reduce-algebra-developers — Discussion of development, administration and support for Reduce

[Reduce-algebra-developers] Homebrew tap for REDUCE [updated]

[Jeff J. <tr...@po...>]

Greetings,

The Homebrew tap for REDUCE is ready for wider testing.

It should work on all systems that Homebrew supports.  As of this writing:

  "macOS 11 (Big Sur) or higher is best and supported, 10.11
  (El Capitan) – 10.15 (Catalina) are unsupported but may work
  and 10.10 (Yosemite) and older will not run Homebrew at all."

The tap currently provides formulae for three packages:

"reduce", a package based on the most recently released stable snapshot.
"reduce-current", a package (usually) based on a more recent SVN commit.
"reduce-static", a special package used for building binary distributions.

See https://github.com/reduce-algebra/homebrew-reduce-algebra for details!

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

If you really want to try it right now, the following should "just work":

$ brew tap reduce-algebra/reduce-algebra
$ brew install reduce --verbose

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

I intend to (eventually) upstream the "reduce" formula to Homebrew-core.
The tap will remain available for "reduce-current" and "reduce-static".

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

I've not yet fully documented the "reduce-static" package, so I'll describe
it more here.  It builds a package with similar contents and layout to the
official packaging, but a few notable differences.  This package does not
include a copy of the REDUCE source code, but does include the HTML version
of the documentation (in addition to the usual PDF version).  This is done
to save some space - the resulting DMG file should be about 70MB or less.
I think it's easy enough to "svn checkout", but I'm open to feedback here.

I've also not yet created a README file for inclusion with these packages.

The binary distributions are fully statically linked.  This means packages
it builds have no external dependencies, and don't require Homebrew to run.

Although currently untested, it should work fine on Apple Silicon machines,
but the PSL distribution will be incomplete (and non-functional), which is
known and currently expected.  The documentation should be updated mention
this, or perhaps the packaging script should not include the PSL files when
running on an ARM64 Macintosh for now?

Packages are built supporting the oldest CPU supported by the OS you are
running, for maximum portability.  If you *really* want to, you can
override this by appending, for example, "--bottle-arch=haswell" to the
"brew install" command.  You can also try building the current SVN commit
instead of the last release, by appending the "--HEAD" argument.

To try the package builder for yourself:

$ brew tap reduce-algebra/reduce-algebra
$ brew install reduce-static --build-bottle --verbose
$ package-reduce.sh
Successfully created Reduce_6547-x86_64-mac_13_ventura-darwin22.3.0.dmg
You can now run "brew remove reduce-static".
$ brew remove reduce-static

The DMG file will be created in your $HOME directory.  To open it:
open $HOME/Reduce_6547-x86_64-mac_13_ventura-darwin22.3.0.dmg

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

While technically unsupported, it is possible to try all this without a full
Homebrew installation.  I've created a `localbrew.sh` script that makes it a
bit easier, inspired by Python's virtualenv.  This method is considerably
slower since many prerequisite packages (and their prerequisites) will need
to compiled from source code, but your system will remain unmolested:

$ cd ~
$ curl -fsSLO https://raw.githubusercontent.com/johnsonjh/localbrew/master/localbrew.sh
$ chmod a+x localbrew.sh
$ ./localbrew.sh
[localbrew] $ exit
$ ./localbrew.sh
[localbrew] $ brew tap reduce-algebra/reduce-algebra
[localbrew] $ brew install reduce-static --build-bottle --verbose
[localbrew] $ package-reduce.sh
Successfully created Reduce_6547-x86_64-mac_13_ventura-darwin22.3.0.dmg
You can now run "brew remove reduce-static".
[localbrew] $ brew remove reduce-static
[localbrew] $ exit

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

The DMG produced does not bundle gnuplot.  You probably already have it,
or you can install your preferred version of it, either manually, from
MacPorts, or with Homebrew by using "brew install gnuplot", etc.

The Homebrew version of gnuplot defaults to using Qt/X11 output, so it will
require XQuartz.  Supposedly, AquaTerm support is available by first doing
"brew install aquaterm --cask".  Then just install (or reinstall) gnuplot.
I've NOT tested AquaTerm support with the Homebrew's version of gnuplot!
There haven't been any AquaTerm updates or releases since 2013, but I know
a lot of people still depend on it, or prefer it to using Qt/X11 output.

