Stephen Pair wrote:
> 
> I also agree with Doug's points (and Thomas' point about fonts).  It seems
> that the most successfull UI cleanup effort to date was the change to make
> the scrollbars more crisp (sorry I forgot who contributed that).

That was actually me. :)

> The change
> was made, and then heavy lobbying occured to get it folded into the base
> realease.  Perhaps that's the same model that should be followed for other
> aspects of the UI.  Is it really necessary to conduct a full top to bottom
> re-examination of the UI?  It would be nice, but it would also mean a lot of
> impact to the existing system, and many more points that people could
> disagree on.
> 
> It seems that small incremental changes to the system have the best chance
> of getting into the base Squeak release...and things seem to die if they
> don't make it into the base release.

It sounds like we're agreed on this... that's why I was going to try to put together a changeset with some cosmetic (Morphic) changes, hopefully the bare minimum of actual code change to maximize it's chance of being incorporated.  (I'll look at Torsten's, Bert's, SMA's etc. changesets as a guide also.  Perhaps there could be some back-and-forth discussion on the list when I post it before it's submitted to SqC.)

> Thomas Kuehne wrote:
> >
> > I am summarizing here:
> >
> > o No native widgets, except for special application development.
> > o Do not try to imitate another look and feel
> >   (allows Squeak to be the role model in UI Design rather than
> > vice versa).
> > o Better font support.
> > o Do some immediate cosmetic changes.
> >
> > I second everything Doug said and would like to add that better
> > font support is really important.

Sounds good.  The only other thing I would add to the list that Stephan Wessels mentioned is:

o Don't rely on skins/themes as the total solution to look & feel issues

Skins/themes are cool and fun to play around with, but aren't really necessary for a basic default look improvement.  Plus I imagine they involve quite a lot of code change to truly support them... integrating bitmapped components and all that.  (Stephan's skins changesets are around 300K... hopefully mine will be less than 10K.)

> > There are very good arguments to counter the native widget demand
> > but there is no good excuse to turn off people with a -- lets say
> > 'very special' -- look which is not easy to change (It took me
> > several unsuccessful attempts with various approaches and
> > changeset compatibility problems to finally get over the font hurdle).

Amen. ;)

> > What about making the system window borders a bit thicker?
> > It is always a bit of a precision job to hit the border with the
> > mouse pointer so that resizing is possible.

I remember someone posting a changeset which implemented this which I'll look up.  I may keep that change separate for now, we'll see.  (You definitely want to strike a balance between ease of resizing and having too-thick borders hogging space.)

- Doug Way
  dway at mat.net