Suggestion components to add onto SAGE
q10
It has come to my attention that SAGE does not have a units-conversion
program component (maybe it has; if it does, please show me). I
recommend adding the GPLed unit conversion program called ConvertAll
(http://www.convertall.bellz.org/) I believe it is written with
Python, so integration of it into SAGE should not be too difficult.
It seems that SAGE is also missing a geometry sketchpad component. It
would be nice to add the GPLed 2D/3D geometry sketchpad program known
as Geogebra (http://www.geogebra.org).
Finally, there is a very good graphing program known as K3DSurf
(http://k3dsurf.sourceforge.net/) that makes very good 3D/4D/5D graph
models. Please consider adding this program into SAGE as well.
Thanks for hearing my suggestions,
Please reply,
-q10
David Joyner
>
> Hello:
>
> It has come to my attention that SAGE does not have a units-conversion
> program component (maybe it has; if it does, please show me). I
> recommend adding the GPLed unit conversion program called ConvertAll
> (http://www.convertall.bellz.org/) I believe it is written with
> Python, so integration of it into SAGE should not be too difficult.
From the above webpage:
"ConvertAll is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either Version 2 of the License, or (at your
option) any later version.
...
ConvertAll requires the following libraries:
* Qt (Version 4.1 or higher - see Trolltech for more information)
* Python (Version 2.3 or higher)
* PyQt (Version 4.0 or higher - see Riverbank for more information)"
Requiring Qt might be a problem. However, I don't see why a units-conversion
package requires a graphics library like Qt. Indeed, the web site says
"Command line options are available to do conversions without the GUI."
>
> It seems that SAGE is also missing a geometry sketchpad component. It
> would be nice to add the GPLed 2D/3D geometry sketchpad program known
> as Geogebra (http://www.geogebra.org).
This has been discussed before. I think it has serious licensing
incompatibility issues and also uses java.
>
> Finally, there is a very good graphing program known as K3DSurf
> (http://k3dsurf.sourceforge.net/) that makes very good 3D/4D/5D graph
> models. Please consider adding this program into SAGE as well.
Were you able to download and compile the source code for this?
The website (or my computer) was very slow for me, so I gae up looking.
If you did so, what OS did you use? Do you know the license for k3dsurf?
mabshoff
On Jul 19, 3:39 pm, "David Joyner" <wdjoy...@gmail.com> wrote:
>
> > Hello:
>
> > It has come to my attention that SAGE does not have a units-conversion
> > program component (maybe it has; if it does, please show me). I
> > recommend adding the GPLed unit conversion program called ConvertAll
> > (http://www.convertall.bellz.org/) I believe it is written with
> > Python, so integration of it into SAGE should not be too difficult.
>
> From the above webpage:
>
> "ConvertAll is free software; you can redistribute it and/or modify it
> under the terms of the GNU General Public License as published by the
> Free Software Foundation; either Version 2 of the License, or (at your
> option) any later version.
>
> ...
>
> ConvertAll requires the following libraries:
>
> * Qt (Version 4.1 or higher - see Trolltech for more information)
> * Python (Version 2.3 or higher)
> * PyQt (Version 4.0 or higher - see Riverbank for more information)"
>
> Requiring Qt might be a problem. However, I don't see why a units-conversion
> package requires a graphics library like Qt. Indeed, the web site says
>
> "Command line options are available to do conversions without the GUI."
introduce a massive amount of overhead for something that should be a
nice and self contained little library.
>
> > It seems that SAGE is also missing a geometry sketchpad component. It
> > would be nice to add the GPLed 2D/3D geometry sketchpad program known
> > as Geogebra (http://www.geogebra.org).
>
> This has been discussed before. I think it has serious licensing
> incompatibility issues and also uses java.
> > Finally, there is a very good graphing program known as K3DSurf
> > (http://k3dsurf.sourceforge.net/) that makes very good 3D/4D/5D graph
> > models. Please consider adding this program into SAGE as well.
>
> Were you able to download and compile the source code for this?
> The website (or my computer) was very slow for me, so I gae up looking.