Because there are so many variants of "gnuplot" being distributed and all
compiled with different default options, I decided it was best to not
include it at all, leaving it up to the user.  I believe this is the same
approach the upstream packaging takes as well.

I've verified the above instructions on my system, and produced:
https://www.dropbox.com/s/0s3b1f8qfrj08sc/Reduce_6547-x86_64-mac_13_ventura-darwin22.3.0.dmg

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

I'd especially like to hear about successes (or failures) from macOS 10.x
users or Apple Silicon system owners.  Any feedback is appreciated!

PS - It *might* be possible to build Universal packages with this system
in the future, but I haven't even started on it, mostly because I don't
have an Apple Silicon machine to test with.

--
Jeffrey H. Johnson
tr...@po...

Re: [Reduce-algebra-developers] macOS Homebrew REDUCE tap

[Jeff J. <tr...@po...>]

I've now updated the "reduce" formula* as well with the same changes that were made to "reduce-current" to (hopefully) fix M1 machines.  I'll get some sleep now and await any feedback.

* https://github.com/reduce-algebra/homebrew-reduce-algebra

--
Jeffrey H. Johnson
tr...@po...

Re: [Reduce-algebra-developers] macOS Homebrew REDUCE tap

[Jeff J. <tr...@po...>]

Greetings,

It has come to my attention that, unfortunately, the Homebrew tap was not yet working on M1 machines - just Intel.

I've pushed some changes which I believe should correct this - see https://sourceforge.net/p/reduce-algebra/discussion/899365/thread/78d7c6c9e3/?limit=25#ddd2/9a9d ...

At this time, only the "reduce-current" formula has been updated with the changes - the "reduce" formula is unchanged.

Once we have a successful build on M1, I'll apply the fixes to the "reduce" formula as well, but for now, M1/Apple Silicon users should use the "reduce-current" formula if they'd like to help testing.

Since I don't have an M1 machine, I'll wait for some feedback.

Thanks,

--
Jeffrey H. Johnson
tr...@po...

[Reduce-algebra-developers] macOS Homebrew REDUCE tap

[Jeff J. <tr...@po...>]

Greetings,

I have a Homebrew formula for macOS, which I eventually plan to upstream, but in the meantime I'd like to get further testing.

You can see details at:

https://github.com/reduce-algebra/homebrew-reduce-algebra

Currently, I have only tested on Intel machines, but it should work for M1/Apple Silicon as well - if not, that's a bug and I'll fix it.

In short, to try it, you'll Homebrew installed - see https://brew.sh/ for that, then:

$ brew tap reduce-algebra/reduce-algebra
$ brew install reduce

or, if you prefer to track SVN:

$ brew install reduce-current

M1/Apple Silicon users and users not on Ventura will need to build REDUCE from source.  This should happen automatically.  However, the way I recommend is:

$ time brew install reduce --verbose --display-times

This just gives you some extra output which might be helpful in case something goes wrong, and times how long things take. The source build can take a *long* time, and at some points it might seem like it's stalled or hung, but that's (almost always) not the case, and is expected.  You should expect it to take about 45 minutes or so. Possibly *MUCH* longer the first time, especially if brew needs to install and configure all bazillion TeXLive prerequisites (since I build the documentation from sources as well).

mini-FAQ:

The motivation for this was three-fold a) to build on a system without requiring MacPorts to be installed and used, b) to keep more updated prerequisites, and c) to have automatic updates. It should work on a system that *also* has MacPorts installed, but I've not tested this.

I know this works on Intel Ventura (and previous Monterey), but I've not tested any M1 machines.  I don't see why it wouldn't work.

I plan to eventually upstream the "reduce" formula, so the tap will not be necessary, but this will probably take awhile.  The tap will however remain available after the formula is upstreamed to provide the "reduce-current" package.

If anyone is successful with building M1 versions or other versions of macOS besides Ventura and wishes to contribute by providing the resulting compiled binaries as "bottles", please contact me and we can arrange this.  It would be appreciated.

If you want to use Run-REDUCE with the Homebrew package, you'll need to adjust some paths accordingly.  Francis provided me build instructions months ago, but I've been busy and haven't yet had time to create a Run-REDUCE formula to enable installation via brew, but I do plan to.  In the meantime, the necessary configuration should be self-explanatory, but I'll assist if it isn't.

