article:

http://en.wikipedia.org/wiki/Lagrange_inversion_theorem#Lagrange.E2.80.93B.C3.BCrmann_formula


I've just verified that the patch applies cleanly to sage
4.6.1.alpha2, and passes all doctests.

best,
Niles

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Using Intel MKL

It would be interesting to compare Intel MKL vs. AMD ACML vs. threaded
ATLAS.

The most useful optimization right now is probably to enable threading
in ATLAS. None of the vendor libraries are free to distribute.

While on the topic of ATLAS, let me plug my spkg at
http://trac.sagemath.org/sage_trac/ticket/10226 which should help with
the recurring build issues ;-)


On Nov 30, 1:21 pm, Mag Gam <magaw...@gmail.com> wrote:
> Has anyone compiled SAGE with Intel MKL ?
>
> Reference:http://www.scipy.org/Installing_SciPy/Linux#head-7ce43956a69ec51c6f2c...http://software.intel.com/en-us/articles/intel-mkl/

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Using Intel MKL

(Please ignore my original response to this thread, which I removed. I
misread the simple one line question!)

I found the following by googling:

http://www.enthought.com/epd/mkl/

A developer once told me that MKL was worse than ATLAS. Perhaps it
depends on the platform? Either that or Intel have really improved it
in the last few years.

Sad that MKL is only "free for personal use", not "Open and Free
Period".

Sorry I didn't really answer your question.

Bill.


On Nov 30, 12:21 pm, Mag Gam <magaw...@gmail.com> wrote:
> Has anyone compiled SAGE with Intel MKL ?
>
> Reference:http://www.scipy.org/Installing_SciPy/Linux#head-7ce43956a69ec51c6f2c...http://software.intel.com/en-us/articles/intel-mkl/

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Using Intel MKL

Sage is not a BLAS/Lapack. It is a large distribution of Open Source
mathematical software (amongst other things). It does however include
a BLAS.

Do you mean has anyone compared Sage linked to ATLAS, etc. with Sage
linked to MKL?

I just googled ATLAS vs MKL and got this page, which seems to say MKL
was faster on at least that platform:

http://www.enthought.com/epd/mkl/

This disagrees with what a developer once told me (that MKL was
slower). Either he was wrong, or it depends what platform you are
using, or Intel have improved MKL considerably since then.

At any rate, the comparison is not so interesting, as MKL is only
"free for personal use", not "open and free PERIOD".

Bill.

On Nov 30, 12:21 pm, Mag Gam <magaw...@gmail.com> wrote:
> Has anyone compiled SAGE with Intel MKL ?
>
> Reference:http://www.scipy.org/Installing_SciPy/Linux#head-7ce43956a69ec51c6f2c...http://software.intel.com/en-us/articles/intel-mkl/

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Buildbot - does not seem to get much use

Its true that backward compatibility is not one of the goals of
Fedora, but rather testing new AND existing code. But the two problems
that prevent Sage from building on Fedora 14 affect everyone: mpir has
a security issue and ecl/maxima contain invalid code that just happens
to compile on older gcc's. The new Fedora release only uncovered
these bugs. Don't shoot the messenger :-)

On Nov 30, 2:01 pm, David Kirkby <david.kir...@onetel.net> wrote:
> On 30 November 2010 09:41, Volker Braun <vbraun.n...@gmail.com> wrote:
>
> > Meanwhile, Sage still fails to build on Fedora, one of the most
> > popular linux distributions, because the ecl/maxima update is stuck in
> > limbo (http://trac.sagemath.org/sage_trac/ticket/10187).
>
> That's not entirely true. The latest "stabe" Sage is 4.6, which was
> released on the 31st of October (30 days ago).
>
> Sage fails to build on the latest Fedora (version 14), which was
> released on the 4th of November - i.e. after the latest Sage.
>
> Ultimately, Fedora is very much a test system for Redhat, in the same
> way that OpenSolaris was for Sun.,  A compromise has to be made
> between having the latest features, and having high stability.
> Backwards compatibility does not appear to be very good with Fedora.
>
> Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On 30 November 2010 09:41, Volker Braun <vbraun.name@gmail.com> wrote:

> Meanwhile, Sage still fails to build on Fedora, one of the most
> popular linux distributions, because the ecl/maxima update is stuck in
> limbo (http://trac.sagemath.org/sage_trac/ticket/10187).

That's not entirely true. The latest "stabe" Sage is 4.6, which was
released on the 31st of October (30 days ago).

Sage fails to build on the latest Fedora (version 14), which was
released on the 4th of November - i.e. after the latest Sage.

Ultimately, Fedora is very much a test system for Redhat, in the same
way that OpenSolaris was for Sun., A compromise has to be made
between having the latest features, and having high stability.
Backwards compatibility does not appear to be very good with Fedora.

Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On 30 November 2010 10:48, Jeroen Demeyer <jdemeyer@cage.ugent.be> wrote:
> On 2010-11-30 00:55, Volker Braun wrote:
>> nobody wants to give a positive review until it has been tried on all
>> platforms
>
> I think "testing on every platform" is not a necessary condition for a
> ticket to receive positive_review.  For reviewing a ticket, I think that
> looking at the code is far more important than testing.  If the ticket
> gets positive_review, it will certainly get tested and any problems will
> be caught.
>
> Jeroen.

IMHO, it should not be too hard for someone to test on one Linux
system, one OS X system and one Solaris system. William has the
hardware and will create accounts for people if they need to test
Sage.

Otherwise I fear too many positively reviewed tickets will end up
getting bounced out for reasons that a bit of testing would have shown
up much earlier. Of course, it also depends on the type of patch, and
how large it is. Some are trivial. But certainly in the case of
updating .spkg's, IMHO it's better if there's confirmation it wiorks
on at least those three systems before giving a positive review.

Testing on every Linux system is a different matter altogether though,
and unfortunately it does seem that some tickets are getting stuck.
because they build on one distribution, but not on another.
Unfortunately, to make matters worst, there are no buildbots for some
of the systems where Sage is breaking on. To be honest, I feel the
only answer to that is that if someone wants Sage to build on a
specific version of an OS, they need to make a buildbot available.

It normally seems to be Fedora and openSUSE which get problems when
people update their operating system.

Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Using Intel MKL

Has anyone compiled SAGE with Intel MKL ?

Reference:
http://www.scipy.org/Installing_SciPy/Linux#head-7ce43956a69ec51c6f2cedd894a4715d5bfff974
http://software.intel.com/en-us/articles/intel-mkl/

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On 2010-11-30 00:55, Volker Braun wrote:
> nobody wants to give a positive review until it has been tried on all
> platforms

I think "testing on every platform" is not a necessary condition for a
ticket to receive positive_review. For reviewing a ticket, I think that
looking at the code is far more important than testing. If the ticket
gets positive_review, it will certainly get tested and any problems will
be caught.

