Re: [Moose-dev] numerical method in smalltalk

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
15 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

SergeStinckwich
Hi all,

we talked recently on the moose mailing-list about having a more
robust library for doing mathematical stuff in Smalltalk
like SciPython or SciRuby (see below). If there is enough interested,
i'm wondering if we could propose a Google Summer of Code project
about that. I could write a draft for the project.

Regards,

On Mon, Mar 5, 2012 at 10:05 AM, Serge Stinckwich
<[hidden email]> wrote:

> On Sat, Mar 3, 2012 at 4:45 AM, Stéphane Ducasse
> <[hidden email]> wrote:
>> could be interesting to see if we need to some numerical analysis in moose and pharo.
>>
>> I will create a configuration
>>
>>
>> MCSqueaksourceRepository
>>        location: 'http://squeaksource.com/DHBNumerical'
>>        user: ''
>>        password: ''
>
> Thank you. I'm definitively interested by having more mathematical
> stuff available in Pharo.
> There is Scipy [1] since a long time for Python and more recently Ruby
> community do the same with SciRuby [2].
>
> Some time ago, i wrote several random number generators [3] that might
> be integrated in a numeral analysis packages.
>
> At the moment, i'm mostly interested by Runge-Kutta methods [4] in
> order to solve some ordinary differential equations.
> I already find some code on the web like this one:
> http://live.exept.de/doc/online/english/programming/goody_stmath.html
>
> What is really important is to be able to test the result of such algorithms.
>
> [1] https://en.wikipedia.org/wiki/SciPy
> [2] http://sciruby.com/
> [3] http://www.squeaksource.com/Random.html
> [4] https://en.wikipedia.org/wiki/Runge%E2%80%93Kutta_methods
>
> Regards,
> --
> Serge Stinckwich
> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
> Every DSL ends up being Smalltalk
> http://doesnotunderstand.org/



--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

Tudor Girba-2
Hi Serge,

There definitely is interest. If you volunteer for mentoring it would be great.

Cheers,
Doru


On 20 Mar 2012, at 04:35, Serge Stinckwich wrote:

> Hi all,
>
> we talked recently on the moose mailing-list about having a more
> robust library for doing mathematical stuff in Smalltalk
> like SciPython or SciRuby (see below). If there is enough interested,
> i'm wondering if we could propose a Google Summer of Code project
> about that. I could write a draft for the project.
>
> Regards,
>
> On Mon, Mar 5, 2012 at 10:05 AM, Serge Stinckwich
> <[hidden email]> wrote:
>> On Sat, Mar 3, 2012 at 4:45 AM, Stéphane Ducasse
>> <[hidden email]> wrote:
>>> could be interesting to see if we need to some numerical analysis in moose and pharo.
>>>
>>> I will create a configuration
>>>
>>>
>>> MCSqueaksourceRepository
>>>        location: 'http://squeaksource.com/DHBNumerical'
>>>        user: ''
>>>        password: ''
>>
>> Thank you. I'm definitively interested by having more mathematical
>> stuff available in Pharo.
>> There is Scipy [1] since a long time for Python and more recently Ruby
>> community do the same with SciRuby [2].
>>
>> Some time ago, i wrote several random number generators [3] that might
>> be integrated in a numeral analysis packages.
>>
>> At the moment, i'm mostly interested by Runge-Kutta methods [4] in
>> order to solve some ordinary differential equations.
>> I already find some code on the web like this one:
>> http://live.exept.de/doc/online/english/programming/goody_stmath.html
>>
>> What is really important is to be able to test the result of such algorithms.
>>
>> [1] https://en.wikipedia.org/wiki/SciPy
>> [2] http://sciruby.com/
>> [3] http://www.squeaksource.com/Random.html
>> [4] https://en.wikipedia.org/wiki/Runge%E2%80%93Kutta_methods
>>
>> Regards,
>> --
>> Serge Stinckwich
>> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
>> Every DSL ends up being Smalltalk
>> http://doesnotunderstand.org/
>
>
>
> --
> Serge Stinckwich
> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
> Every DSL ends up being Smalltalk
> http://doesnotunderstand.org/
>

--
www.tudorgirba.com

"Not knowing how to do something is not an argument for how it cannot be done."


Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

SergeStinckwich
On Tue, Mar 20, 2012 at 12:45 PM, Tudor Girba <[hidden email]> wrote:
> Hi Serge,
>
> There definitely is interest. If you volunteer for mentoring it would be great.

Yes sure i can do it. I can even maybe propose some local students.
Do you want to co-mentor ?

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

Tudor Girba-2
I would more want to see what can be done with it :). I have very little knowledge in this area and a bit too little time, but I can try to help.

Doru


On 20 Mar 2012, at 07:12, Serge Stinckwich wrote:

> On Tue, Mar 20, 2012 at 12:45 PM, Tudor Girba <[hidden email]> wrote:
>> Hi Serge,
>>
>> There definitely is interest. If you volunteer for mentoring it would be great.
>
> Yes sure i can do it. I can even maybe propose some local students.
> Do you want to co-mentor ?
>
> Regards,
> --
> Serge Stinckwich
> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
> Every DSL ends up being Smalltalk
> http://doesnotunderstand.org/
>

--
www.tudorgirba.com

"Don't give to get. Just give."






Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] Re: numerical method in smalltalk

SergeStinckwich
In reply to this post by SergeStinckwich
2012/3/20 Steffen Märcker <[hidden email]>:
> Hi,
>
> There's a package "Numerical Methods" in Cincoms Public Repository which is
> - I think - based on the book "Object-Oriented Implementation of Numerical
> Methods". Is this package already ported? If not, it might be a good
> starting point. But it needs more documentation...

Yes there is already the same version for Pharo/Squeak on Squeaksource.
There is actually a lot of code dispersed in several projects with
some duplication also.
The idea is to consolidate everything.

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] Re: numerical method in smalltalk

Stéphane Ducasse

On Mar 20, 2012, at 10:34 AM, Serge Stinckwich wrote:

> 2012/3/20 Steffen Märcker <[hidden email]>:
>> Hi,
>>
>> There's a package "Numerical Methods" in Cincoms Public Repository which is
>> - I think - based on the book "Object-Oriented Implementation of Numerical
>> Methods". Is this package already ported? If not, it might be a good
>> starting point. But it needs more documentation...
>
> Yes there is already the same version for Pharo/Squeak on Squeaksource.
> There is actually a lot of code dispersed in several projects with
> some duplication also.
> The idea is to consolidate everything.

Yes!

>
> Regards,
> --
> Serge Stinckwich
> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
> Every DSL ends up being Smalltalk
> http://doesnotunderstand.org/
>


Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] Re: numerical method in smalltalk

SergeStinckwich
In reply to this post by SergeStinckwich
2012/3/20 Steffen Märcker <[hidden email]>:
> It can imagine that it would be useful for the ST community to have this
> cross dialect. Are there any plans in this direction?

Yes in a perfect world, but having already something working on Pharo
is a first step.
After that people could adapt to other flavors if they want.

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

R: [Moose-dev] numerical method in smalltalk

Lorenzo
In reply to this post by SergeStinckwich
Hi Serge,

I have developed a package for fuzzy logic computation (in ST).
Do you think could be helpful?

Grazie

Lorenzo

-----Messaggio originale-----
Da: [hidden email]
[mailto:[hidden email]] Per conto di Serge
Stinckwich
Inviato: martedì 20 marzo 2012 7.13
A: [hidden email]
Cc: Moose-related development
Oggetto: Re: [Pharo-project] [Moose-dev] numerical method in smalltalk

On Tue, Mar 20, 2012 at 12:45 PM, Tudor Girba <[hidden email]> wrote:
> Hi Serge,
>
> There definitely is interest. If you volunteer for mentoring it would be
great.

Yes sure i can do it. I can even maybe propose some local students.
Do you want to co-mentor ?

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/


Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

Stéphane Ducasse

On Mar 20, 2012, at 6:31 PM, Lorenzo Schiavina wrote:

> Hi Serge,
>
> I have developed a package for fuzzy logic computation (in ST).
> Do you think could be helpful?