Finally, if anything goes wrong, blame me directly, not anyone else.

--
Jeffrey H. Johnson
tr...@po...

[Reduce-algebra-developers] "RosettaVM" on a Mac-m1 (macbook air early version)

[Arthur N. <ac...@ca...>]

I have just been trying this (do a search for RosettaVM to understand 
more!)

I have put an aarch-Ubuntu on the Macbook and then as part the blurb about 
RosettaVM set it up so that via binfmt and Rosetta 2 it is able to execute
x86-64 code.
I have been able to build csl.img/x86_64 on that but at present when I try 
to make bootstrapreduce.img things crasj half way through. So I build a 
regulaar aarch64-linux version of Reduce there and copied an x86_64 chunk 
from the cslbuild directory of a different machine where that was native.
Then using the version of testall.sh that I have very recently checked in 
I went "scripts/testall.sh --noregressions --csl --csl=vubuntu" where 
cslbuild/vubuntu is where I have put the copied stuff.
The end of the output shows

Summary of test runs
csl                 52020 ms CPU        620 ms GC
csl:vubuntu         76930 ms CPU        1170 ms GC

so in broad terms the emulated x86_64 version takes about 1.5 times as 
long to run the tests. When (and if) I can coax the building to run in 
that world the build step will have even less overhead, since (cross) 
compilation of C++ code has no reason to be slower that compilation to the 
native architecture - the only slowdown is when Reduce is run to make 
image files etc.

For comparison if I run "scripts/testall.sh --csl" directly on the Mac (ie 
not in the arm-linux VM) I see (after installing macports "diffutils" and 
messing with scripts/test1/sh to use "gdiff" - I think that something 
there has changed since I had previously tested!)
csl                  46160 ms CPU       430 ms GC
so there is (as one would expect) a bit of overhead running the 
aarch64-Linux VM - but you might view that as fairly modest. And the
amazing take awas is that this may mean that the Mac-m1 can run (almost?) 
any linux-x86_64 software with an overhead that is made less obnoxious by 
the possibility that the m1 processor may be faster than the Intel/AMD one 
you are used to!

I will have a go at sorting out the "diff" issue tomorrow and sometime I 
need to understand why the attempt to fully build bootstrapreduce.img 
crashed. But maybe some of you will be interested in these numbers.

Arthur

Re: [Reduce-algebra-developers] Error building csl in reduce trunk

[Andrey G. G. <A.G...@in...>]

On Fri, 3 Feb 2023, Arthur Norman wrote:
> That was a transient error for some of yesterday and you clerarly dopwnloaded 
> at a time it was there and that I had left bad messed up code there, I 
> checked in a correction at around 6:20 (UK time) yesterday!
Thanks! 6511 compiles fine.

Andrey

Re: [Reduce-algebra-developers] Error building csl in reduce trunk

[Arthur N. <ac...@ca...>]

That was a transient error for some of yesterday and you clerarly 
dopwnloaded at a time it was there and that I had left bad messed up code 
there, I checked in a correction at around 6:20 (UK time) yesterday!
But thanks for the report. Arthur

On Fri, 3 Feb 2023, Andrey G. Grozin wrote:

