2018-02-22 2:36 GMT+01:00 tim Rowledge <tim at rowledge.org>:

> At the latest board meeting we got to discussing the relative quietness of
> the squeak list(s) recently. We were wondering what you folks out there are
> doing with Squeak, what you'd like to be able to use it for, the things
> that you think would be important to improve it for wider use and so on.
>
> Please, whether you're a frequent user or an occasional look-at-the-list
> type, take a moment to let us know your opinions.
>

There's no reason for board members to not answer

I'm an engineer in Navigation/Guidance/Control, formerly in aerospace
domain, and now in naval.
I'm interested in signal processing, sensor data fusion, modelling,
prediction and anticipation of disturbances (predictive control) and plenty
of other cool things, like augmenting the autonomy of unmaned vehicles.

>From 1988 to 2003 I have developped a tool more or less like Matlab, within
a micro company (3 to 5 engineers at most).
I should have said more AND less.
More because built around a higher level concept: the model of a system,
not just the poor Matlab concept of "oh, cool everything is matrix of
complex even a string!" (you know the genericity of assembler what if
everything was a blob of bits?)
It was able to manipulate symbolic expressions (ODE, transfer function,
etc...) all the way down, unlike an exogen Symbolic Toolbox plugged to
nothing.
The first effect of having symbolic expressions was the ability to handle
symbols, not just indices in an array (let's not connect output 3 of block
2 with input 4 of block 5, but simply connect the roll angle phi).
It was able to generate code for simulation or for exporting control
command algorithm.
Plugging an external user system was much easier and efficient than writing
a Simulink sfunction.
User interface was well ahead Matlab: all operations on systems were
available from a Simulink like block diagam browser
And our graphics were quite better, more professional layout with nicer
PostScript output.
Less than Matlab too, quantity-of-available-Toolbox-wise.

In less than two years, the prototype were in production in several EADS
GNC design offices (Aerospatiale company at that time) with about 30 users.
And you know what, all this was written in Smalltalk... (after learning
with STV, developped in st80 v2.5 then objectworks/visualworks)
What else could have been as productive at that time?

We then failed to grow, but did we really want to?
We were engineers, and happy to eat our own dog food everyday,
We wrote the tool for us before thinking to others.


>
> What do you use Squeak for?
>
- keeping programming in Smalltalk as a hobby
- learning from participation to an opensource project
- giving back some contribution to what I consider a potentially productive
environment (one of the most productive)


> If you don't use Squeak, why not?
>
Professionnally: see below


> If you used Squeak in the past and don't now, what pulled you away?
>
Not applicable


> What does Squeak lack that you think might make you use it for 'regular'
> development?
>
First a decent library for developping engineering tools, then acceptation
as a standard.

The standards for my domain are things like Matlab/Python/... (numpy,
scipy, ...)
We more and more need database connection for storing more and more data.
And good and live plots

If I want to deviate from the standard, I know that I will have to spend
tons of energy for convincing.
But if there isn't even a start of decent library, that ain't gonna be
easy...
- OK, OK, so you have something better, show me what you can do with Squeak
  Oh you mean that we'll have to rewrite all the available Python tools
first, but then will have something better?

Acceptation as a standard comes later, if you start having some materials.
It means that new employees have been exposed to the language/tools before.
And we don't have to teach them from scratch. On the contrary, they can
teach us too.

As you might guess, I'm interested in SciSmalltalk (now Polymath), but did
not devote enough time to integrate the community.
I'd like to see more good packages from Pharo ported back or even better
shared by Squeak, rather than simply switching to Pharo because I'm far
more comfortable with Squeak tools (they just work) and community too.


What things are too hard or annoying to do?
>
Gathering know-how on using advanced features of FFI recently cost me too
much
Gathering knowledge about what is up-to-date, what is abandonware, where is
the working version...
I'd like a socially successful SqueakMap.
The growing gap with Pharo is not helping and is really annoying me



> What would you like to be able to use Squeak for?
>
As I said above: a replacement for those tools used by (too) many
engineers: Matlab/Python/...
Maybe musing with web apps, I'm sure it's possible, but it's more a problem
of my own ignorance than of Squeak.
Demonstrating that Squeak has a future requires playing well in those
ecosystems too


> tim
> --
> tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
> C for sinking, java for drinking, Smalltalk for thinking
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20180427/32d95a2d/attachment.html>