> If you did so, what OS did you use? Do you know the license for k3dsurf?
>
>
>
> > Thanks for hearing my suggestions,
> > Please reply,
> > -q10
Michael
Jason Grout
> On Sat, Jul 19, 2008 at 6:10 PM, q10 <benso...@gmail.com> wrote:
>> Hello:
>>
>> It has come to my attention that SAGE does not have a units-conversion
>> program component (maybe it has; if it does, please show me). I
>> recommend adding the GPLed unit conversion program called ConvertAll
>> (http://www.convertall.bellz.org/) I believe it is written with
>> Python, so integration of it into SAGE should not be too difficult.
>
> From the above webpage:
>
> "ConvertAll is free software; you can redistribute it and/or modify it
> under the terms of the GNU General Public License as published by the
> Free Software Foundation; either Version 2 of the License, or (at your
> option) any later version.
>
> ...
>
> ConvertAll requires the following libraries:
>
> * Qt (Version 4.1 or higher - see Trolltech for more information)
> * Python (Version 2.3 or higher)
> * PyQt (Version 4.0 or higher - see Riverbank for more information)"
>
> Requiring Qt might be a problem. However, I don't see why a units-conversion
> package requires a graphics library like Qt. Indeed, the web site says
>
> "Command line options are available to do conversions without the GUI."
>
Enthought also has a units python module, which might be interesting to
look at. It's not very well documented, though.
>
>> It seems that SAGE is also missing a geometry sketchpad component. It
>> would be nice to add the GPLed 2D/3D geometry sketchpad program known
>> as Geogebra (http://www.geogebra.org).
>
> This has been discussed before. I think it has serious licensing
> incompatibility issues and also uses java.
>
>> Finally, there is a very good graphing program known as K3DSurf
>> (http://k3dsurf.sourceforge.net/) that makes very good 3D/4D/5D graph
>> models. Please consider adding this program into SAGE as well.
>
> Were you able to download and compile the source code for this?
> The website (or my computer) was very slow for me, so I gae up looking.
> If you did so, what OS did you use? Do you know the license for k3dsurf?
k3dsurf is a standard package that can be installed on ubuntu through
the package manager. It is nice, but I think it's a kde thing, which
would be a problem on Mac OSX, Windows, and other platforms we do or
would like to support.
Jason
William Stein
From the ConvertAll webpage: "Why write another unit converter? There
are plenty of them out there. Well, I couldn't find one that worked
quite the way I wanted.
With ConvertAll, you can combine the units any way you want. If you
want to convert from inches per decade, that's fine. Or from
meter-pounds. Or from cubic nautical miles. The units don't have to
make sense to anyone else."
The justification for ConvertAll would be exactly the justification for us to
*not* use it in Sage, i.e., for Sage unit conversion we would surely want
something that works well in the context of sage itself, e.g., the coercion
model, symbolic calculus, etc. Whatever unit conversion sage eventually
gets, it will have to be done probably as a bunch of new code that might
use something like ConvertAll or enthought's unit code to do some of
the real work behind the scenes. I think something like this will happen
when a student comes to me (or whoever) asking for a several-months
long Sage development project, and unit conversion is something they
find interesting. It's the sort of thing where one would have to:
(a) research exactly what Maple, Mathematica, and Matlab do
(b) come up with a SEP inspired by (a) for Sage, and get it
discussed on sage-devel,
(c) survey the range of existing unit conversion libraries to see
if any are helpful in implementing (b).
(d) implement something.
-- William
William Stein
>>> would be nice to add the GPLed 2D/3D geometry sketchpad program known
>>> as Geogebra (http://www.geogebra.org).
>>
>> This has been discussed before. I think it has serious licensing
>> incompatibility issues and also uses java.
Making it easy to include Geogebra based applets in the Sage notebook
would be a worthwhile thing to do, and I'm sure can be done in a way that
avoids licensing problems. The Geogebra project itself is confused
about software licenses -- they write on their download page "You are
free to copy, distribute and transmit GeoGebra for non-commercial
purposes. Please see the GeoGebra license for details." The license
itself on code is according
to the license:"GeoGebra's source code is subject to the GNU General
Public License:" So in fact they cannot restrict its use to
non-commercial
purposes. I'll write to the project for clarification in a moment.
