Pink Plane vs Blue Plane

Steven Swerling sps2000 at
Thu Feb 13 17:11:10 UTC 2003

Bijan Parsia wrote:
 > [megasnip]

First of all, thanks for not mentioning that I mixed up pink and blue 
plane (and thanks to those who did mention it tactfully, off list).

 > I notice that when some of these things are actually
 > provided, they are not picked up and maintained/supported/used
 > by many of the "requestors" (BobsUI is a nice counterexample

I know I'm guilty of this, but I've never had a request to update 
something, so posting updates hasn't been a big priority. (SPSBrowser, 
SPSMethodSearcher, SVI, my widgets, and the ModalEnforcer stuff live on 
in my own image, evolve, and they are in constant use, if anyone cares). 
If my contributions aren't good enough for people to use, too bad for 
me. But I sense a cheap shot being taken there, and I think you are 
giving short shrift to how hard it is to create UIs and maintain them in 
Squeak if you're not in the inner circle. Have you tried it? I'm talking 
UI, here. It's really hard to keep up. I believe that even BobsUI 
languished for years until recently. I keep an eye on this list when I 
can, and I swear it looks like some of these guys treat squeak as a full 
time job with overtime. Does Ned Konz get paid for this? And it's not 
just the pace of change that is making it hard. There is just something 
unwieldy about all that extra stuff in Morphic -- by extra stuff I mean 
everything beyond basic handling of buffered drawing, user events, and 
stepping for the basic widgets. It's stiffling -- there is a 
psychological barrier to writing new widgets for Morphic.

 > But I really don't understand. You can build straightforward apps using
 > the pluggablefoo classes.

The PluggableFoo widgets are not of acceptable quality for shareware, 
the kind of software I would like to be able to put together. It would 
be too much to put together a UI, since I'd have to develop many of the 
widgets -- and if nobody else were to like those widgets, I would have 
to maintain them by myself, too. MVC is out of the question, they do not 
look professional. Good looking widgets, in Morphic, are required to 
write shareware. And shrinking the image below 6MB (w/ app) is beyond by 
abilities. That's too big for, say, a Turkish Vocabulary Builder -- the 
competing apps are about 150k, crummy as they are. And at least they 
still *look* good.

 > Squeak is a big, old, evolved system.

Word. So what are the fruits? Wait. Let me make the point this way:

1. Take all of the Applications written in Squeak. Not Smalltalk -- Squeak.
2. Subtract the ones created specifically for Squeak developers
3. Subtract those in use primarily by Squeakers
4. Subtract the ones used primarily by the captive audience of a 
Squeaker (ie. students or focus groups or other groups that have to use 
a squeak application for one reason or another)

What apps are you left with? How old is squeak? Something like 4 years. 
And it's based on a mature technology far older. So where are the 
applications? I'm not talking about the ones for the gearheads, or being 
forced on students. Just the simple shareware applications, developed 
for people who have never and will never meet a Squeak developer, and 
think that smalltalk is what you do at cocktail parties. The number that 
would shake out, I beleive, is too low to chalk it up to the 
inadequacies of people like me.

Saying that the cruftyness of Morphic is inherent to smalltalk, or that 
it doesn't matter, or that Morphic should be more like eToys, it's 
getting carried away. But these happen to be the opinions of one of the 
brightest stars in the Squeak firmament, with The Great Creator himself 
and some almost as bright stars like yourself agreeing. What they (and 
you) are saying is likely to have a big effect on the direction things 
go, and it dampens my hopes for the efforts of the SqueakMappers to 
reduce Morphic down to a 'base' and a 'rest' ('like the skins of an 
onion'), as suggested on another thread. That would be a lot easier with 
the blessing of SqueakC. Who is going to "take ownership" of EToys if 
not someone from SqueakC?

More information about the Squeak-dev mailing list