Good, thorough Smalltalk reference
Julian Fitzell
julian at beta4.com
Tue Jan 17 08:56:42 UTC 2006
I won't say documentation isn't needed, but I think you are missing the
point. It isn't that smalltalk code is easy to read (though some may
argue it is), it's that smalltalk code is easier to *navigate*. Senders,
implementers, references; these things make it easier than in any other
system I've used to find your way around quickly. Not to mention the
method finder. If more of our code had good tests, a simple senders
check would give you a very good idea of how to use the method and of
whether it is public/private/etc. A few choice class comments to give
slightly broader information and we'd be in good shape.
Overview information is still great to have as well but we're talking
pretty broad here and I'd say at the level that could easily be handled
on a wiki.
I won't deny that many newcomers crave more printed documentation and
part of that is because they're new and need more documentation, but
another part of it is that they're new and can't picture a world where
you don't need the documentation (I can't even fathom working with, say,
MFC classes without printed docs). So "read the code" is not just a cop
out; it's partly based on the experience of seeing a lot of new people
come onto the list with the same concerns and most of them catching on
pretty quick and becoming as productive in Squeak as the rest of us.
You have to remember that resources are limited and most of us would
rather spend a few minutes pointing newcomers to the tools (teaching
them how to fish, if you will) rather than spending a year writing a
book that will be incomplete and out of date a year down the road.
Like I said, I won't say documentation is a bad idea; I'd love to do it
myself, in theory. But there just doesn't seem to be anyone, currently,
for whom having the largest possible number of squeak users is a big
enough priority. And until there is, most people will prefer to spend
their time just being 10x more productive in their language of choice.
Oh, and I learned more of Squeak faster without docs than I did
Borland's OWL library with about a half-dozen books (and I think OWL was
*easier* than MFC :) ).
Julian
joshscholar at nightstudies.net wrote:
> I agree that there's no substitute for documentation written by humans for
> humans. It's good that smalltalk code is easier to read than other code,
> but even the best code only gives you local information (unless you read
> every line in the damn system, inspect every varible, trace every routine) -
> you need overview information too. And that needs to be organized and
> complete. "Read the code" is a cop out that ignores that fact.
>
> Josh Scholar
>
> ----- Original Message -----
> From: "Lynn Hales" <lhales at earthlink.net>
> To: "The general-purpose Squeak developers list"
> <squeak-dev at lists.squeakfoundation.org>
> Sent: Monday, January 16, 2006 10:25 AM
> Subject: Re[2]: Good, thorough Smalltalk reference
>
>
>
>>Hello Gary and All, for me documentation is paramount. By looking at how
>
> many Smalltalk programmers there are and how many "newbies" there seem to be
> there is a large barrier to entry that I think many can't or don't want to
> see. It's documentation, how to's, cook books, best practices all in book
> form - electronic or paper. Class comments are very important but they are
> not even close to a substitute. Up to date writings are mandatory for wider
> adoption and participation. All the talk about learning the class library,
> how to program in Smalltalk, how to create GUI's in Morphic by reading Class
> comments is nuts. That's not a best practices approach and by just looking
> around it is easy to see it isn't working.
>
>>If we want the benefits of wider participation in the Squeak world then we
>
> need to write and keep what has been written up to date. Lynn
>
>>Monday, January 16, 2006, 4:22:29 AM, you wrote:
>
>
>
More information about the Squeak-dev
mailing list
|