[REQUEST] updated documentation

Bijan Parsia bparsia at email.unc.edu
Tue Nov 23 04:48:15 UTC 1999


[Ok, ok, I'm behind :)]
At 11:37 AM -0500 11/21/99, Alan Kay wrote:

>Bijan --
>
>I agree that this is the way it *should* be -- but, at present, the
>rhetoric for running code and being able to understand how it got there --
>and just what is there -- is quite different.

I agree. Well, I agree to a certain extent. I think Morphic is trickier to
understand, at the moment, that lots of other parts of the system. I'm not
entirely sure if this is merely because there's more, say, MVC
documentation, or because Morphic isn't as "settled" yet.

> Also, none of the browsers
>are good at showing the logic of the larger "application" artifact that the
>user encounters.

No quibbles there!

>(This is what led me/us long ago to want to have
>prototypical instances as exemplars of classes,

I'll add that I've been bit with the XP bug, and think that, at least from
a *programmers* perpective, unit tests can make for effective documentation
(eh, but perhaps not *standalone* documentation). Functional tests,
esepcially UI based ones, might work well for users.

> and to have an "application
>browser". The latter would be greatly facilitated if there were more of a
>theory for doing such.

Theory! Yes! More theory! We need it! (The trained philosopher in me
drools. :))

> E.g. Morphic has a theory of how a morph should be
>constructed, but there is still not a real Morph browser -- moreover, I

Now, even without a brower, having a clear articulation of the theory of
morph construction would be nice. I believe I've read something like,
"Well, you can do things the normal Smalltalky way, or you could do things
in a kinda of Selfishy way." That's confusing, even if true!

>think all would agree that there could be much more of a theory for larger
>useful tools, and then we could build a very nice
>browser/inspector/explorer/debugger that would make crystal clear how
>things are organized and used.

I'm agreeing. I thought the Syntax Browser was an interesting experiment.

>     This is now -- IMO -- a very pressing need. Please contribute ideas.

Er. Maybe after slogging through everyone else's comments :)

Cheers,
Bijan Parsia.





More information about the Squeak-dev mailing list