> Hello *,
>
> I've downloaded reduce-algebra-master.zip from 
> https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Freduce-algebra%2Freduce-algebra&data=05%7C01%7Cacn1%40universityofcambridgecloud.onmicrosoft.com%7C68d6bf6ca6a64b2768d308db05b9ae6e%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638110067282106477%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=CvTkNk0%2BiyL4K4HYeZWDx4OjLy7eG7QGaZABfc4vIPI%3D&reserved=0 
> (the last commit is 398cf181d422697d646c6c754e92dd37a8aea07b, Feb 2, 2023, 
> Merge branch 'svn/trunk'), unzipped it, and run
>
> ./rebuild.sh
>
> From rebuild.log:
>
> [x86_64-unknown-gentoo2.13] CXX   csl-arith01.o
> In file included from 
> /home/grozin/reduce-algebra-master/csl/cslbase/headers.h:82,
>                 from 
> /home/grozin/reduce-algebra-master/csl/cslbase/arith01.cpp:41:
> /home/grozin/reduce-algebra-master/csl/cslbase/machine.h: In function 
> ‘int64_t ASR(int64_t, int)’:
> /home/grozin/reduce-algebra-master/csl/cslbase/machine.h:452:40: error: 
> expected ‘;’ before ‘)’ token
>  452 |     return a/static_cast<int64_t>(1)<<n);
>      |                                        ^
>      |                                        ;
> /home/grozin/reduce-algebra-master/csl/cslbase/machine.h:452:40: error: 
> expected primary-expression before ‘)’ token
> /home/grozin/reduce-algebra-master/csl/cslbase/machine.h: In function 
> ‘int128_t ASR(int128_t, int)’:
> /home/grozin/reduce-algebra-master/csl/cslbase/machine.h:457:41: error: 
> expected ‘;’ before ‘)’ token
>  457 |     return a/static_cast<int128_t>(1)<<n);
>      |                                         ^
>      |                                         ;
> /home/grozin/reduce-algebra-master/csl/cslbase/machine.h:457:41: error: 
> expected primary-expression before ‘)’ token
> gmake[2]: *** [Makefile:2668: csl-arith01.o] Error 1
> gmake[2]: Leaving directory 
> '/home/grozin/reduce-algebra-master/cslbuild/x86_64-unknown-gentoo2.13/csl'
>
> This is Gentoo linux, gcc-12.
>
> psl reduce builds successfully. But there is no redfront, because it is built 
> as a part of csl reduce.
>
> Andrey
>

[Reduce-algebra-developers] Error building csl in reduce trunk

[Andrey G. G. <A.G...@in...>]

Hello *,

I've downloaded reduce-algebra-master.zip from 
https://github.com/reduce-algebra/reduce-algebra (the last commit is 
398cf181d422697d646c6c754e92dd37a8aea07b, Feb 2, 2023, Merge branch 
'svn/trunk'), unzipped it, and run

./rebuild.sh

>From rebuild.log:

[x86_64-unknown-gentoo2.13] CXX   csl-arith01.o
In file included from 
/home/grozin/reduce-algebra-master/csl/cslbase/headers.h:82,
                  from 
/home/grozin/reduce-algebra-master/csl/cslbase/arith01.cpp:41:
/home/grozin/reduce-algebra-master/csl/cslbase/machine.h: In function 
‘int64_t ASR(int64_t, int)’:
/home/grozin/reduce-algebra-master/csl/cslbase/machine.h:452:40: error: 
expected ‘;’ before ‘)’ token
   452 |     return a/static_cast<int64_t>(1)<<n);
       |                                        ^
       |                                        ;
/home/grozin/reduce-algebra-master/csl/cslbase/machine.h:452:40: error: 
expected primary-expression before ‘)’ token
/home/grozin/reduce-algebra-master/csl/cslbase/machine.h: In function 
‘int128_t ASR(int128_t, int)’:
/home/grozin/reduce-algebra-master/csl/cslbase/machine.h:457:41: error: 
expected ‘;’ before ‘)’ token
   457 |     return a/static_cast<int128_t>(1)<<n);
       |                                         ^
       |                                         ;
/home/grozin/reduce-algebra-master/csl/cslbase/machine.h:457:41: error: 
expected primary-expression before ‘)’ token
gmake[2]: *** [Makefile:2668: csl-arith01.o] Error 1
gmake[2]: Leaving directory 
'/home/grozin/reduce-algebra-master/cslbuild/x86_64-unknown-gentoo2.13/csl'

This is Gentoo linux, gcc-12.

psl reduce builds successfully. But there is no redfront, because it is 
built as a part of csl reduce.

Andrey

[Reduce-algebra-developers] New release of REDUCE IDE for GNU Emacs

[Francis W. <f.j...@li...>]

REDUCE IDE version 1.10 is now available 
<https://reduce-algebra.sourceforge.io/reduce-ide/>. Here are some of 
the key updates:

  * Parse comment statements robustly starting from the top of the buffer.
  * Better support for |/**/| comments and indentation.
  * A new option |autoload-reduce-run| controls whether to autoload
    REDUCE Run mode, and if not whether to install an abbreviated Run
    REDUCE menu that can load REDUCE Run mode.
  * *Please note that the function |require-reduce-run| , intended to be
    added to reduce-mode-load-hook, is now deprecated and will be
    removed in the next release*; please use the |autoload-reduce-run|
    option instead.

