Bravo!

[bstjean]

Just watched you presentation "Doing Math & Science in Pharo with PolyMath" on YouTube!  Wow!  Great!  Too short, too much stuff to cover.  But then, you cannot make a 16-hour presentation! But a few comments..

Regarding the Rump's polynomials, one attendee asked the question I was hurling to SHOUT: why not use fractions?  In fact, I've always told fresh & new Smalltalkers : You think you'll Float?  Nah!  You'll sink!  When you have to deal with precision or big exponents, use Fractions, NOT Floats... As I wrote here: https://endormitoire.wordpress.com/2019/11/24/smalltalk-et-les-nombres/.

The ONLY negative-more-or-less thing I have to say about your presentation is the omission of the name of Didier H. Besset.

Now, where do we go from there?  I've submitted SplitMix64.  Are you into combinatorics?  I've got a SequenceableCollection>>#permutationsDo: WITHOUT repetition that is faster that the original (with repetitions then aggregating into a Set, it gets soooooooo slow when dealing with millions of combinations!).  What about cubic roots, getting the digits of Pi 14 digits at a time, etc?

You got me started.

I can see NO reason why the best programming language in the world cannot eclipse all others in any domain/field.

Food for thought.

Why limit ourselves to a finite set of primitives?  String, Collection and Integer could benefit some C code in the right places!  I guess that's an Eliot Miranda & Clément Béra question!

P.S.  We need more "Statistiques" !  ThatM's what rules the world these days!

[Serge Stinckwich]

On Sun, Dec 22, 2019 at 1:14 PM 'bstjean' via PolyMath <polymath...@googlegroups.com> wrote:

Just watched you presentation "Doing Math & Science in Pharo with PolyMath" on YouTube!  Wow!  Great!  Too short, too much stuff to cover.  But then, you cannot make a 16-hour presentation! But a few comments..

Thank you Benoît.

Regarding the Rump's polynomials, one attendee asked the question I was hurling to SHOUT: why not use fractions?  In fact, I've always told fresh & new Smalltalkers : You think you'll Float?  Nah!  You'll sink!  When you have to deal with precision or big exponents, use Fractions, NOT Floats... As I wrote here: https://endormitoire.wordpress.com/2019/11/24/smalltalk-et-les-nombres/.

Yes you are right. I can adapt the example like that:

f := [ :x :y |
(((1335/4) - (x raisedTo: 2)) * (y raisedTo:6)) + ((x raisedTo: 2)*
((11 * (x raisedTo: 2)* (y raisedTo: 2)) - (121/1*(y raisedTo: 4)) - 2)) + ((11/2) * (y raisedTo: 8)) + (x/(2*y)) ].
f value: 77617/1 value: 33096/1.
(-54767/66192) asFloat. "-0.8273960599468214"

In fact this example was more to illustrate that in some case the precision impact the results.

The ONLY negative-more-or-less thing I have to say about your presentation is the omission of the name of Didier H. Besset.

I already talk about him in previous presentation of PolyMath.

Now, where do we go from there?  I've submitted SplitMix64.  Are you into combinatorics?  I've got a SequenceableCollection>>#permutationsDo: WITHOUT repetition that is faster that the original (with repetitions then aggregating into a Set, it gets soooooooo slow when dealing with millions of combinations!).  What about cubic roots, getting the digits of Pi 14 digits at a time, etc?

You got me started.

I was busy recently, so your contribution SplitMix64 is still not integrated in PolyMath:

https://github.com/PolyMathOrg/PolyMath/issues/157

We are missing more people to assess PR and integrate them.

Please join the #polymath channel on Pharo's discord in order to discuss about that.

Every contribution is valuable :-)

 

I can see NO reason why the best programming language in the world cannot eclipse all others in any domain/field.

Food for thought.

Why limit ourselves to a finite set of primitives?  String, Collection and Integer could benefit some C code in the right places!  I guess that's an Eliot Miranda & Clément Béra question!

At the moment, all the code is in pure Pharo, but we can imagine use FFI to use C libraries or PythonBridge to use Numpy or other Python libraries.

P.S.  We need more "Statistiques" !  ThatM's what rules the world these days!

Yes definitively, but we have already a lot of tools like PCA or T-SNE.

Thank you your contributions.

--

Serge Stinckwic

h

Int. Research Unit

 on Modelling/Simulation of Complex Systems (UMMISCO)

Sorbonne University

 (SU)

French National Research Institute for Sustainable Development (IRD)

U

niversity of Yaoundé I, Cameroon

"Programs must be written for people to read, and only incidentally for machines to execute." https://twitter.com/SergeStinckwich

[bstjean]

First and foremost,

Thanks for all the hard work Serge!  In other words (that we each understand in our native language), "un gros merci pour tout ce travail, ces efforts et les résultats qui, naturellement, découlent de toutes ces parcelles de passion qui s'agglutinent autout d'un projet commun et cher".

En tant que "baveux" de nature (je suis Québécois quand même!!), je ne vois AUCUNE raison de ne pas torcher Python en ce qui a trait à QUOI QUE CE SOIT!

;)

[Serge Stinckwich]

Yes sure, we just need more people in the project and more money to support the developers.

You can be a sponsor here: https://github.com/sponsors/SergeStinckwich

Regards,