[Self-interest] goals for Self

[Jecel Assumpcao Jr]

Nicklas,
> Creating a more accessible Self:
> Defining a Self standard:

Those are very interesting goals. Having participated in an effort to do
a Smalltalk standard (it would be a follow-on to the previous ANSI
Smalltalk) I can say it isn't easy even if we have many years of
experience with the system, and it is much harder if we are still
exploring. In that project there were multiple Smalltalks that were
independently created while there is only one Self that would be a
reasonable starting point (all other versions are very dead).

Russell,
> artefact instead of a project

Sure, but the discussion is about if we want to make it a project. I an
old Apple II clone that I built myself back in 1984. It currently
doesn't work and is just gathering dust. It might be like that forever
or I could decided to add missing parts and even having it use an SD
card instead of floppies.

Thanks for the ideas on apps.

In fact, one reason for my to use Self would be to take advantage of
existing code. It is very tempting for me to instead start from scratch
on some successor to Self only using the lessons learned from it. In my
particular case I have the following Self code (which are also mostly
interesting artefacts instead of useful for the future):

- CMOS Simulator (runs on Self 1.2 and possibly Self 2)
- Albertina Lourenci's Ecodesign CAD system (runs on Self 4.0 with
Smalltalk extension - this was never released to the public at her
request but I could always do a demo)
- simple physics simulation by Emilene Morais (runs in Self 4.0 like all
that follow, needs to be changed to use staggered time steps)
- Petri net simulator by Luis Campos de Carvalho (I can't seem to find
this code)
- tinySelf 1 (which I mentioned previously)

I see these applications that were part of the Self distribution:
avExtensions, crypto, javaClient	, liveJava, mySQL, preprocessor,
serverDemo, sliceOutliner, sorting, cecil, fastCGI, javaServer, mango	,
pep, selfParser, sha1, smalltalk, webBrowser

Mango is the parser generator by Ole Agesen, Pep is Java-in-Self, Cecil
is Craig Chambers' language and most of the rest have reasonable names.

Steve,
> 1. useful
> 2. ideal

There is some overlap between your 1 and Russell's suggestions (access
to modern libraries would lead to better integrations with the current
computing world, for example). Most of what you list for 2 could be
added easily to existing Self but at some point the "starting over"
alternative starts to look interesting.

-- Jecel