Francis

[Reduce-algebra-developers] New release of REDUCE IDE for GNU Emacs

[Francis W. <f.j...@li...>]

REDUCE IDE version 1.9 is now available 
<https://reduce-algebra.sourceforge.io/reduce-ide/>. Here are some of 
the key updates:

  * Highlight fluid and global forms as variable type declarations
    rather than as quoted data.
  * Some menu improvements.
  * Indentation is now more reasonable.
  * *INCOMPATIBLE CHANGE:* The command |reduce-indent-line| is now
    idempotent and no longer successively indents. Hence, the option
    |reduce-indent-line-conservative| no longer exists.
  * The new command |reduce-indent-line-always| bound to |C-TAB| indents
    by one step, and replaces the effect of executing
    |reduce-indent-line| repeatedly.

Francis

Re: [Reduce-algebra-developers] Please unscubscribe me

[Francis W. <f.j...@li...>]

Mailing list management is self-service. If you click on the link at the 
bottom of the email, namely

https://lists.sourceforge.net/lists/listinfo/reduce-algebra-developers

there is a link to unsubscribe. You don't even need to log in to use this.

Francis

On 03/10/2022 11:37 pm, Gary Knight wrote:
> I am no longer following reduce-algebra development. Please 
> unsubscribe me;  thanks.
>  Gary Knight
>
>
> _______________________________________________
> Reduce-algebra-developers mailing list
> Red...@li...
> https://lists.sourceforge.net/lists/listinfo/reduce-algebra-developers

[Reduce-algebra-developers] Please unscubscribe me

[Gary K. <gdk...@gm...>]

I am no longer following reduce-algebra development. Please unsubscribe
me;  thanks.
 Gary Knight

[Reduce-algebra-developers] New release of REDUCE IDE for GNU Emacs

[Francis W. <f.j...@li...>]

REDUCE IDE version 1.8 is now available 
<https://reduce-algebra.sourceforge.io/reduce-ide/>. Here are some of 
the key updates:

  * Use lexical scoping, which might be a little faster.
  * Operations based on procedures now support |matrixproc|, |listproc|,
    and procedure type declarations. They report a user error if they fail.
  * Syntax Highlighting:
      o More robust highlighting of comment statements.
      o There are now three strictly inclusive levels: “Symbolic”
        includes “Algebraic” includes “Basic”.
      o Highlight group delimiters the same as block delimiters.
      o Highlight named constants such as Catalan.
      o Highlight symbolic-mode functions such as get and put as builtin
        functions.
      o Highlight lambda parameters the same as procedure parameters.

Francis

Re: [Reduce-algebra-developers] Accessing coefficients of a Taylor series

[Rainer S. <rai...@gm...>]

A new operator,

  taylorcoefflist

extracts a list of pairs (exponents, coefficient) from a Taylor kernel. If
there is only one variable exponents is a list of one element.


yy := taylor (e**y, y, 0, 4);


               1   2    1   3    1    4      5
yy := 1 + y + ---*y  + ---*y  + ----*y  + O(y )
               2        6        24


taylorcoefflist yy;


{{{0},1},

 {{1},1},

       1
 {{2},---},
       2

       1
 {{3},---},
       6

       1
 {{4},----}}
       24


To extract the coefficient for exponent 3 you may use the select operator, e.g.,

select(first(~w)={3},taylorcoefflist yy);


       1
{{{3},---}}
       6


Regards,
   Rainer




On Sat, 10 Sep 2022 at 19:36 +0200, Thomas Baruchel wrote:

 > Hi,
 >
 > what is the most reliable way of accessing coefficients of a power series
 > (built with the taylor() function)?
 >
 > I initially had some success with "part", but as soon as some coefficnets are
 > zero, the "part" function returns wrong terms.
 >
 > Also, I am only interested in the coefficient, not the variable; for insance,
 > for   1 + 2*x - 3*x^2 + O(x^3), I would like to get the -3 coefficient for
 > degree 2 rather than the whole -3*x^2 term.
 >
 > Thank you by advance,
 >
 > best regards,
 >
 > --
 > Thomas Baruchel
 >
 >
 > _______________________________________________
 > Reduce-algebra-developers mailing list
 > Red...@li...
 > https://lists.sourceforge.net/lists/listinfo/reduce-algebra-developers
 >

