Andrew,

<mnt/soapbox>

I definitely don't want to get into a slanging match about this,
and I appreciate the frank opinions from this group. The funny
thing is, I agree with nearly all of your sentiments. There is
one point I want to make very strongly however - It is
important to make the base of the pyramid as large as
possible, by whatever means we have.

What I mean by this is, neither you nor anyone else knows
where the next great programming ideas will come from. Out
of 5000 people that look at Squeak, only 10 may turn out
to usefully enhance it. The point is, you don't know who those
ten will be, or what attracted them. Worse, you won't even
be able to predict what sort of background these people will
have. (Heck, Alan Kay has a background in molecular biology,
and I assume you are a lawyer. And me, I do neuroscience).

So, to get those 10 potential Smalltalk gurus,
unfortunately, you need the 4990 others.
To say that we need to attract 'techies' is a dangerous
assumption. In 'The cathedral and the bazaar' Eric
Raymond mentions that one of the most fundamental
improvements in the architecture of fetchmail came
about because of a suggestion from a user.

Look at evolution - you have to try all those random mutations
before you will get some kind of improvement. If you knew a priori
which mutations were going to be successful, cool,
you could throw evolution away.

Critical Mass is incredibly important. More users create a
synergistic effect - the rate of improvement increases,
which in turn increases the rate of new users adopting the
language. Don't underestimate this. This is exactly the
effect by which languages (and even operating systems
:-) with perhaps lesser merit propagate. If you have enough users,
there will always be someone to fix the shortcomings.
Look at the rapid progress of Java and C++ against far more elegant
languages like lisp, haskell, and smalltalk.

My extreme attitude is that it actually doesn't matter
who you attract, as long as you attract enough of them.
(self organisation and emergent properties will take
care of the rest). The scary thing is that the converse
may also be true. That is, even if you attract the most
brilliant people in the world, if you don't have a
critical mass in the user base, a project can
wither and die.

>By the same analysis, we should long
>ago have abandoned the antiquated Smalltalk syntax, adopted a purely
>C++ style of coding and documentation and undertaken a host of other
>things by which "a number" of "new users" are put off.


Remember the separation of Model and View. What we are trying
to promote is the Model - the Smalltalk mindset. This Model can
have any number of Views attached to it. If one particular View
attracts more people to the model then that is a good thing.
Model := C++ defeats the purpose.

The irony in your sarcastic suggestion is that Sun took it to heart
when they came up with Java. Without that 'bridge' to the
C syntax it is unlikely they would ever have got so much developer
interest. While Java started behind the eight-ball, it's rate of
improvement is snowballing. How many good Smalltalkers
have been lost to the 'Java effect'?

>Unlike Python, which has been very stable for quite some time; Squeak
>really isn't ready for new users with thin skins -- those guys are
>bound to be disillusioned with each major change, and get in the way
>of necessary changes that need to be done.
>
>The worst thing that could happen is to develop a community of
>marginally committed newbies who become an excuse for Squeak to
>ossify in the interest of "upward compatibility."  At this time, this
>open source project appears to be looking for committed innovators
>rather than "customers."


Like I said above, it is actually irrelevant who you attract, if you believe
that no-one knows where the next great advancement will come from.
Self selection does the job here. Let people decide for themselves
whether or not to get dissillusioned. Don't assume to do it for them.
Did I get dissillusioned when the early versions of Squeak crashed
time after time?

>Looks are important only for marketing.  Real techies appreciate what
>it is from the spec sheets.  Right now the community doesn't need
>people who aren't real techies.
>Elitist though it may sound, I'm not sure the Squeak community really
>needs a raft of eyes-agog newbies who won't or can't dig deep enough
>to see Squeak for what it is.  There's plenty of use for such people
>later, but let them be the second and third generation, after things
>have settled in a bit.


Very dangerous attitude. The great thing about newbies is they see
things from a different perspective. Let nature take it's course. If
Squeak is truly unsuitable for 'unsophisticated' users they will soon
leave. However, that is for evolution to decide, not us.

>I'm not sure that's really a big deal, however.  Squeak doesn't need
>new users who would be turned off, but rather great minds who will
>make it what Peter envisions it can be.


And where exactly do you think these great minds will come from?

>>Anyway, I have attached a screenshot of one potential Squeak...
>>(8 bit color to save space)
>
>Why not add images of exotic dancers?  :-)
>


I am now preparing the limited Black Label edition of Squeak featuring
none other than Pamela Anderson and (exclusive to Squeak) Naomi
Campbell. Immerse yourself in the this true multimedia feast for only
$99. (First 5000 callers only).

</umnt/soapbox>


Peter