Flexibility and Application Programming (Re: irc summary)

[Chris Jantzen <chris maybe net>]

Kenneth R . Kinder writes:
 > I get the feeling here, Chris, that basicly you're arguing for rapid
 > development through shortcuts. Didn't the KDE do this by using Qt?  Gnome
 > should take the time to do things right.

(This is not what I'm 'basically arguing for'.)

Huh? Not at all. I am saying, in part, that at some distant future,
application developers should have rapid development (the holy grail
in the real world). But the initial development of the GUI libraries
will certainly not be rapid.  As I have said or implied, GNOME should
have a rich set of well implemented libraries that conform to and
encourage SG compliance. 

 >     In the perfect world, which definitly doesn't exist, and I'm not
 >     asking you to make it - in the perfect world, we'd present INFORMATION
 >     to the *UI, and expect that user interface to figure out how to

Well, yes. But, here, again you misunderstood (common theme in this
mailing list): I'm not saying you should necessarily be relieved of
placing controls and what-not. I'm saying decisions about matters like 
"Should I put Exit on the File menu? Or should I put Quit on the
Session menu?" should already be resolved. Common application options, 
preferences, and menu items should be a cinch to implement.

 >  2. X's suffering from standards.  Right, well, that's true.  It's mostly
 >     due to Motif being proprietary, that a code base isn't commonly
 >     shared.  X programs behave differently, right?  So whip them into
 >     shape, and force them to behave, right?  Wrong.

No, right. Confused users loses market share. Isn't this the basis for 
this entire list? Create standards and guidelines that whip all future 
applications into shape? I'm confused...

 >     I think window managers are a class-act example of something done
 >     wonderfully elegant on X.

Agreed.

 >  Consistancy is the goal, but only no each
 >     individual's desktop.

Hmmmm.... Hrrmmmm.... There is much to be said here. If things are too
configurable, could you ever hope to use someone else's machine? What
if you screw up your own beyond repair?  Still, we do like to
customize, don't we? :-)

 >  It's better to make a standard so hopelessly
 >     flexible for the user's configuration and additional
 > programming.

Am I alone, that the word "hopelessly" sounds warning sirens?

 >     Projects which have followed this philosophy - like fvwm and E, emacs,
 >     and sendmail mark themselves and great standards.

Oooooh. Each of those is a subject of bitter flame wars that range far
and wide across the `net. Needless to say, many users cringe at the
thought of configuring, much less learning each and every one of those
programs. Many, indeed, would spit and cuss at 'mark themselves as
great standards'. (Though I, personally, quite like most of those
programs.)

Also, my experience is that emacs, for example, is in fact so
hopelessly configurable that nobody does. Configure it, that
is. [Nobody referring to most regularish users I know that use
emacs.]

 >  3. Doing what others are doing?  I think you implied that we ought to do
 >     what Apple did because they did it so well - make a GUI.  Sure, they
 >     did okay, but the Mac's lack of flexibility is so painfull, I won't
 >     use it, and if that happened so badly to gnome/*nix, I wouldn't use
 >     it.

Don't throw the baby out with the bathwater. Just because you don't
like what they did overall, doesn't mean there isn't something to be
learned. My point in all of this is that the Macintosh certainly did
one thing right: rigid human interface guidlines. Certain things must
be present in certain places with certain names. This gives users
comfort and eases the learning curve.

Never have I said that visual appearance should be strictly
standardized. Themes and customization are greatly welcomed. Logical
flow and presentation should be standardized.

(I have seen a Macintosh tool to give MacOS "themes" of a sort, BTW.)

 >     Steve Jobs is the software dude behind Apple and the Mac, right?
 >     Taken any note of his later indevours toward perfection?  OpenStep
 >     maybe?  OpenStep is very theme-inclined, and a far better model for
 >     anything than MacOS.

Steve Jobs is not doing anything New and Different with any of his
NeXT/OpenStep work that he hasn't done with MacOS. Look to the
fundamentals, and then judge. The face may have changed somewhat, but
the heart is the same.

-- 
chris jantzen kb7rnl =-> 
---------------------------------------------------
systems administrator |    __O   |          student
maybe.net             |  _`\<,_  |     oregon state
possibly the best     | (*)/ (*) | computer science
---------------------------------------------------
  http://www.maybe.net/ - mailto:chris@maybe.net