The fact that Geogebra applets use java is certainly not a show stopper
for them having something to do with Sage, since jmol also uses java.
It might make a lot of sense e.g., to have an optional geogebra spkg
that makes it trivial for people to create a bunch of geogebra applets
in the sage notebook...
>>> Finally, there is a very good graphing program known as K3DSurf
>>> (http://k3dsurf.sourceforge.net/) that makes very good 3D/4D/5D graph
>>> models. Please consider adding this program into SAGE as well.
>>
>> Were you able to download and compile the source code for this?
>> The website (or my computer) was very slow for me, so I gae up looking.
>> If you did so, what OS did you use? Do you know the license for k3dsurf?
>
>
> k3dsurf is a standard package that can be installed on ubuntu through
> the package manager. It is nice, but I think it's a kde thing, which
> would be a problem on Mac OSX, Windows, and other platforms we do or
> would like to support.
k3dsurf is a kde gui program, hence completely unsuitable for inclusion
in Sage. Plus sage already has 3d surface plotting.
William
Robert Dodier
> The justification for ConvertAll would be exactly the justification for us to
> *not* use it in Sage, i.e., for Sage unit conversion we would surely want
> something that works well in the context of sage itself, e.g., the coercion
> model, symbolic calculus, etc.
which exploits its symbolic functions to make working with units
convenient, extensive, and flexible. I claim to have gotten some
useful & interesting code at this point.
A dimensional quantity like 10 meters is represented as 10 ` m,
i.e. a Maxima expression with the operator ` and arguments 10 and m.
As usual (for the most part) Maxima carries along such expressions
when they appear in algebraic operations, which leaves the door open
for us to define results appropriate for dimensional quantities.
E.g. product of dimensional quantities => product of nondimensional
parts times product of units. Unit conversions are indicated by
the `` operator, e.g. 10 ` m `` ft => 12500/381 ` ft. Units are
converted by constructing and solving a system of linear equations
(after taking logarithms).
So far the package as it stands in CVS doesn't have anything beyond
algebra, but I was just trying some calculus operations and it looks
like integrals and derivatives should work OK. E.g. stuff like
diff(x(t) ` m, t ` s) => dx(t)/dt ` m/s,
integrate(v(t) ` m/s, t`s, a`s, b`s) => integrate(v(t), t, a, b) ` m.
I haven't worked out the details yet.
Ezunits also has some functions for dimensional analysis which were
adapted from an existing Maxima package, and also a collection of
physical constants (CODATA 2006) from NIST.
Ezunits exists in previous versions of Maxima but there will be a
substantial revision in the 5.16 release which will appear next month.
Anyway maybe this is some use or interest. Comments on the CVS
version are much appreciated.
Robert Dodier
William Stein
Since you've clearly been thinking of this from a developer's perspective,
and maybe even spending a lot of time writing code, is there any
chance you could just dump some of your design thoughts in an
email here? As a frame of reference, the last time *I* personally
thought a lot about units was when I was taking a physics course
in 1992, so you can imagine that I'm actually pretty clueless about
what people really want. Some questions:
(1) Are the list of all units one uses pretty standard? Is there
a table, say in Wikipedia, with pretty much all of them? Or do
people make up new units in the course of their work or research?
Obviously I know about converting between fahrenheit
and celcius or between Euros and Dollars -- both are "units"
computations for me, where of course the Euro/Dollar FX
rate varies every second, which is kind of amusing.
(2) Are there *any* difficult algorithms that involve units or is this
mostly a notational and representation problem plus some algebra
and simplification?
(3) Does Maxima, Maple, Mathematica, Matlab or Axiom do anything
particularly cool, surprising or clever involving units?
-- William
Robert Dodier
> Since you've clearly been thinking of this from a developer's perspective,
> and maybe even spending a lot of time writing code, is there any
> chance you could just dump some of your design thoughts in an
> email here?