   Rainer Schöpf

[Reduce-algebra-developers] New release of REDUCE IDE for GNU Emacs

[Francis W. <f.j...@li...>]

REDUCE IDE version 1.7 is now available 
<https://reduce-algebra.sourceforge.io/reduce-ide/>. Here are some of 
the key updates:

  * *INCOMPATIBLE CHANGE:* The command |run-reduce| is now the only way
    to run REDUCE. It prompts for the REDUCE command name to run,
    defaulting to the last one used.
  * The commands |run-csl-reduce| and |run-psl-reduce| have been
    removed; use |run-reduce| and select the appropriate command name
    instead.
  * The customizable option |reduce-run-commands| now accepts arbitrary
    key strings and an arbitrary number of REDUCE versions.
  * |<RET>| in REDUCE Run mode now adds a |;| statement terminator
    automatically as appropriate (whereas |S-<RET>| never adds a
    terminator).
  * Initial support for C-style |/**/| comments: they are now
    highlighted as comments and movement commands should skip them.

Francis

Re: [Reduce-algebra-developers] Accessing coefficients of a Taylor series

[Alan B. <ala...@gm...>]

Arthur has already replied regarding accessing coefficients of a power 
series built with the Taylor package.

For series built with the TPS package, there is a slightly more direct 
way, for example:

load tps;
s := ps(sin x, x, 0);
psterm(s,9);

One may also use pstruncate to truncate a power series after a specified 
number of terms and then usecoeff orcoeffn to extract the coefficient(s) 
of the the desired term(s) as Arthur indicated.

Sections 16.1 & 16.2 of the manual give more info about the facilities 
of the two packages and give a brief discussion of the relative 
advantages/disadvantages of the two systems.

Alan Barnes

On 10/09/2022 18:36, Thomas Baruchel wrote:
> Hi,
>
> what is the most reliable way of accessing coefficients of a power 
> series (built with the taylor() function)?
>
> I initially had some success with "part", but as soon as some 
> coefficnets are zero, the "part" function returns wrong terms.
>
> Also, I am only interested in the coefficient, not the variable; for 
> insance, for   1 + 2*x - 3*x^2 + O(x^3), I would like to get the -3 
> coefficient for degree 2 rather than the whole -3*x^2 term.
>
> Thank you by advance,
>
> best regards,
>
> -- 
> Thomas Baruchel
>
>
> _______________________________________________
> Reduce-algebra-developers mailing list
> Red...@li...
> https://lists.sourceforge.net/lists/listinfo/reduce-algebra-developers

[Reduce-algebra-developers] CALI

[Donald S. <sit...@gm...>]

Anyone know any easy readings on package CALI?

Re: [Reduce-algebra-developers] Accessing coefficients of a Taylor series

[Arthur N. <ac...@ca...>]

One possibility might start

load_package taylor;
aa := taylor(tan x, x, 0, 7);
bb := taylortostandard aa;
coeff(bb, x);
which yields the list
         1     2      17
{0,1,0,---,0,----,0,-----}
         3     15     315
where part may do what you would like?

Arthur


On Sat, 10 Sep 2022, Thomas Baruchel wrote:

> Hi,
>
> what is the most reliable way of accessing coefficients of a power series 
> (built with the taylor() function)?
>
> I initially had some success with "part", but as soon as some coefficnets are 
> zero, the "part" function returns wrong terms.
>
> Also, I am only interested in the coefficient, not the variable; for insance, 
> for   1 + 2*x - 3*x^2 + O(x^3), I would like to get the -3 coefficient for 
> degree 2 rather than the whole -3*x^2 term.
>
> Thank you by advance,
>
> best regards,
>
> --
> Thomas Baruchel
>

[Reduce-algebra-developers] Accessing coefficients of a Taylor series

[Thomas B. <bar...@gm...>]

Hi,

what is the most reliable way of accessing coefficients of a power series (built with the taylor() function)?

I initially had some success with "part", but as soon as some coefficnets are zero, the "part" function returns wrong terms.

Also, I am only interested in the coefficient, not the variable; for insance, for   1 + 2*x - 3*x^2 + O(x^3), I would like to get the -3 coefficient for degree 2 rather than the whole -3*x^2 term.

