Prev: Ephemeral ports
Next: ioctl or api to get size of fdisk partition p0 (entire physical disk) on solaris x86
From: David Kirkby on 19 Feb 2010 17:28
On Feb 19, 8:22 pm, Waldek Hebisch <hebi...(a)math.uni.wroc.pl> wrote:
> In sci.math.symbolic David Kirkby <drkir...(a)gmail.com> wrote:
>
>
>
> > There is not much serious maths software for Solaris, and that which
> > does exist, costs a small fortune. In contrast, Sage is free and open-
> > source.
>
> Dave, it is very nice that you work on porting Sage.  But the sentence
> above is misleading.  At least FriCAS and Maxima work on Solaris.
> And I think that few other serious open source systems work too.
>
> Your sentence suggests that up to now only serious systems on
> Solaris were proprietary, which is not the case.
>
> --
>                               Waldek Hebisch
> hebi...(a)math.uni.wroc.pl

I take your point - what I said was misleading.

What I meant was there is no general purpose maths package - something
that aims to cover a wide area of mathematics, like Mathematica does
for example. Maxima & FriCAS cover computer algebra only. You can't
use them to plot graphs for example.

Maxima is included in Sage, along with about 100 other open-source
packages. Many have had to be modified in some way to work on Solaris
- usually as a result of assumptions that the build system is Linux.

Sorry for any confusion.

Dave
From: Raymond Toy on 19 Feb 2010 18:25
On 2/19/10 5:28 PM, David Kirkby wrote:
> On Feb 19, 8:22 pm, Waldek Hebisch <hebi...(a)math.uni.wroc.pl> wrote:
>> In sci.math.symbolic David Kirkby <drkir...(a)gmail.com> wrote:
>>
>>
>>
>>> There is not much serious maths software for Solaris, and that which
>>> does exist, costs a small fortune. In contrast, Sage is free and open-
>>> source.
>>
>> Dave, it is very nice that you work on porting Sage. But the sentence
>> above is misleading. At least FriCAS and Maxima work on Solaris.
>> And I think that few other serious open source systems work too.
>>
>> Your sentence suggests that up to now only serious systems on
>> Solaris were proprietary, which is not the case.
>>
>> --
>> Waldek Hebisch
>> hebi...(a)math.uni.wroc.pl
>
> I take your point - what I said was misleading.
>
> What I meant was there is no general purpose maths package - something
> that aims to cover a wide area of mathematics, like Mathematica does
> for example. Maxima & FriCAS cover computer algebra only. You can't
> use them to plot graphs for example.

Maxima can plot 2D and 3D graphs. Maxima can do more than just computer
algebra. Since Maxima has its own language, you can make it do whatever
you want. Maxima also uses Lisp, so you can do whatever Lisp program
you want too.

Ray
From: Waldek Hebisch on 19 Feb 2010 21:44
In sci.math.symbolic David Kirkby <drkirkby(a)gmail.com> wrote:
>
> What I meant was there is no general purpose maths package - something
> that aims to cover a wide area of mathematics, like Mathematica does
> for example. Maxima & FriCAS cover computer algebra only. You can't
> use them to plot graphs for example.

FriCAS can plot graphs. In fact some folks prefered to use FriCAS
instead of Mathematica because they liked more FriCAS graphs.

--
Waldek Hebisch
hebisch(a)math.uni.wroc.pl
From: David Kirkby on 20 Feb 2010 20:59
On Feb 19, 11:25 pm, Raymond Toy <toy.raym...(a)gmail.com> wrote:
> On 2/19/10 5:28 PM, David Kirkby wrote:
>
>
>
> > On Feb 19, 8:22 pm, Waldek Hebisch <hebi...(a)math.uni.wroc.pl> wrote:
> >> In sci.math.symbolic David Kirkby <drkir...(a)gmail.com> wrote:
>
> >>> There is not much serious maths software for Solaris, and that which
> >>> does exist, costs a small fortune. In contrast, Sage is free and open-
> >>> source.
>
> >> Dave, it is very nice that you work on porting Sage.  But the sentence
> >> above is misleading.  At least FriCAS and Maxima work on Solaris.
> >> And I think that few other serious open source systems work too.
>
> >> Your sentence suggests that up to now only serious systems on
> >> Solaris were proprietary, which is not the case.
>
> >> --
> >>                               Waldek Hebisch
> >> hebi...(a)math.uni.wroc.pl
>
> > I take your point - what I said was misleading.
>
> > What I meant was there is no general purpose maths package - something
> > that aims to cover a wide area of mathematics, like Mathematica does
> > for example. Maxima &  FriCAS cover computer algebra only. You can't
> > use them to plot graphs for example.
>
> Maxima can plot 2D and 3D graphs.  Maxima can do more than just computer
> algebra.  Since Maxima has its own language, you can make it do whatever
> you want.  Maxima also uses Lisp, so you can do whatever Lisp program
> you want too.
>
> Ray

I was not aware Maxima could plot graphs.

