Re: levels of compliance



John R Sheets wrote:
> 
> sungod wrote:
> 
> > can someone give me an example of how a _rule_ for human interface would
> > meet, say, level 3 compliance or level 4 compliance but not level 1 or
> > 2?
> 
> I believe I was pushing for the compliance levels to be an indication of how
> important each style element/rule was towards being a good GNOME app.  It
> would be a measure of 1) how uniquely "GNOME-y" a rule was, and 2) how much
> it contributed towards being a universally good UI.  I think it would be
> good to separate (or at least make note of) which style elements are
> universal, and which are GNOME-specific.  This will have the side effect of
> distilling "what it means to be GNOME" into a quicker, more concise read (a
> step toward sun's other post about learning by browsing).  Developers
> already familiar with good UI design could skip over the universal elements,
> but they should definitely be included somewhere.

this is pretty reasonable, but i guess i ought to mention that many of
the rules i've been thinking of and am about to propose should mostly be
"level 1" compliant, then: they should all be general rules for layout
and widget choice, to make an application "user-centric" and taking into
accounts the human psychology and tendencies to categorize, build
relationships, and err.

> The previous proposal was to put these universal elements into the core-most
> compliance level, and have the GNOME-specific elements in the following
> (less critical) compliance levels.  However, now I am beginning to question
> that idea, too.

right, well, according to this model i guess the "levels"-based
hierarchy of organization isn't necessarily _broken_; i think we should
proceed with the foundation as it is rather than calling for a complete
rewrite at this point. let's go ahead and get those level-one rules
hashed out; if we find ourselves reasonably satisfied with the
completeness of the user interface guide at the end and discover that
all the rules for good design are in 1 and 2 and we have nothing for
levels 3 or 4, we can decide then whether the levels are really called
for.

> > good human interface design is not limited to whether a feature is in
> > or out; it involves determining _where_ and _how_ the features (_any_
> > features really) are implemented.
> 
> Yep.  I think we should also make it clear where to find a given style
> element, whether it's in gnome-libs (and what the API looks like), or
> supported by some external library, or even if it's some experimental new
> feature that isn't fully implemented yet, but that people should be free to
> make use of (and so they're aware of it in the first place).

agreed, although i'm afraid i won't be much help in this area. i can't
read code (yet; still working on that class). this will be an area where
developer assistance will be required.

> So maybe three conceptual types of style elements: Universal, GNOMEish, and
> Experimental.  Should they be intermixed throughout the style guide, or
> should they be separated into each of three discrete parts/chapters/books?

rather than dividing them into three separate parts/chapters/books we
can, as i've proposed, use the "level" hierarchy as it exists now, put
them into the categories that exist now (like "menus" and "toolbars"
etc) and determine the suitability of organization after the rules
exist. let's not get too hung up on this, though: as you can see, we got
too carried away with organizing the structure first last time and still
have nothing of substance to show for all our efforts. :)

> P.S.--Hi, sun!  Good to see ya again.

ps. good to see you too. it's encouraging to find people who really _do_
want to get some work done on this list. :) i'm glad to be back.
-- 
 ______(sungod)_____________________________________
| To ensure privacy and data integrity this message |
| has been encrypted by using dual rounds of ROT-13 |
 --------------------------(as387@yfn.ysu.edu)------



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]