Jeroen.

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Buildbot - does not seem to get much use

Well fortunately there are exceptions to the rule, otherwise we would
never update any spkg :-)

Meanwhile, Sage still fails to build on Fedora, one of the most
popular linux distributions, because the ecl/maxima update is stuck in
limbo (http://trac.sagemath.org/sage_trac/ticket/10187).

On Nov 30, 9:50 am, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote:
> On 2010-11-30 00:55, Volker Braun wrote:> the release manager will not run in on the build bot
> > until it has been positively reviewed...
>
> False.  There are 4 spkgs in sage-4.6.1.alpha2 in the needs_review
> status (two of them now have a positive_review, cvxopt (#6456) and
> readline (#9523) remain to be reviewed).

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: #8972 was re-opened months ago, but the resolution is still "fixed"

On 30 Nov., 10:03, Minh Nguyen <nguyenmi...@gmail.com> wrote:
> Done.

Thank you!

Cheers,
Simon

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] #8972 was re-opened months ago, but the resolution is still "fixed"

Hi Simon,

On Tue, Nov 30, 2010 at 6:56 PM, Simon King <simon.king@uni-jena.de> wrote:
> Can one of the administrators please take care of it and remove the
> wrong resolution, so that it is officially open?

Done. Here's the procedure for reopening a closed ticket:

1. You have a closed ticket.
2. Reopen it so its status would be "new".

Do not follow this procedure:

1. You have a closed ticket.
2. You reopen it by directly setting its status to "needs work" or a
status other than "new".

--
Regards
Minh Van Nguyen

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On 2010-11-30 00:55, Volker Braun wrote:
> the release manager will not run in on the build bot
> until it has been positively reviewed...
False. There are 4 spkgs in sage-4.6.1.alpha2 in the needs_review
status (two of them now have a positive_review, cvxopt (#6456) and
readline (#9523) remain to be reviewed).

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: how to make pdf reference from a single file

Thanks John, I'll try that.

John

On Mon, Nov 29, 2010 at 11:50 PM, John H Palmieri
<jhpalmieri64@gmail.com> wrote:
> On Nov 29, 2:24 pm, Niles <nil...@gmail.com> wrote:
>> On Nov 29, 9:11 am, John Cremona <john.crem...@gmail.com> wrote:
>>
>> > Can anyone tell me how to make just part of the pdf reference manual,
>> > specifically the part from a single .py file?
>>
>> I'd like to know how to do this too, and also how to produce the html
>> version!
>
> This will work for files in the Sage library (although you could add
> them temporarily if they weren't there).  This is also kind of
> destructive, so I would suggest working in a clone.  But: edit the
> file
>
> SAGE_ROOT/devel/sage/doc/en/reference/index.rst
>
> so that it looks something like
>
>
> .. toctree::
>
>   sage/matrix/matrix0
>   sage/rings/integer
>   sage/graphs/graph_latex
>
>
>
>
> Of course, replace sage/matrix/matrix0, etc., with whatever files you
> want.  Then run "sage -docbuild reference html" or "sage -docbuild
> reference pdf".  You'll probably get lots of warnings, but the docs
> should look okay; they'll be stored in SAGE_ROOT/devel/sage/doc/
> output/. It will have some extra stuff in it, though: table of
> contents, indices, etc.  I don't know a way to get rid of them without
> just selecting a page-range from the pdf file.  Oh, you could also
> edit the latex file, SAGE_ROOT/devel/sage/doc/output/latex/en/
> reference/reference.tex.  Delete the stuff you don't want and then try
> running pdflatex on the file.
>
> --
> John
>
> --
> To post to this group, send an email to sage-devel@googlegroups.com
> To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/sage-devel
> URL: http://www.sagemath.org
>

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

วันจันทร์ที่ 29 พฤศจิกายน พ.ศ. 2553

[sage-devel] #8972 was re-opened months ago, but the resolution is still "fixed"

Hi!

The history of #8972:

Opened 7 months ago

It was merged into sage-4.4.4.alpha0 by Mike Hansen 6 months ago, the
resolution set to "fixed". But the patch was immediately backed out by
Mike Hansen, because (apparently in combination with other patches) it
created doctest failures.

However, the resolution was not changed.

Subsequently, I solved the doctest problems and found that sage-4.4.2
with the page passes "sage -testall". So, I changed it to "needs
review".

Since then, nothing happened. I guess the reason is that the
resolution is still "fixed", which is plainly wrong. Namely, the
initial problem addressed by the ticket persists:
sage: R.<x> = ZZ[[]]
sage: (1/x).parent()
Laurent Series Ring in x over Integer Ring
sage: (x/x).parent()
Power Series Ring in x over Integer Ring

I think both parents are wrong - it should instead be "Laurent Series
Ring in x over Rational Field".

Can one of the administrators please take care of it and remove the
wrong resolution, so that it is officially open? I doubt that the
patch would still apply to sage-4.6, but let's try...

Cheers,
Simon

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Joint Math Meetings Sage talks

I will give at least a small demo of my Sage package for the abelian
sandpile model in the session on "Laplacian growth":

http://www.ams.org/meetings/national/jmm/2125_program_mipgroab.html#title

--Dave

On Nov 29, 5:53 pm, kcrisman <kcris...@gmail.com> wrote:
> (Posted primarily to sage-edu, cc:ed to sage-devel):
>
> Once again the Joint Meetings have a fair amount of Sage activity in
> the talks.  This year, there are three I found in a very cursory
> search:
>
> http://www.ams.org/temp/mtgs-2125-1067-z1-1926.pdf(Don Krug, from the
> summer PREP workshop, talking about "Teaching Introductory Computer
> Programming to Mathematics Majors with SAGE.")http://www.ams.org/temp/mtgs-2125-1067-l1-972.pdf(Our friends from
> Sunkyunkwan University who have presented in the past, I think, on
> "Mobile Sage-Math for Linear Algebra and its Application.")http://www.ams.org/amsmtgs/2125_abstracts/1067-z1-1891.pdf(Gregory
> Bard from Fordham on "Exploring Game Theory with SAGE, the open-source
> competitor to Maple, Mathematica, Matlab and MAGMA.")
>
> I hope I can go to some of these, especially the last one, since I'll
> be teaching that this spring to some extent.  I find it particularly
> interesting that the last talk essentially has the mission statement
> in the title, though I'm not sure if I'm +1 or -1 on that.  Eventually
> we'd want it to be the case that no explanation was needed!
>
> Any other talks known that don't have 'sage' in the title?  Also, I
> wonder if similar conferences in Europe (or elsewhere) are starting to
> have such abstracts submitted...
>
> - kcrisman

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On Mon, Nov 29, 2010 at 3:55 PM, Volker Braun <vbraun.name@gmail.com> wrote:
> Being able to test a single spkg on the most recent alpha/beta/rc
> would be very useful and not produce much load. I take it that this in
> not possible right now, but could probably be done easily.
>
> I find the current spkg review process is essentially a catch-22:
> nobody wants to give a positive review until it has been tried on all
> platforms but the release manager will not run in on the build bot
> until it has been positively reviewed...