> (1) Are the list of all units one uses pretty standard? Is there
> a table, say in Wikipedia, with pretty much all of them? Or do
> people make up new units in the course of their work or research?
relatively small set of units which are conventional for the field.
The union of all units in all fields is fairly extensive though.
Any computerized units system would have to include the SI units.
Beyond that, there are so-called customary units (feet, pounds, etc)
which are still widely used in the US; I am not aware of other non-SI
systems in extensive use. I think it's fairly common to see semi-SI
units such as MJ/day or mg/kg. I think it's common enough to invent
to invent new units like those for convenience.
The implication, to me, is that any units system has to be easily
extensible, since it's likely that random users would want units
that escaped our notice.
> (2) Are there *any* difficult algorithms that involve units or is this
> mostly a notational and representation problem plus some algebra
> and simplification?
complex operation in the ezunits package is to solve a system of
linear equations. Most of the fun seems to be in the notation and
representation. Maybe there would be some interesting algebra
problems if only I had more imagination.
> (3) Does Maxima, Maple, Mathematica, Matlab or Axiom do anything
> particularly cool, surprising or clever involving units?
the way of units, but from what I have seen, there is nothing very
exciting. Which is OK --- something boring which just works right
would be very useful.
FWIW
Robert Dodier
Tim Lahey
>
> William Stein wrote:
>> (3) Does Maxima, Maple, Mathematica, Matlab or Axiom do anything
>> particularly cool, surprising or clever involving units?
>
> I have looked around to see what Maple and Mathematica have in
> the way of units, but from what I have seen, there is nothing very
> exciting. Which is OK --- something boring which just works right
> would be very useful.
I've used the Maple units package and it works fairly well for most
of the calculations I've done with it. It's an add-on package but is
still useful enough to handle most things I've thrown at it. The biggest
problem I've had is the appropriation of symbols for use in units when
I've wanted to use them for something else. The best example is
the use of A for amps when I wanted to use A to represent an area.
Cheers,
Tim Lahey
---
Tim Lahey
PhD Candidate, Systems Design Engineering
University of Waterloo
root
>the way of units, but from what I have seen, there is nothing very
>exciting. Which is OK --- something boring which just works right
>would be very useful.
Frink is the standard for unit conversion.
<http://futureboy.homeip.net/frinkdocs>
I cannot find any mention of a license.
It seems to be implemented in Java.
Tim
David Joyner
>
>>I have looked around to see what Maple and Mathematica have in
>>the way of units, but from what I have seen, there is nothing very
>>exciting. Which is OK --- something boring which just works right
>>would be very useful.
>
> Frink is the standard for unit conversion.
> <http://futureboy.homeip.net/frinkdocs>
>
> I cannot find any mention of a license.
At http://futureboy.us/frinkdocs/faq.html under the heading "Philosophy"
there is the question "Why isn't this open source?"
Robert Dodier
> I've used the Maple units package and it works fairly well for most
> of the calculations I've done with it. It's an add-on package but is
> still useful enough to handle most things I've thrown at it. The biggest
> problem I've had is the appropriation of symbols for use in units when
> I've wanted to use them for something else. The best example is
> the use of A for amps when I wanted to use A to represent an area.
Maxima
at least to prevent evaluation of units (so you could assign something
to A and still have A appear as a symbol in units), but that would
also
prevent stuff like ev(12 ` goats + 5 ` sheep, goats = sheep/2), i.e.
evaluate w/ some conversion indicated by binding goats to a value.
Maybe that kind of assignment is less important than preventing
unintended evaluation of units symbols. Undecided at this point.
best
Robert Dodier
Robert Dodier
stuff wrapped in a reimplementation of various wheels ...
Frink would work a lot better as a Python package, or Maxima for that
matter, although ezunits covers most of the stuff covered by Frink.
One thing that I hadn't thought about before is the conversion of time
units (timezones). It's fairly tricky to get it right, and it's widely
useful.
FWIW
Robert Dodier
Harald Schilly
about units. I think (altough never implemented) the best way is some
kind of intelligent wrapper around a value. This wrapper is able to
work with other unit wrappers and is more or less "transparent" to
other calculations or returns new unit wrappers as the situation
requires.