Thank you by advance,

best regards,

--
Thomas Baruchel

[Reduce-algebra-developers] REDUCE IDE 1.6 for GNU Emacs

[Francis W. <f.j...@li...>]

REDUCE IDE is a GNU Emacs package that provides an *I*ntegrated 
*D*evelopment *E*nvironment for REDUCE consisting of major modes for 
editing REDUCE source code and running a *command-line version* of 
REDUCE in an Emacs window.

A new release is now available from 
https://reduce-algebra.sourceforge.io/reduce-ide/. REDUCE IDE version 
1.6 (June 2022) provides better syntactic highlighting (fontification) 
and better delimiter matching, which are provided by two new files. It 
also repairs support for running PSL REDUCE on Microsoft Windows. I have 
tested REDUCE IDE 1.6 with REDUCE revision 6339 and GNU Emacs 28 on 
Microsoft Windows and Ubuntu.

Francis

Re: [Reduce-algebra-developers] Reduce algebra: pls advise how to get an effectively 'random' seed for random number generator

[Prof T. R. <Pro...@pr...>]

Hi Arthur,

thanks,  "lisp random_new_seed systo_datestamp();"   works easily
and well,

Tony

On 6/6/2022 4:24 pm, Arthur Norman wrote:
> At one time I think we had considered making the random seed unpredictable
> at startup, but that really hurt testability.
> As you note, random_new_seed sets it.
>
> systo_datestamp() returns a value which is liable to be seconds since the
> epoch, and time() may give (with potential granularity) milliseconds of
> CPU time thus far. I may suggest
>     lisp random_new_seed systo_datestamp();
>
> Some people also play with passing in seeds via environment variables and
> whatever functions specific to the underlying Lisp they can find! In which
> spirit...
>
> If you are using the CSL version you could also try
>     symbolic operator random!-fixnum;
>     random!-fixnum();
> and by default that is randomised as best I can (see code in
> csl/cslvase/arith06.cpp for details). Also in that case
>     redcsl -r NNN
> can be used to override the initial random seed for when you need
> repeatability because you are bug-hunting. It uses a stronger random
> generator than the standard version too. But for CSL/PSL/other-systems
> portability you would want to use the regular version.
>
>
> Arthur
>
> On Mon, 6 Jun 2022, Prof Tony Roberts via Reduce-algebra-developers wrote:
>> Hi,
>>
>> to generate some random numbers in Reduce, I see 
>> random_new_seed(n)  starts a new sequence of random numbers. 
>> Question: is there some way to choose n that is effectively random,
>> such as something from the time of day, or the number of millisecs
>> since startup, or the O/S jot, or ...?
>>
>> I have looked in the User Manual but could see nothing useful.  I am
>> executing Reduce on Mac OSX 12.1
>>
>> Tony

--
---------------------------------------------------------------------
Four books and a toolbox:

* (2020) Linear Algebra for the 21st Century
Oxford University Press. isbn: 978-0-19-885640-5, 978-0-19-885639-9
https://global.oup.com/academic/product/linear-algebra-for-the-21st-century-9780198856399

* (2015) Model emergent dynamics in complex systems
SIAM, Philadelphia. isbn: 9781611973556.
http://bookstore.siam.org/mm20

* (2009) Elementary calculus of financial mathematics
SIAM, Philadelphia. isbn: 978-0-898716-67-2.
http://bookstore.siam.org/mm15

* (1994) A one-dimensional introduction to continuum mechanics
World Sci. isbn: 978-981-02-1913-0.

* (2020) with John Maclean, and J. E. Bunder; Equation-Free
function toolbox for Matlab/Octave.
http://github.com/uoa1184615/EquationFreeGit

Emeritus Professor A.J. Roberts
School of Mathematical Sciences, University of Adelaide
https://profajroberts.github.io/
mailto:Pro...@pr...
http://orcid.org/0000-0001-8930-1552

Re: [Reduce-algebra-developers] Reduce algebra: pls advise how to get an effectively 'random' seed for random number generator

[Arthur N. <ac...@ca...>]

At one time I think we had considered making the random seed unpredictable 
at startup, but that really hurt testability.
As you note, random_new_seed sets it.

systo_datestamp() returns a value which is liable to be seconds since the 
epoch, and time() may give (with potential granularity) milliseconds of 
CPU time thus far. I may suggest
    lisp random_new_seed systo_datestamp();

Some people also play with passing in seeds via environment variables and 
whatever functions specific to the underlying Lisp they can find! In which 
spirit...