+1

I've got 90% of a framework in place for testing individual patches,
but it doesn't handle spkgs yet, which are usually the most helpful to
test out on a large number of systems.

- Robert

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Joint Math Meetings Sage talks

(Posted primarily to sage-edu, cc:ed to sage-devel):

Once again the Joint Meetings have a fair amount of Sage activity in
the talks. This year, there are three I found in a very cursory
search:

http://www.ams.org/temp/mtgs-2125-1067-z1-1926.pdf (Don Krug, from the
summer PREP workshop, talking about "Teaching Introductory Computer
Programming to Mathematics Majors with SAGE.")
http://www.ams.org/temp/mtgs-2125-1067-l1-972.pdf (Our friends from
Sunkyunkwan University who have presented in the past, I think, on
"Mobile Sage-Math for Linear Algebra and its Application.")
http://www.ams.org/amsmtgs/2125_abstracts/1067-z1-1891.pdf (Gregory
Bard from Fordham on "Exploring Game Theory with SAGE, the open-source
competitor to Maple, Mathematica, Matlab and MAGMA.")

I hope I can go to some of these, especially the last one, since I'll
be teaching that this spring to some extent. I find it particularly
interesting that the last talk essentially has the mission statement
in the title, though I'm not sure if I'm +1 or -1 on that. Eventually
we'd want it to be the case that no explanation was needed!

Any other talks known that don't have 'sage' in the title? Also, I
wonder if similar conferences in Europe (or elsewhere) are starting to
have such abstracts submitted...

- kcrisman

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: LaTeX font issue with notebook

Hi,

Thanks Rob and Reg. Turns out that I was missing texlive-latex-extra.
LaTeX cells render fine now with these packages installed:

$ dpkg -l|grep texlive|grep ^ii|awk '{print $2}'|sort
texlive-base
texlive-binaries
texlive-common
texlive-doc-base
texlive-extra-utils
texlive-fonts-extra
texlive-fonts-extra-doc
texlive-fonts-recommended
texlive-fonts-recommended-doc
texlive-font-utils
texlive-generic-recommended
texlive-latex-base
texlive-latex-base-doc
texlive-latex-extra
texlive-latex-extra-doc
texlive-latex-recommended
texlive-latex-recommended-doc
texlive-luatex
texlive-metapost
texlive-metapost-doc
texlive-pictures
texlive-pictures-doc
texlive-pstricks
texlive-pstricks-doc


--
Carlo Hamalainen
http://carlo-hamalainen.net

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Buildbot - does not seem to get much use

Being able to test a single spkg on the most recent alpha/beta/rc
would be very useful and not produce much load. I take it that this in
not possible right now, but could probably be done easily.

I find the current spkg review process is essentially a catch-22:
nobody wants to give a positive review until it has been tried on all
platforms but the release manager will not run in on the build bot
until it has been positively reviewed...

Volker

On Nov 30, 12:08 am, David Kirkby <david.kir...@onetel.net> wrote:
> On 29 November 2010 13:09, Volker Braun <vbraun.n...@gmail.com> wrote:
>
> > Hi Dave,
>
> > I would be interested in using the buildbot occasionally, but I don't
> > know how to start a new build. Did you document the necessary steps
> > somewhere?
>
> > Best wishes,
> > Volker
>
> I did not set the buildbot up, so have not documented it.
>
> But I think in general it would be a bad idea if every developer could
> run a test whenever they want. It would lead to too much duplication
> and too much load on machines.
>
> Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how to make pdf reference from a single file

On Nov 29, 2:24 pm, Niles <nil...@gmail.com> wrote:
> On Nov 29, 9:11 am, John Cremona <john.crem...@gmail.com> wrote:
>
> > Can anyone tell me how to make just part of the pdf reference manual,
> > specifically the part from a single .py file?
>
> I'd like to know how to do this too, and also how to produce the html
> version!

This will work for files in the Sage library (although you could add
them temporarily if they weren't there). This is also kind of
destructive, so I would suggest working in a clone. But: edit the
file

SAGE_ROOT/devel/sage/doc/en/reference/index.rst

so that it looks something like


.. toctree::

sage/matrix/matrix0
sage/rings/integer
sage/graphs/graph_latex


Of course, replace sage/matrix/matrix0, etc., with whatever files you
want. Then run "sage -docbuild reference html" or "sage -docbuild
reference pdf". You'll probably get lots of warnings, but the docs
should look okay; they'll be stored in SAGE_ROOT/devel/sage/doc/
output/. It will have some extra stuff in it, though: table of
contents, indices, etc. I don't know a way to get rid of them without
just selecting a page-range from the pdf file. Oh, you could also
edit the latex file, SAGE_ROOT/devel/sage/doc/output/latex/en/
reference/reference.tex. Delete the stuff you don't want and then try
running pdflatex on the file.

--
John

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Buildbot - does not seem to get much use

On 29 November 2010 13:09, Volker Braun <vbraun.name@gmail.com> wrote:
> Hi Dave,
>
> I would be interested in using the buildbot occasionally, but I don't
> know how to start a new build. Did you document the necessary steps
> somewhere?
>
> Best wishes,
> Volker

I did not set the buildbot up, so have not documented it.

But I think in general it would be a bad idea if every developer could
run a test whenever they want. It would lead to too much duplication
and too much load on machines.

Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Buildbot - does not seem to get much use

On 29 November 2010 14:27, Jeroen Demeyer <jdemeyer@cage.ugent.be> wrote:
> On 2010-11-29 13:59, David Kirkby wrote:
>> The addition of the buildbot to Sage should be a good thing. But it
>> seems to get very little use.

> Let me answer this from a release manager point-of-view.
>
> Currently, the buildbot is essentially only used to test new Sage
> (alpha) releases and I must say it is extremely for that.

You missed an adjective there - I assume you mean it is extremely
"good" for that, but correct me if I'm wrong.