It's important to understand, that units are not just values in a
table, because they can be built up upon other more basic values: e.g.
a division of two units returns another unit or two different units
can be added in some cases, in others not. I've not thought about the
details, but this behaviour is something I've seen before: I have an
old HP 48/49 pocket calculator that is able to work with units in that
way.
So, in this sense units are a new kind of number that wraps basic
numbers, intervals or symbolic expressions.
Maybe, it should also be possible to work with units without a value.
In applied physics and engeneering, sometimes it is nice to check a
formula just by inserting the units. e.g.
>>> f(a,b,c) = ...
>>> f(Unit('kg'), Unit('s')^2, Unit('m')^2) is Unit('ampere')
True
Harald
Gary Furnish
I have a prototype based on the old system somewhere). There is no
need for wrappers because you can just treat the units as unevaluated
symbolic variables, and do substitution as needed to convert between
unit systems.
Carl Witty
> ... for Sage unit conversion we would surely want
> model, symbolic calculus, etc.
could work in Sage, but I didn't get something I was happy with.
My goal was to get something that used the coercion model. I decided
there should be a parent for every combination of unit and base ring,
so if foo were "5 feet" you might see
sage: foo.parent()
Dimension feet over Integer Ring
sage: foo
5 feet
Then there would be coercions between different units of length, etc.;
(and there would also be a conversion to strip off the units):
sage: a
3 in
sage: b
254 mm
sage: a+b
13 in
sage: ZZ(a)
3
sage: ZZ(b)
254
One problem is that polynomials assume that the variable is
dimensionless. For instance, it would be nice to have something like
this work:
sage: a
9.8 m/s^2
sage: v
35 m/s
sage: t = polygen(...)
sage: 1/2 * a * t^2 + v*t
but it can't, because all the coefficients of a polynomial must have
the same parent, but a is an acceleration and v is a velocity.
> (a) research exactly what Maple, Mathematica, and Matlab do
written in Emacs Lisp with arbitrary-precision arithmetic, symbolic
integration, and all sorts of other cool features. That's what I use
for all my unit-conversion needs.
Carl
Carl Witty
> (1) Are the list of all units one uses pretty standard? Is there
> a table, say in Wikipedia, with pretty much all of them? Or do
> people make up new units in the course of their work or research?
make up new units and define their conversions; things like "1
gallon_of_paint is 150 feet^2" or "100 paces is 237 feet".
Carl
Gary Furnish
consider any units system useless.
Dan Drake
It's also useful because some places use units you've never heard of.
Here in Korea, it's very common to use "pyeong" as a unit of area.
The government wants everyone to use square meters, but still everyone
talks about how big their apartments are in terms of pyeong.
To use Carl's syntax, "1 pyeong is 3.3058 meters^2", or "1 pyeong is
35.586 feet^2".
Also, in Japan, they traditionally use tatami as a unit of area: 1
tatami is 1.62 meters^2. (Except for the parts of Japan where it's 1.82
meters^2: http://en.wikipedia.org/wiki/Tatami)
Dan
--
--- Dan Drake <dr...@mathsci.kaist.ac.kr>
----- KAIST Department of Mathematical Sciences
------- http://math.kaist.ac.kr/~drake
DuaneKaufman
Another package that can handle units (plus do a lot more) is Konrad
Hinsen's ScientificPython package:
http://dirac.cnrs-orleans.fr/plone/software/scientificpython
It has units (and unit conversion/computation capability). I had asked
about its inclusion a while ago, but there didn't seem to be much
interest.
Duane
William Stein
Maybe we should reconsider. There isn't even an optional package of
ScientificPython for sage at present? Is it a good project, good code, etc?
What does it do?
I tried downloading the devel version (released a year ago) and:
teragon-2:ScientificPython-2.7.8 was$ sage -python setup.py install
netCDF not found, the netCDF module will not be built!
If netCDF is installed somewhere on this computer,
please set NETCDF_PREFIX to the path where
include/netcdf.h and lib/netcdf.a are located
and re-run the build procedure.