Yes!
It would be great to organize all the libraries we already have.


> Grazie
>
> Lorenzo
>
> -----Messaggio originale-----
> Da: [hidden email]
> [mailto:[hidden email]] Per conto di Serge
> Stinckwich
> Inviato: martedì 20 marzo 2012 7.13
> A: [hidden email]
> Cc: Moose-related development
> Oggetto: Re: [Pharo-project] [Moose-dev] numerical method in smalltalk
>
> On Tue, Mar 20, 2012 at 12:45 PM, Tudor Girba <[hidden email]> wrote:
>> Hi Serge,
>>
>> There definitely is interest. If you volunteer for mentoring it would be
> great.
>
> Yes sure i can do it. I can even maybe propose some local students.
> Do you want to co-mentor ?
>
> Regards,
> --
> Serge Stinckwich
> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
> Every DSL ends up being Smalltalk
> http://doesnotunderstand.org/
>
>


Reply | Threaded
Open this post in threaded view
|

Re: R: [Moose-dev] numerical method in smalltalk

SergeStinckwich
In reply to this post by Lorenzo
On Wed, Mar 21, 2012 at 12:31 AM, Lorenzo Schiavina <[hidden email]> wrote:
> Hi Serge,
>
> I have developed a package for fuzzy logic computation (in ST).
> Do you think could be helpful?

Dear Lorenzo,

yes might be interesting for SciSmalltalk. Is your code available
somewhere ? What is the licence of your code ?
Are you willing to integrate/adapt your work in a bigger package ?

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

Stéphane Ducasse

On Mar 21, 2012, at 3:49 AM, Serge Stinckwich wrote:

> On Wed, Mar 21, 2012 at 12:31 AM, Lorenzo Schiavina <[hidden email]> wrote:
>> Hi Serge,
>>
>> I have developed a package for fuzzy logic computation (in ST).
>> Do you think could be helpful?
>
> Dear Lorenzo,
>
> yes might be interesting for SciSmalltalk. Is your code available
> somewhere ? What is the licence of your code ?
> Are you willing to integrate/adapt your work in a bigger package ?

serge not everything has to be in one package but in one repository
we could have

        Algo-FusyLogic
        Algo-Math
        Algo-…
        Algo-…

>
> Regards,
> --
> Serge Stinckwich
> UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
> Every DSL ends up being Smalltalk
> http://doesnotunderstand.org/
>


Reply | Threaded
Open this post in threaded view
|

R: R: [Moose-dev] numerical method in smalltalk

Lorenzo
In reply to this post by SergeStinckwich
Hi Serge,

if you can understand Italian, you can see a presentation at:
http://www.edor.it/index.php?option=com_content&view=article&id=30&Itemid=65
This application has been developed in VSE, but a student of mine has
already done (a partial) porting in Squeak; he already told me that he
wanted to send it to Squeak community so, if you want, I can send it to you.
At present the code has no special license and I will be pleased to
integrate it into a bigger package with any kind of license.
My application has a specific fuzzyfication/defuzzyfication method but, if
you consider it useful, I can extend it to some other theoretical approach.
As a matter of fact, my application (not my student's) is an environment for
creating expert systems that have been used in many different fields
(medical, forecasting, sport, environment control).

Ciao

-----Messaggio originale-----
Da: [hidden email]
[mailto:[hidden email]] Per conto di Serge
Stinckwich
Inviato: mercoledì 21 marzo 2012 3.50
A: [hidden email]
Oggetto: Re: [Pharo-project] R: [Moose-dev] numerical method in smalltalk

On Wed, Mar 21, 2012 at 12:31 AM, Lorenzo Schiavina <[hidden email]> wrote:
> Hi Serge,
>
> I have developed a package for fuzzy logic computation (in ST).
> Do you think could be helpful?

Dear Lorenzo,

yes might be interesting for SciSmalltalk. Is your code available
somewhere ? What is the licence of your code ?
Are you willing to integrate/adapt your work in a bigger package ?