> I don't know that much about the buildbot, but I believe it is not easy
> to allow for testing of different tarballs (like, user 1 wants to test
> tarball X and user 2 wants to test tarball Y).  Right now, the path is
> actually hardcoded and must be changed for each release.

I think allowing individual developers to test would be a bit of a
disaster. I suspect the load on the buildbots would be too high.

I think it needs to be either

1) Run on a continuous basis. That is fine if hosts are dedicated to
just be buildbots, but is probably not acceptable for most hosts. I
for one would not want my machine "hawk" building Sage 24/7 !! So I
doubt that is very practical - at least not for all hosts.

2) Release manager test regularly (seems the best solution to me). The
frequency would dictated by

* Need - there's no point testing if you have not changed anything.
* Speed of hosts
* Limits agreed with the individual administrators of machines.
* etc

> It would be good for me to test pre-releases more often while they are
> prepared.  I almost continuously build new Sage distributions, testing
> various tickets.  For example right now there is a sage-4.6.1.alpha3
> tarball which will of course not be the final one.

That's what I'd expect to see more. I actually downloaded one of your
alpha3's and it built + tested ok, but I think the buildbot should be
giving you this information on a more regular basis.

> I proposed the following to Mitesh Patel: to have the hardcoded path
> point to a location which is easily changed, there could be a symbolic
> link /home/buildbot pointing to some sage release.  This would still not
> allow users to easily test their tarballs, but it would certainly help
> the release manager to test tickets.
>
> Jeroen.

Something along that line seems logical.

Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how to make pdf reference from a single file

On Nov 29, 9:11 am, John Cremona <john.crem...@gmail.com> wrote:
> Can anyone tell me how to make just part of the pdf reference manual,
> specifically the part from a single .py file?

I'd like to know how to do this too, and also how to produce the html
version!

-Niles

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

This is now Trac 10355:

http://trac.sagemath.org/sage_trac/ticket/10355

I've uploaded a patch which simply deletes the definition of
__contains__ for multivariate polynomials, thus falling back to the
definition from Parent. Unfortunately, this causes some doctest
failures in schemes/generic/toric_variety.py and misc/trace.py. I'll
have to look into those later (unless someone else wants to jump
in!). In any case, I guess we should take any further discussion to
that ticket.

best,
Niles


On Nov 29, 12:54 pm, Niles <nil...@gmail.com> wrote:
> On Nov 29, 9:07 am, Simon King <simon.k...@uni-jena.de> wrote:
>
> > > sage: PolynomialRing(QQ,'a').gen() in PolynomialRing(QQ,'a,b')
> > > True
>
> > Yes. But note that there must be a pushout of the parents:
> >   sage: QQ['c','a']('a')==ZZ['b','a','d']('a')
> > is false, will be false, and should be false.
>
> agreed.
>
> > > "UNIQUENESS and IMMUTABILITY: In Sage there is exactly one single-
> > > variate polynomial ring over each base ring in each choice of
> > > variable, sparseness, and implementation. There is also exactly one
> > > multivariate polynomial ring over each base ring for each choice of
> > > names of variables and term order. ..."
>
> > I don't see how these two details are related, but anyway, I am +1 for
> > having standard containment tests for multivariate polynomial rings.
>
> Oh, I was thinking that there is a polynomial subring of QQ['a','b']
> generated by 'a', and that according to the uniqueness/immutability
> idea, this subring ought to be THE (dense) polynomial ring in 'a' over
> QQ.  This contradicts the description of the current (undesirable)
> containment for multivariate polynomial rings (as John mentioned).
> That docstring reads:
>
>         """
>         This definition of containment does not involve a natural
>         inclusion from rings with less variables into rings with more.
>         """
>
> Since the ticket will be reversing this behavior, I thought it should
> have some justification.
>
> -Niles

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

On Nov 29, 9:07 am, Simon King <simon.k...@uni-jena.de> wrote:

> > sage: PolynomialRing(QQ,'a').gen() in PolynomialRing(QQ,'a,b')
> > True
>
> Yes. But note that there must be a pushout of the parents:
>   sage: QQ['c','a']('a')==ZZ['b','a','d']('a')
> is false, will be false, and should be false.

agreed.

> > "UNIQUENESS and IMMUTABILITY: In Sage there is exactly one single-
> > variate polynomial ring over each base ring in each choice of
> > variable, sparseness, and implementation. There is also exactly one
> > multivariate polynomial ring over each base ring for each choice of
> > names of variables and term order. ..."
>
> I don't see how these two details are related, but anyway, I am +1 for
> having standard containment tests for multivariate polynomial rings.
>

Oh, I was thinking that there is a polynomial subring of QQ['a','b']
generated by 'a', and that according to the uniqueness/immutability
idea, this subring ought to be THE (dense) polynomial ring in 'a' over
QQ. This contradicts the description of the current (undesirable)
containment for multivariate polynomial rings (as John mentioned).
That docstring reads:

"""
This definition of containment does not involve a natural
inclusion from rings with less variables into rings with more.
"""

Since the ticket will be reversing this behavior, I thought it should
have some justification.

-Niles

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Buildbot - does not seem to get much use

On 2010-11-29 13:59, David Kirkby wrote:
> The addition of the buildbot to Sage should be a good thing. But it
> seems to get very little use. I know the Wireshark projects runs their
> buildbots 24/7. Any failures are emailed to the developers, along with
> a list of who has committed source since the last successful build. I
> appreciate that its not practical to do this on all machines (I for
> one would not appreciate if 'hawk' was constantly running Sage tests..
> But its 10 days since the buildbot was last run, yet in that time,
> there has clearly been tickets committed.

Let me answer this from a release manager point-of-view.

Currently, the buildbot is essentially only used to test new Sage
(alpha) releases and I must say it is extremely for that.

I don't know that much about the buildbot, but I believe it is not easy
to allow for testing of different tarballs (like, user 1 wants to test
tarball X and user 2 wants to test tarball Y). Right now, the path is
actually hardcoded and must be changed for each release.

It would be good for me to test pre-releases more often while they are
prepared. I almost continuously build new Sage distributions, testing
various tickets. For example right now there is a sage-4.6.1.alpha3
tarball which will of course not be the final one.

I proposed the following to Mitesh Patel: to have the hardcoded path
point to a location which is easily changed, there could be a symbolic
link /home/buildbot pointing to some sage release. This would still not
allow users to easily test their tarballs, but it would certainly help
the release manager to test tickets.

Jeroen.

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] how to make pdf reference from a single file

Can anyone tell me how to make just part of the pdf reference manual,
specifically the part from a single .py file?

