Re: irc summary

["William R. Tipton" <tiptonwi ENGR ORST EDU>]

/lurker off

Chris Jantzen writes:
>Speaking as a developer, and estimating what I think most level-headed
>developers I know would say: If there is not one inch for
>customization in most areas of the Style Guide, that is fine by me. As
>programmers, we're supposed to be writing programs that do
>something. Not conducting independent user interface style
>research.

Exactly.  Any developer who says otherwise isn't likely to abide by the
user interface standards anyway.  While the Macintosh sure pissed off a lot
of programmers with its Human Interface Guidelines (correct?), I think
you'd be hard pressed to find a user who would say that the standard
screwed them up.  The Mac's FDI really made this a requirement: not having
a rigid guideline/standard/whatever simply makes it too difficult to
manuever when you can so easily (and accidentally) switch applications.

Consider the "X-like" mouse focus: if, by accident, I slide my cursor over
to another window and then hit the keystroke to save for the app that I
*thought* had focus, what would you expect the other app to do?  If the
keystroke for save in the first app was actually "shred" in the one that
had accidental focus, how would I be expected to know that?

>programmer here.] If I can make library calls to create a standardized
>appearance style tab in my Preferences notebook, sure, I'll do
>it. This is why About dialogs in Gnome are already standardized:
>they're dead simple.


Again, I agree on this one.
I do a lot of development in MFC under Win32, and I must say that I spend
far too much time balancing widgets and trying to figure out exactly how
everything should look.  I spend too much time looking at other programs
that I find flow properly and trying to figure out what they do right.  I
spend too much time testing the application against people who haven't used
it before.

The little things really do count, but in Win32 it's harder than it should
be to do the little things (as John Sheets pointed out).  Code that abides
by the constantly changing "standard" in Windows (which I now believe to be
IE 4-like) gets obfuscated real quick.

>If the Style Guide suffers because some coders
>complain about loss of control, Gnome *will* suffer. A kite does not
>soar to heights without being held to the ground.


Exactly.  Gnome is in some serious trouble if it gives coders too much
control.

I asked a question similar to this in the IRC meeting thing, and the quick
response was, "what came first?  user or the coder?"

I think Fred Brooks (author of The Mythical Man-Month) has a lot to say on
this subject (as he does just about everything related to programming).  In
"Aristocracy, Democracy, and System Design" Brooks addresses some of the
issues directly related to the concept of "architects will get all the
creative fun and shut out the inventiveness of the implementers" when he
says flat out that it is nothing more than an illusion.  He writes:
"The opportunity to be creative and inventive in implementation is not
significantly diminished by working within a given external specification,
and the order of creativity may even be enhanced by that discipline."

I fear that the Gnome GUI guidelines will not be tough enough to say
something as absolute as, "forget control of programmers, give consistency
and control to users."

>reactions and flames, but I believe that I am speaking for a majority
>of real world programmers here. Think carefully before replying (let


I, for one, like to think of myself as a real world programmer.  I agree
with you whole-heartedly.

My fifteen cents.

william r. tipton