If you are using the CSL version you could also try
    symbolic operator random!-fixnum;
    random!-fixnum();
and by default that is randomised as best I can (see code in 
csl/cslvase/arith06.cpp for details). Also in that case
    redcsl -r NNN
can be used to override the initial random seed for when you need 
repeatability because you are bug-hunting. It uses a stronger random 
generator than the standard version too. But for CSL/PSL/other-systems 
portability you would want to use the regular version.


Arthur

On Mon, 6 Jun 2022, Prof Tony Roberts via Reduce-algebra-developers wrote:
> Hi,
>
> to generate some random numbers in Reduce, I see 
> random_new_seed(n)  starts a new sequence of random numbers. 
> Question: is there some way to choose n that is effectively random,
> such as something from the time of day, or the number of millisecs
> since startup, or the O/S jot, or ...?
>
> I have looked in the User Manual but could see nothing useful.  I am
> executing Reduce on Mac OSX 12.1
>
> Tony

[Reduce-algebra-developers] Reduce algebra: pls advise how to get an effectively 'random' seed for random number generator

[Prof T. R. <Pro...@pr...>]

Hi,

to generate some random numbers in Reduce, I see 
random_new_seed(n)  starts a new sequence of random numbers. 
Question: is there some way to choose n that is effectively random,
such as something from the time of day, or the number of millisecs
since startup, or the O/S jot, or ...?

I have looked in the User Manual but could see nothing useful.  I am
executing Reduce on Mac OSX 12.1

Tony

--
---------------------------------------------------------------------
Four books and a toolbox:

* (2020) Linear Algebra for the 21st Century
Oxford University Press. isbn: 978-0-19-885640-5, 978-0-19-885639-9
https://global.oup.com/academic/product/linear-algebra-for-the-21st-century-9780198856399

* (2015) Model emergent dynamics in complex systems
SIAM, Philadelphia. isbn: 9781611973556.
http://bookstore.siam.org/mm20

* (2009) Elementary calculus of financial mathematics
SIAM, Philadelphia. isbn: 978-0-898716-67-2.
http://bookstore.siam.org/mm15

* (1994) A one-dimensional introduction to continuum mechanics
World Sci. isbn: 978-981-02-1913-0.

* (2020) with John Maclean, and J. E. Bunder; Equation-Free
function toolbox for Matlab/Octave.
http://github.com/uoa1184615/EquationFreeGit

Emeritus Professor A.J. Roberts
School of Mathematical Sciences, University of Adelaide
https://profajroberts.github.io/
mailto:Pro...@pr...
http://orcid.org/0000-0001-8930-1552

Re: [Reduce-algebra-developers] reduce-mode.el: bugs in syntax highlighting

[Andrey G. G. <A.G...@in...>]

On Thu, 19 May 2022, Francis Wright wrote:
> I have just checked in a version of reduce-mode.el that handles the two examples below and the examples in your previous email
> correctly, or at least more consistently.
Many thanks!!

Now it is much more pleasant to edit .red files.

Andrey

Re: [Reduce-algebra-developers] reduce-mode.el: bugs in syntax highlighting

[Francis W. <f.j...@li...>]

Dear Andrey,

I have just checked in a version of reduce-mode.el 
<https://sourceforge.net/p/reduce-algebra/code/HEAD/tree/trunk/generic/emacs/reduce-mode.el> 
that handles the two examples below and the examples in your previous 
email correctly, or at least more consistently. If you spot any more 
problems please let me know. I haven't released an updated Emacs package 
yet; that will take a bit longer.

Francis

On 03/05/2022 12:12 pm, Andrey G. Grozin wrote:
> Dear Francis,
>
> for all n1,n2 match x1^n1*x2^n2=f(n1,n2);
>
> is also highlighted incorrectly. I suppose match should have the same 
> highlighting properties as let and clear.
>
> a where {f(~x)=>0,y=>0} when x>0;
>
> where is just black; it would be better to highlight it.
> when is blue, as if it is a function; it would be better to highlight it.
> I think let, match, clear, where, when should be in magenta.
>
> Andrey