The intention is to do this for two specific files and include the
result as an appendix to a PhD thesis (about to be submitted by one of
my students).

I know that I can take the full 6000+ page manual and extract a page
range from it; and can look at individual classes and modules using
browse_sage_doc, but these do not (I think) either give me pdf output
or allow to view an entire file (module).

Thanks,

John

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

Hi Niles,

On 29 Nov., 14:22, Niles <nil...@gmail.com> wrote:
> Thanks for the pointers.  If I'm understanding correctly, everyone is
> in favor of using the default containment inherited from Parent;

I, for one, am.

> so
> that we would have
>
> sage: 3 in PolynomialRing(QQ,'a,b')
> True
>
> sage: PolynomialRing(QQ,'a').gen() in PolynomialRing(QQ,'a,b')
> True

Yes. But note that there must be a pushout of the parents:
sage: QQ['c','a']('a')==ZZ['b','a','d']('a')
is false, will be false, and should be false.

> The proposed behavior would be consistent with the uniqueness and
> immutability of polynomial rings:
>
> http://www.sagemath.org/doc/reference/sage/rings/polynomial/polynomia...
>
> "UNIQUENESS and IMMUTABILITY: In Sage there is exactly one single-
> variate polynomial ring over each base ring in each choice of
> variable, sparseness, and implementation. There is also exactly one
> multivariate polynomial ring over each base ring for each choice of
> names of variables and term order. ..."

I don't see how these two details are related, but anyway, I am +1 for
having standard containment tests for multivariate polynomial rings.

Cheer,
Simon

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

Thanks for the pointers. If I'm understanding correctly, everyone is
in favor of using the default containment inherited from Parent; so
that we would have

sage: 3 in PolynomialRing(QQ,'a,b')
True

sage: PolynomialRing(QQ,'a').gen() in PolynomialRing(QQ,'a,b')
True

(currently both return False)

The proposed behavior would be consistent with the uniqueness and
immutability of polynomial rings:

http://www.sagemath.org/doc/reference/sage/rings/polynomial/polynomial_ring_constructor.html#sage.rings.polynomial.polynomial_ring_constructor.PolynomialRing

"UNIQUENESS and IMMUTABILITY: In Sage there is exactly one single-
variate polynomial ring over each base ring in each choice of
variable, sparseness, and implementation. There is also exactly one
multivariate polynomial ring over each base ring for each choice of
names of variables and term order. ..."

So I'll open a ticket for this later today.

-Niles

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Buildbot - does not seem to get much use

Hi Dave,

I would be interested in using the buildbot occasionally, but I don't
know how to start a new build. Did you document the necessary steps
somewhere?

Best wishes,
Volker


On Nov 29, 1:59 pm, David Kirkby <david.kir...@onetel.net> wrote:
> The addition of the buildbot to Sage should be a good thing. But it
> seems to get very little use. I know the Wireshark projects runs their
> buildbots 24/7. Any failures are emailed to the developers, along with
> a list of who has committed source since the last successful build. I
> appreciate that its not practical to do this on all machines (I for
> one would not appreciate if 'hawk' was constantly running Sage tests..
> But its 10 days since the buildbot was last run, yet in that time,
> there has clearly been tickets committed.
>
> Perhaps a list should be drawn up on how often its acceptable to run
> tests on hosts, and perhaps make better use of the resources.
>
> For the two machines I'm responsible for, I would say:
>
> t2 - no limits
> hawk - up to 3 builds per day.
>
> Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Buildbot - does not seem to get much use

The addition of the buildbot to Sage should be a good thing. But it
seems to get very little use. I know the Wireshark projects runs their
buildbots 24/7. Any failures are emailed to the developers, along with
a list of who has committed source since the last successful build. I
appreciate that its not practical to do this on all machines (I for
one would not appreciate if 'hawk' was constantly running Sage tests..
But its 10 days since the buildbot was last run, yet in that time,
there has clearly been tickets committed.

Perhaps a list should be drawn up on how often its acceptable to run
tests on hosts, and perhaps make better use of the resources.

For the two machines I'm responsible for, I would say:

t2 - no limits
hawk - up to 3 builds per day.

Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Ticket #10326 (change sage --help, clean up sage-sage)

On Mon, Nov 29, 2010 at 01:09, Ivan Andrus <darthandrus@gmail.com> wrote:
On Nov 28, 2010, at 7:51 PM, Jeroen Demeyer wrote:

> Hello all,
>
> While working on testing command line options to sage (#10300,
> positive_review), I had a look at SAGE_ROOT/local/bin/sage-sage and
> found many things which could be improved.  There is now a patch at
> #10326 (needs_review).
>
> In particular, I reorganized the help messages (sage --help and sage
> --advanced).  Since this is a high-visibility change, please have a look
> and say what you think.

Is there a reason you used e.g. -gap instead of --gap in the help?  AFAIU they both work and -- is the future.  I'm just thinking that while we are changing it we might as well change that too.  Or perhaps this change won't matter once trac #21 is merged.  That said I don't really know the status of trac #21, but it looks close.
http://trac.sagemath.org/sage_trac/ticket/21

-Ivan

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

other than the new functionality of #10300, and the constantly changing documentation, #21 is basically done - I know that some other command line option requests have also been rolled into #21 as well, so it would be nice to have that closed

 -- Andrew

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Ticket #10326 (change sage --help, clean up sage-sage)

On Nov 28, 2010, at 7:51 PM, Jeroen Demeyer wrote:

> Hello all,
>
> While working on testing command line options to sage (#10300,
> positive_review), I had a look at SAGE_ROOT/local/bin/sage-sage and
> found many things which could be improved. There is now a patch at
> #10326 (needs_review).
>
> In particular, I reorganized the help messages (sage --help and sage
> --advanced). Since this is a high-visibility change, please have a look
> and say what you think.

Is there a reason you used e.g. -gap instead of --gap in the help? AFAIU they both work and -- is the future. I'm just thinking that while we are changing it we might as well change that too. Or perhaps this change won't matter once trac #21 is merged. That said I don't really know the status of trac #21, but it looks close.
http://trac.sagemath.org/sage_trac/ticket/21

-Ivan

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: Unifying, if it is not too troublesome ...

> advantages are obvious, e.g. one will be able to get, say, all the
> known to DB regular graphs on 15 vertices of degree 6 and diameter
> 2...
> That's a standard DB query then. Now one would have to browse the
> source to answer this, it seems...

Well, my question about a large number of methods was intended to the
actual Graph classes, no to the generators.... This is where the
number is becoming a problem (around 250 methods), and we need to find
a way around that would feel both natural enough and wouldn't prevent
new users from working with this class...