The point about having its own language does not detract from the fact
that to implement certain things in lisp or Maxima would be a huge
task. Would it be practical to extend Maxima to have the statistical
capabilities of R for example? Sage integrates Maxima, R and many
other packages.

I have not used Sage much myself, so do not claim to know the
capabilities of its constituent parts. Clearly there is some overlap
between them. Sage aims to integrate them as closely as possible so
the capabilities of all of them can be accessed in a reasonably
consistent way.

Here's a list of the components

http://www.sagemath.org/links-components.html

I'd be keen to hear from anyone who either downloads the Solaris
binary, or builds the source code on Solaris.

Dave
From: Richard Fateman on 21 Feb 2010 12:45
David Kirkby wrote:
> On Feb 19, 11:25 pm, Raymond Toy <toy.raym...(a)gmail.com> wrote:
Since Maxima has its own language, you can make it do whatever
>> you want. Maxima also uses Lisp, so you can do whatever Lisp program
>> you want too.
>>
>> Ray
>
> I was not aware Maxima could plot graphs.
>
> The point about having its own language does not detract from the fact
> that to implement certain things in lisp or Maxima would be a huge
> task.

Certain things? Like "building a computer program with human-like
intelligence?" Sure. But it would probably be no harder than in
some other language, and maybe easier.

Would it be practical to extend Maxima to have the statistical
> capabilities of R for example? Sage integrates Maxima, R and many
> other packages.

Of course. Maxima already includes many scientific floating-point
packages, and could include more. To some extent the ease of importing
additional facilities not written in Lisp depends on which of several
possible Lisp implementations are being used under the particular
Maxima. Some lisp systems allow for the calling of programs written in
C, FORTRAN, Assembler, Java, Python.

Not all Lisps; this has been a bit of a stumbling block since the
older GCL has poor support for this; there are more supportive lisps
that also runs on (almost?) all platforms. In fact, the version
of Maxima in Sage is not in GCL.)

Without leaving Lisp, programs have been incorporated in Maxima,
including GCL Maxima by translating FORTRAN automatically to lisp by
"f2cl", and then compiling.

As for R, Octave, or for that matter, the commercial Matlab, it is
hardly obvious how to best attach two systems each with its own
interactive top-level interpreter. Switching between them? How to
relate results in one to the other? Passing commands as quoted strings
from one place to another, and re-parsing results?

But can one include statistical computing in Maxima? Sure. Can
statistical computing be done in Lisp? Sure. Look up XLISP-STAT.
(This is not an entirely happy story, but see
http://www.jstatsoft.org/v13/i07/paper )

I don't know what you are thinking, in particular. You do know that the
Common Lisp standard has extensive support for arrays, floating-point,
etc.

Sage does not "integrate" packages so much as lash them together. I have
not tested Sage, but my suspicion from reading the developer message
board is that there are problems. For example, one can "divide by zero"
in several of its packages. Is this treated in an "integrated" fashion?
Or how about the treatment of infinity and NaNs?
>
> I have not used Sage much myself, so do not claim to know the
> capabilities of its constituent parts.

I daresay there are very few people who "know" the details of most of
the programs.

Clearly there is some overlap
> between them. Sage aims to integrate them as closely as possible so
> the capabilities of all of them can be accessed in a reasonably
> consistent way.

That is perhaps the aim. You apparently are not in a position to judge
how effectively the project is moving towards achieving that aim, nor
how it compares in effectiveness to (say) just using one of its
components for some set of tasks, nor even the extent to which
Mathematica covers the same ground as some of the components of Sage.
>
> Here's a list of the components
>
> http://www.sagemath.org/links-components.html
>
> I'd be keen to hear from anyone who either downloads the Solaris
> binary, or builds the source code on Solaris.

I would not be surprised if you were alone, or nearly alone, in enjoying
trying to build Sage on Solaris. Since you say you don't use it much
yourself or know much about it, this strikes me as being particularly
generous of your time to the Sage people and the Solaris users.

RJF
 |  Next  |  Last
Pages: 1 2 3 4
Prev: Ephemeral ports
Next: ioctl or api to get size of fdisk partition p0 (entire physical disk) on solaris x86