Traceback (most recent call last):
File "setup.py", line 163, in <module>
headers.append(netcdf_h_file)
NameError: name 'netcdf_h_file' is not defined
teragon-2:ScientificPython-2.7.8 was$
DuaneKaufman
On Jul 27, 10:45 pm, "William Stein" <wst...@gmail.com> wrote:
(and maybe the -dev package as well?)
Perhaps it is no longer being actively worked on by the author.
Duane
DuaneKaufman
I took the liberty of contacting Konrad:
On 28.07.2008, at 13:23, Duane Kaufman wrote:
>Lately there has been chatter on the Sage forums about including a units facility in the
>default install. A few collections have been forwarded, including ScientificPython.
>There were a few concerns raised, such as its dependence on a NetCDF installation, plus
>the fact that not much has happened to it recently.
>Is ScientificPython still being developed? Is there a flag to omit the netCDF
>dependency? Will it build cleanly against numpy?
Yes, yes, and yes. The development site is at
http://sourcesup.cru.fr/projects/scientific-py/
and a public Mercurial archive of the latest developments is available
at
http://dirac.cnrs-orleans.fr/hg/ScientificPython/main/
As you can see, development continues steadily. The 2.7 version line
is compatible with
NumPy, the next "stable" line (2.8) will use NumPy by default. If no
netCDF library is
found during installation, the netCDF interface is not built, so
netCDF is not a strict
dependency.
Hope this helps,
Konrad.
--
---------------------------------------------------------------------
Konrad Hinsen
Centre de Biophysique Moléculaire, CNRS Orléans
Synchrotron Soleil - Division Expériences
Saint Aubin - BP 48
91192 Gif sur Yvette Cedex, France
Tel. +33-1 69 35 97 15
E-Mail: hinsen (at) cnrs-orleans.fr
On Jul 28, 6:09 am, DuaneKaufman <duane.kauf...@gmail.com> wrote:
> On Jul 27, 10:45 pm, "William Stein" <wst...@gmail.com> wrote:
>
>
>
> > On Tue, Jul 22, 2008 at 3:59 AM, DuaneKaufman <duane.kauf...@gmail.com> wrote:
>
> > > Hi,
>
> > >http://dirac.cnrs-orleans.fr/plone/software/scientificpython
>
William Stein
>
> Hi,
>
> I took the liberty of contacting Konrad:
>
> On 28.07.2008, at 13:23, Duane Kaufman wrote:
>
>>Lately there has been chatter on the Sage forums about including a units facility in the
>>default install. A few collections have been forwarded, including ScientificPython.
>>There were a few concerns raised, such as its dependence on a NetCDF installation, plus
>>the fact that not much has happened to it recently.
>
>>Is ScientificPython still being developed? Is there a flag to omit the netCDF
>>dependency? Will it build cleanly against numpy?
>
> Yes, yes, and yes. The development site is at
>
> http://sourcesup.cru.fr/projects/scientific-py/
>
> and a public Mercurial archive of the latest developments is available
> at
>
> http://dirac.cnrs-orleans.fr/hg/ScientificPython/main/
>
> As you can see, development continues steadily. The 2.7 version line
> is compatible with
> NumPy, the next "stable" line (2.8) will use NumPy by default. If no
> netCDF library is
> found during installation, the netCDF interface is not built, so
> netCDF is not a strict
> dependency.
>
> Hope this helps,
> Konrad.
Could you post a tarball of a version of ScientificPython with the
property that if I do:
tar zxvf ScientificPython-x.y.z.tar.gz
cd ScientificPython-x.y.z
sage -python setup.py install
then it will work? I want to try it it. Right now,
teragon-2:ScientificPython-2.7.8 was$ sage -python setup.py install
netCDF not found, the netCDF module will not be built!
If netCDF is installed somewhere on this computer,
please set NETCDF_PREFIX to the path where
include/netcdf.h and lib/netcdf.a are located
and re-run the build procedure.
Traceback (most recent call last):
File "setup.py", line 163, in <module>
headers.append(netcdf_h_file)
NameError: name 'netcdf_h_file' is not defined
-- William