Perhaps just a better-designed web-based help ? I already tried to
sort Sage's graph methods into several categories, but as usual I can
not really make these categories exclusive..... Some flow-related
methods are also matching-related, some flow problems are also
existence problems or planarity problems -- still haven't found the
way out :-/

Dima : the database examples I see on your notebook aren't exactly as
natural as typing graphs.<tab>, but of course they could be hidden
beneath a nice interface and the user would not have to deal with it
directly if he doesn't know how to
... But the problem is the same : with a database, you are still
reduced to know what you are looking for, otherwise you don't find it
(except if you are just looking for some special strongly regular
graphs whose parameters you already know). I would like to find a
answer which lets the user browse the graphs available :-/

Nathann

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

วันอาทิตย์ที่ 28 พฤศจิกายน พ.ศ. 2553

Re: [sage-devel] Re: NIST FOIA Re:DLMF (The "Digital Library of Mathematical Functions")

On Sun, Nov 28, 2010 at 3:26 PM, rjf <fateman@gmail.com> wrote:

> ...

I agree with RJF completely on this one.

It can happen, apparently. ;)

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Unifying, if it is not too troublesome ...

On Nov 29, 10:54 am, Dima Pasechnik <dimp...@gmail.com> wrote:
> > P.S. (if anybody has any smart idea about how to deal with classes
> > having *far too many* different methods, I am also very interested.
> > Something will have to be decided soon for the Graph class anyway...
> > :-/ )
>
> not that my reply qualifies as very smart, but IMHO it's high time
> Sage has a database of
> graphs, rather than a zillion of different methods in Graph...
>
> http://uw.sagenb.org/home/pub/40/

advantages are obvious, e.g. one will be able to get, say, all the
known to DB regular graphs on 15 vertices of degree 6 and diameter
2...
That's a standard DB query then. Now one would have to browse the
source to answer this, it seems...

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] ZODB: remove it from Sage?

On Sun, Nov 28, 2010 at 7:40 PM, Dr. David Kirkby
<david.kirkby@onetel.net> wrote:
> On 11/29/10 01:45 AM, William Stein wrote:
>>
>> Hi,
>>
>> I'm curious who uses ZODB in Sage.     It's a standard package.
>>
>> I don't think it is a dependency for anything except some number
>> theory databases I created.  Also, I think it doesn't work at all with
>> SageObjects, which means its of very limited use in the context of
>> Sage.   Does anybody actually use it?
>>
>> I opened two related tickets:
>>
>>    http://trac.sagemath.org/sage_trac/ticket/10352
>
> I get the
>
> ---------------------------------------
>  Trac detected an internal error:
>
> ProgrammingError: could not create temporary file
> "base/pgsql_tmp/pgsql_tmp28162.4": No space left on device
>
> --------------------------------------
>
>
> I was going to ask if you know the ticket where it was added.

It was added before trac or justifications.

> i.e. what
> justification there was given. Just out of interest really. From what you
> say, it seems a bit pointless. Perhaps move it to experimental.

It should either be standard in Sage, or definitely not a package at
all. There is no point making it a package if it isn't standard,
since it is trivial to easy_install it.

-- William

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] ZODB: remove it from Sage?

On 11/29/10 01:45 AM, William Stein wrote:
> Hi,
>
> I'm curious who uses ZODB in Sage. It's a standard package.
>
> I don't think it is a dependency for anything except some number
> theory databases I created. Also, I think it doesn't work at all with
> SageObjects, which means its of very limited use in the context of
> Sage. Does anybody actually use it?
>
> I opened two related tickets:
>
> http://trac.sagemath.org/sage_trac/ticket/10352

I get the

---------------------------------------
Trac detected an internal error:

ProgrammingError: could not create temporary file
"base/pgsql_tmp/pgsql_tmp28162.4": No space left on device

--------------------------------------


I was going to ask if you know the ticket where it was added. i.e. what
justification there was given. Just out of interest really. From what you say,
it seems a bit pointless. Perhaps move it to experimental.

Dave

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: Unifying, if it is not too troublesome ...

> P.S. (if anybody has any smart idea about how to deal with classes
> having *far too many* different methods, I am also very interested.
> Something will have to be decided soon for the Graph class anyway...
> :-/ )
not that my reply qualifies as very smart, but IMHO it's high time
Sage has a database of
graphs, rather than a zillion of different methods in Graph...

http://uw.sagenb.org/home/pub/40/

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] ZODB: remove it from Sage?

Hi,

I'm curious who uses ZODB in Sage. It's a standard package.

I don't think it is a dependency for anything except some number
theory databases I created. Also, I think it doesn't work at all with
SageObjects, which means its of very limited use in the context of
Sage. Does anybody actually use it?

I opened two related tickets:

http://trac.sagemath.org/sage_trac/ticket/10352
http://trac.sagemath.org/sage_trac/ticket/10353

Proposal: We remove it from Sage, after somebody (e.g., me)
transitions all the existing databases that use it to a different
format.

I think sqlite would be a good choice for those database. The
performance would even be better, at least, and the data in the
databases would be usable without Sage. E.g., it is stupid that
after installing the large optional cremona database package, we have:

sage: time EllipticCurve('30001a1')
CPU times: user 0.38 s, sys: 0.35 s, total: 0.73 s
Wall time: 2.03 s
Elliptic Curve defined by y^2 + y = x^3 + x^2 - 171*x - 6408 over Rational Field
sage: time EllipticCurve('30001a1')
CPU times: user 0.01 s, sys: 0.00 s, total: 0.01 s
Wall time: 0.01 s

That should be instant, but ZODB has a big startup cost when you open
the database and it loads some indexes. Using sqlite, the same thing
would be nearly instant.

I think even shelve (which uses dbm) would be better, but I'm too
nervous to actually use it, since shelve's backend is too platform
dependent for my taste.

Reasons to remove ZODB:

* It's test suite fails miserably in Sage (see trac 10352)
* It isn't the right solution for how we currently use it.
* It's spkg is weird/hard to update, due to all the easy_install dependencies
* Almost nobody uses it. (?)

Reasons to keep it:

* Several people actually use it. (I don't know.)
* I'm missing something and it is a dependency for some key parts of Sage.


I think these days using SQLAlchemy + sqlite (or postgresql) is
probably a better approach than ZODB.

If people want to use ZODB after we remove it, they could always do:

sage -sh
easy_install zodb3

Note that the ZODB spkg has a bunch of other ZOPE-related packages
inside of it, so removing it will remove some other cruft.