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/


Reply | Threaded
Open this post in threaded view
|

Re: [Moose-dev] numerical method in smalltalk

SergeStinckwich
In reply to this post by Stéphane Ducasse
On Wed, Mar 21, 2012 at 1:50 PM, Stéphane Ducasse
<[hidden email]> wrote:

>
> On Mar 21, 2012, at 3:49 AM, Serge Stinckwich wrote:
>
>> On Wed, Mar 21, 2012 at 12:31 AM, Lorenzo Schiavina <[hidden email]> wrote:
>>> Hi Serge,
>>>
>>> I have developed a package for fuzzy logic computation (in ST).
>>> Do you think could be helpful?
>>
>> Dear Lorenzo,
>>
>> yes might be interesting for SciSmalltalk. Is your code available
>> somewhere ? What is the licence of your code ?
>> Are you willing to integrate/adapt your work in a bigger package ?
>
> serge not everything has to be in one package but in one repository
> we could have
>
>        Algo-FusyLogic
>        Algo-Math
>        Algo-…
>        Algo-…


Yes sure you are right. I mean 1 repo in fact.

--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

Re: R: R: [Moose-dev] numerical method in smalltalk

SergeStinckwich
In reply to this post by Lorenzo
On Wed, Mar 21, 2012 at 2:50 PM, Lorenzo Schiavina <[hidden email]> wrote:
> Hi Serge,
>
> if you can understand Italian, you can see a presentation at:
> http://www.edor.it/index.php?option=com_content&view=article&id=30&Itemid=65
> This application has been developed in VSE, but a student of mine has
> already done (a partial) porting in Squeak; he already told me that he
> wanted to send it to Squeak community so, if you want, I can send it to you.

Yes please.

> At present the code has no special license and I will be pleased to
> integrate it into a bigger package with any kind of license.

I think we should have the same licence than Pharo, MIT licence.

> My application has a specific fuzzyfication/defuzzyfication method but, if
> you consider it useful, I can extend it to some other theoretical approach.
> As a matter of fact, my application (not my student's) is an environment for
> creating expert systems that have been used in many different fields
> (medical, forecasting, sport, environment control).

The idea of SciSmalltalk is to build a library not a tool. Maybe some
of part of your fuzzy operators you provide could be reused to be part
of this library.

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/

Reply | Threaded
Open this post in threaded view
|

R: R: R: [Moose-dev] numerical method in smalltalk

Lorenzo
Hi Serge,

I got in touch with Davide Arrigo, the guy who developed SqueakFuzzyWorld;
at present he is in Russia  and he will be back at the end of the month and
he will be pleased to send you directly the program application.

Ciao

Lorenzo

-----Messaggio originale-----
Da: [hidden email]
[mailto:[hidden email]] Per conto di Serge
Stinckwich
Inviato: mercoledì 21 marzo 2012 16.23
A: [hidden email]
Oggetto: Re: [Pharo-project] R: R: [Moose-dev] numerical method in smalltalk

On Wed, Mar 21, 2012 at 2:50 PM, Lorenzo Schiavina <[hidden email]> wrote:
> Hi Serge,
>
> if you can understand Italian, you can see a presentation at:
>
http://www.edor.it/index.php?option=com_content&view=article&id=30&Itemid=65
> This application has been developed in VSE, but a student of mine has
> already done (a partial) porting in Squeak; he already told me that he
> wanted to send it to Squeak community so, if you want, I can send it to
you.

Yes please.

> At present the code has no special license and I will be pleased to
> integrate it into a bigger package with any kind of license.

I think we should have the same licence than Pharo, MIT licence.

> My application has a specific fuzzyfication/defuzzyfication method but, if
> you consider it useful, I can extend it to some other theoretical
approach.
> As a matter of fact, my application (not my student's) is an environment
for
> creating expert systems that have been used in many different fields
> (medical, forecasting, sport, environment control).

The idea of SciSmalltalk is to build a library not a tool. Maybe some
of part of your fuzzy operators you provide could be reused to be part
of this library.

Regards,
--
Serge Stinckwich
UMI UMMISCO 209 (IRD/UPMC), Hanoi, Vietnam
Every DSL ends up being Smalltalk
http://doesnotunderstand.org/