[IAEP] squeak/etoys accepted as free software...

[Jecel Assumpcao Jr]

Holger Levsen wrote:
> On Friday 07 November 2008 19:45, Jonas Smedegaard wrote:
> > The Squeak image "Etoys" (the only one currently packaged officially for
> > Debian) is in "non-free" due to ftpmasters judging it not possible for
> > the security team to maintain throughout the (multiple year long)
> > lifespan of a Debian release.
> 
> IIRC/IIUC this is one aspect why the ftpmasters didnt accept it in main. More 
> generally said, (IIRC) it's because the impossibility to bootstrap etoys.

Is the subject correct? I mean I know we are talking about a directory
called "non-free" but is there anyone out there that after what has been
said still doesn't accept etoys as Free Software? It seems to me that
the discussion has moved on to the issue of etoys being "unsafe"
instead. It is just that Debian doesn't have an "unsafe" directory.

> Even though the etoys developers "don't do it" and the stateful VM (or rather 
> patches to it) is/are the prefered form of modification. 

Note that several Smalltalks can be built entirely from a set of text
files: Self, GNU Smalltalk, Slate, Little Smalltalk and others. There is
no technical difficulty. But as you said, the people who can do it don't
have any reason to do so. This leads us to the situation where there is
a group of people who want to do something which they feel would be very
important but they can't do it themselves and another group that could
do it but are busy with other things. It is very easy for discussions to
get heated under such circumstances.

> And while I dont agree with the position  they are taking (should I 
> say, "anymore"..) I can understand why they do: because it makes sense and 
> (probably also) because this is like it always was: traditional software has 
> to have the ability to be bootstrapped or build.

It is a matter of being able to check out things for yourself instead of
having to take some other person's word that all is ok. I don't see
anything unreasonable about this position in theory. In practice the
volume of things to be checked might be such that your resources will
not be enough and you will have to do some random sampling and hope that
will be enough to scare evil doers from trying to slip in bad stuff.

> Squeak is special in this case and I dare to say "new". (I know it was started 
> in the 70ties :) But not all people do.)
> 
> So my planned approach to get it into main in the long run, is to start a 
> general discussion in Debian about this kind of software, thus stopping to 
> special case squeak.

That supposes the people involved are not fully informed and might
change their opinions given more facts, but as Jonas Smedegaard wrote
this is probably not be the case. I probably shouldn't be adding
gasoline to the fire, but the fact is that the current plan for Squeak 5
will give us something far more alien than what you have seen so far -
http://netjam.org/spoon/naiad

I will confess that I don't personally care much about getting Squeak
into any Linux distribution. Don't get me wrong - I love Debian, Ubuntu,
Slackware, Red Hat, Gentoo and all the rest and have used them all
myself. But consider the following groups of people:

- children and teachers not familiar with Linux. For these the only
practical options are either having it pre-installed like on the XO or
getting the web plugin version from squeakland.org. Having it be just an
apt-get away doesn't help these people.

- Linux users who might try out different languages, including Squeak.
This is the group for whom having a package in their distribution will
make a difference.

- Linux users who really are interested in Squeak for some reason. This
group would be helped by a package in their distribution but Squeak is
simple enough that they can probably get started by downloading from
squeak.org.

So the most convencing argument would be to say that getting Squeak into
Debian main will allow it to come pre-installed on Edubuntu for the
first group. In practice, however, all large scale educational
deployments using Linux that I am aware of have used really obscure or
even home grown distributions. I am just trying to evaluate the
practical results of putting in some effort to make this kind of thing
happen. As you pointed out, nothing at all needs to be done about Squeak
to get Sugar into any Linux since Sugar doesn't depend on Squeak in any
way. So the issue is making Sugar the same on all platforms by having
the same set of pre-installed stuff, right?

Let me give you an example of what I mean by results vs effort - the
original Squeak License had three things that bothered different groups
of people. One of these was the export clause that didn't allow Squeak
to be shipped to countries like Cuba. Now that the three issues have
been fixed (at a very high cost in terms of effort diverted from
development) will we see happy Cuban children playing with eToys on
their XOs? If not then all we have achieved was making a tiny group of
very vocal people happy that a certain phrase was removed from some
text. Nothing else was changed.

-- Jecel