---

I wrote a little lecture (=Sage worksheet) about basic non-SQL
non-object databases in Sage, by the way, which ends we me concluding
ZODB is pretty useless in its current state for Sage:

http://uw.sagenb.org/home/pub/40


--
William Stein
Professor of Mathematics
University of Washington
http://wstein.org

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

On Nov 28, 3:14 pm, Mike Hansen <mhan...@gmail.com> wrote:
> On Sun, Nov 28, 2010 at 3:09 PM, John H Palmieri <jhpalmier...@gmail.com> wrote:
>
> > I think it's this code in sage/rings/polynomial/
> > multi_polynomial_ring_generic.pyx:
>
> >    def __contains__(self, x):
> >        """
> >        This definition of containment does not involve a natural
> >        inclusion from rings with less variables into rings with more.
> >        """
> >        try:
> >            return x.parent() == self
> >        except AttributeError:
> >            return False
>
> I think this code should be removed in favor of the default implementation.
>
> --Mike

That looks okay to me. I don't know why "This definition of
containment does not involve...", but if you delete the code, then

sage: A1.<a> = PolynomialRing(QQ)
sage: A2.<a,b> = PolynomialRing(QQ)
sage: A1(a) in A2
True

I don't know if the original implementers didn't want this sort of
containment, but deleting the code will change this behavior.

--
John

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: NIST FOIA Re:DLMF (The "Digital Library of Mathematical Functions")

On Nov 27, 11:58 am, Donald Alan Morrison <donmorri...@gmail.com>
wrote:

> Yes, I started a new thread because the topic had changed.  You are
> quite fond of pointing out topic changes on sage-polemic.

Ah, my search for prior references (to DLMF) somehow did not bring to
my attention the OEIS thread. I only spotted notes from 2008. But
now
I see.
>

>
> Please top post a reply, and we can have Dave chime in on that
> too. :-)

Sure. middle, too.

>
> > why don't you just look at the copyright info.
> > It says
>
> >  Bulk copying,reproduction, or redistribution in any form is not
> > permitted.
>
> >http://dlmf.nist.gov/about/notices#S4
>
> Because the central point that Robert Dodier made is that it should be
> in the public domain as it was funded with federal money (not sure if
> that's necessarily true);

It certainly was not the case for Macsyma, initially, which was
funded by federal money. Much of the work published in scientific
journals is funded by federal money, and yet journals do not
give away copies..

Now the situation
with NIST is not that it was federally funded, but that the work was
done
under the direction of a federal agency (with consultants presumably
doing work for hire.) Does that place the work in the public domain?
Probably not, but I am not a lawyer.

> but since I'm interested in such an
> eventuality, that's why I would want to ask them for a formal
> explanation on public record.  FOIA requests get answered by lawyers,
> not webmasters.

As I see it, FOIA makes documents available for inspection or copying.
Since NIST appears to have nothing much hidden wrt DLMF, my guess
is that FOIA will not be useful ... you will get copies of the pages
you
already have, including the page with the copyright on it.

(for $0.50/page.)

Now I suppose that there is a possibility that if you requested the
DLMF via
FOIA, then that copy would not be covered by copyright, regardless of
what
it says. That would be interesting. But wouldn't it be better to
just contact
Dan Lozier and ask him nicely for permission to redistribute DLMF?

I would personally object strongly if DLMF were restricted by GPL.
Of course you could get a copy of DLMF and restrict YOUR copy.


RJF

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

On 29 Nov., 00:09, John H Palmieri <jhpalmier...@gmail.com> wrote:
> I think it's this code in sage/rings/polynomial/
> multi_polynomial_ring_generic.pyx:
>
>     def __contains__(self, x):
>         """
>         This definition of containment does not involve a natural
>         inclusion from rings with less variables into rings with more.
>         """
>         try:
>             return x.parent() == self
>         except AttributeError:
>             return False

Is this really what we want? Simply testing equality of the parents? I
don't think that one should drop any coercion!

See Parent.__contains__, which is
def __contains__(self, x):
r"""
True if there is an element of self that is equal to x under
==, or if x is already an element of self. Also, True in
other
cases involving the Symbolic Ring, which is handled specially.

For many structures we test this by using :meth:`__call__` and
then testing equality between x and the result.

The Symbolic Ring is treated differently because it is
ultra-permissive about letting other rings coerce in, but
ultra-strict about doing comparisons.
...
"""
P = parent_c(x)
if P is self or P == self:
return True
try:
x2 = self(x)
EQ = (x2 == x)
if EQ is True:
return True
elif EQ is False:
return False
elif EQ:
return True
else:
from sage.symbolic.expression import is_Expression
if is_Expression(EQ): # if comparing gives an
Expression, then it must be an equation.
# We return *true* here, even though the equation
# EQ must have evaluated to False for us to get to
# this point. The reason is because... in practice
# SR is ultra-permissive about letting other rings
# coerce in, but ultra-strict about doing
# comparisons.
return True
return False
except (TypeError, ValueError):
return False


Cheers,
Simon

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Re: how does "in" work?

On Sun, Nov 28, 2010 at 3:09 PM, John H Palmieri <jhpalmieri64@gmail.com> wrote:
> I think it's this code in sage/rings/polynomial/
> multi_polynomial_ring_generic.pyx:
>
>    def __contains__(self, x):
>        """
>        This definition of containment does not involve a natural
>        inclusion from rings with less variables into rings with more.
>        """
>        try:
>            return x.parent() == self
>        except AttributeError:
>            return False

I think this code should be removed in favor of the default implementation.

--Mike

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

Hi Niles,

On 28 Nov., 23:49, Niles <nil...@gmail.com> wrote:
> I've noticed the following strange behavior (bug):
>
> sage: A.<a> = PolynomialRing(QQ)
> sage: 3 in A
> True
>
> sage: A2.<a,b> = PolynomialRing(QQ)
> sage: 3 in A2
> False
>
> So apparently "in" for multivariate polynomial rings is slightly
> broken, but I don't know where to look to fix it.  I did spend some
> time looking for documentation or previous email threads, but "in"
> makes for a tough search.

I remember that someone has explained to me that membership testing is
specified in Sage as
e in P <=> e==P(e)
where e==P(e) means in particular that no error is raised.

However, as your example demonstrates, that specification is not
always used.

IIRC, this default implementation of the __contains__ could be
inherited from the class sage.structure.parent.Parent

Cheers,
Simon

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: how does "in" work?

On Nov 28, 2:49 pm, Niles <nil...@gmail.com> wrote:
> I've noticed the following strange behavior (bug):
>
> sage: A.<a> = PolynomialRing(QQ)
> sage: 3 in A
> True
>
> sage: A2.<a,b> = PolynomialRing(QQ)
> sage: 3 in A2
> False
>
> So apparently "in" for multivariate polynomial rings is slightly
> broken, but I don't know where to look to fix it.

I think it's this code in sage/rings/polynomial/
multi_polynomial_ring_generic.pyx:

def __contains__(self, x):
"""
This definition of containment does not involve a natural
inclusion from rings with less variables into rings with more.
"""
try:
return x.parent() == self
except AttributeError:
return False

--
John

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] how does "in" work?

I've noticed the following strange behavior (bug):

sage: A.<a> = PolynomialRing(QQ)
sage: 3 in A
True

sage: A2.<a,b> = PolynomialRing(QQ)
sage: 3 in A2
False

So apparently "in" for multivariate polynomial rings is slightly
broken, but I don't know where to look to fix it. I did spend some
time looking for documentation or previous email threads, but "in"
makes for a tough search. (This seems to be called "membership
testing" sometimes, but I haven't found documentation for that
either.)

thanks,
Niles

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Ticket #10326 (change sage --help, clean up sage-sage)

Hello all,

While working on testing command line options to sage (#10300,
positive_review), I had a look at SAGE_ROOT/local/bin/sage-sage and
found many things which could be improved. There is now a patch at
#10326 (needs_review).

In particular, I reorganized the help messages (sage --help and sage
--advanced). Since this is a high-visibility change, please have a look
and say what you think.

Jeroen.

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Re: [sage-devel] Cython elements and categories

Hi David!

On Fri, Nov 26, 2010 at 08:19:22PM +0100, David Roe wrote:
> Xavier Caruso and I are trying to implement some methods on matrices by
> using the category code. I think some of the functions in matrix2.py
> should get moved to categories, but that's a topic for another day...
>
> For now, we're trying to implement smith_form for matrices over certain
> kinds of base rings by writing a new category for matrices over those base
> rings and defining smith_form in the ElementMethods. Unfortunately,
> matrix2.Matrix implements smith_form already, so if we want to call our
> new smith form code, we need to do something like
>
> return self.parent().category().element_class.smith_form(self)
>
> When we try this, Python complains:
> TypeError: unbound method smith_form() must be called with
> MatrixAlgebrasOverQuasiDVR.element_class instance as first argument (got
> Matrix_modn_dense instance instead).
>
> There must be a way to get around this, since we can implement other
> methods in the category's ElementMethods class that work correctly. Does
> anyone know how to deal with this case that the elements of the category
> don't actually inherit from element_class?

Did you try a super call? It could possibly work. See
e.g. Parent._an_element_ in which I introduced a similar trick for a
similar purpose. Oops, that made me realize that I did not test or
document this trick there. Shame on me. Luckily, it's used and tested
in e.g. ModulesWithBasis.

Cheers,
Nicolas
--
Nicolas M. Thiéry "Isil" <nthiery@users.sf.net>
http://Nicolas.Thiery.name/

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: #9940: need help with parents and comparisons (and additive abelian groups)

I got
G < H
to alternately evaluate to true/false, for maybe half dozen trials,
then it went to many true.
My GUESS is that it is either due to some"residue" in something
uninitialized,
or a bug - which could also be an uninitialized issue ...somewhere.


On Nov 28, 1:35 am, Simon King <simon.k...@uni-jena.de> wrote:
> Hi John,
>
> On 27 Nov., 22:37, John H Palmieri <jhpalmier...@gmail.com> wrote:
>
> > I'm completely confused about the following behavior:
>
> So am I. I only found an __eq__ method, but no __cmp__ or __richcmp__,
> even though there is a _richcmp and _richcmp_helper. Can you test if
> these _rich... methods are called?
>
> Cheers,
> Simon

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

[sage-devel] Re: LaTeX font issue with notebook

Your example works on my virtual box encased
UBUNTU 10.10, Sage 4.6.
Without knowing exactly which ones do what, I offer the list of mine;
dpkg -l|grep texlive|grep ^ii|awk '{print $2}'
texlive
texlive-base
texlive-binaries
texlive-common
texlive-doc-base
texlive-extra-utils
texlive-font-utils
texlive-fonts-recommended
texlive-fonts-recommended-doc
texlive-generic-recommended
texlive-latex-base
texlive-latex-base-doc
texlive-latex-extra
texlive-latex-extra-doc
texlive-latex-recommended
texlive-latex-recommended-doc
texlive-luatex
texlive-pictures
texlive-pictures-doc
texlive-pstricks
texlive-pstricks-doc


On Nov 27, 2:04 am, Carlo Hamalainen <carlo.hamalai...@gmail.com>
wrote:
> Hi,
>
> I can't render a LaTeX cell in the notebook. I am running Sage 4.6 on
> Ubuntu 10.10 64bit. I made a cell in the notebook that just says
>
> ------------------------------------------------------------------------------------------
> %latex
>
> boo
> ------------------------------------------------------------------------------------------
>
> When I try to evaluate the cell I get an error:
>
> ------------------------------------------------------------------------------------------
> File: t1enc.def 2005/09/27 v1.99g Standard LaTeX file
> LaTeX Font Info:    Redeclaring font encoding T1 on input line 43.
> )
> ! Font T1/cmr/m/n/10=ecrm1000 at 10.0pt not loadable: Metric (TFM) file not fou
> nd.
> <to be read again>
>                    relax
> l.100 \fontencoding\encodingdefault\selectfont
>
> I wasn't able to read the size data for this font,
> so I will ignore the font specification.
> [Wizards can fix TFM files using TFtoPL/PLtoTF.]
> You might try inserting a different font spec;
> e.g., type `I\font<same font id>=<substitute font name>'.
> ------------------------------------------------------------------------------------------
>
> Here is the full output:http://carlo-hamalainen.net/sage/full_output_ubuntu1010_sage4.6.txt
>
> And here are the installed texlive packages on my system:
>
> carlo@r500:~$ dpkg -l|grep texlive|grep ^ii|awk '{print $2}'
> texlive-base
> texlive-binaries
> texlive-common
> texlive-doc-base
> texlive-extra-utils
> texlive-font-utils
> texlive-fonts-extra
> texlive-fonts-extra-doc
> texlive-fonts-recommended
> texlive-fonts-recommended-doc
> texlive-generic-extra
> texlive-generic-recommended
> texlive-latex-base
> texlive-latex-base-doc
> texlive-luatex
> texlive-math-extra
> texlive-pstricks
> texlive-pstricks-doc
>
> Am I missing another texlive package? Or some other kind of preamble
> for my LaTeX cell?
>
> Thanks,
>
> --
> Carlo Hamalainenhttp://carlo-hamalainen.net

--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to sage-devel+unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org