Re: [Usability] HIG: Fixing the sort arrow direction guideline



On Mon, 2009-05-18 at 17:53 +0100, Calum Benson wrote:
> On 17 May 2009, at 13:58, Celeste Lyn Paul wrote:
> 
> > Also, the language "natural" order is very strange, does the GNOME  
> > HIG mean
> > "logical" order? There should never be a table that isn't sorted in  
> > some way.
> > Usually a good default is the first column is increasing by default  
> > unless
> > there is a better column to sort by.
> 
> "Logical" is probably a little better than "natural"; the main point  
> the HIG is trying to convey here is that it doesn't always mean  
> "increasing".  E.g. sorting a checkbox column, the goal is really to  
> group the items in the list such that (usually) all the checked items  
> come first, then all the unchecked columns.  There isn't really any  
> "increasing" or "decreasing" order.

Having the checked items come first is inherently neither "logical" nor
"natural".  In some circumstances, that might feel "right" and
"defaulty" to have the checked items first; while in others the opposite
might easily be true.  Shall we conclude that one of these instances
should have an arrow pointing in one direction when the checked items
are at the top, while the other should have an arrow pointing the
opposite direction when the checked items are in the same position?

I would hope the answer is "no"; but it seems to me that unless the HIG
comes out and says, "checked-items-first corresponds to a(n) [ascending|
descending] sort," that is the logical outcome.

FWIW, having checked items listed first strikes me as a descending sort
order. In addition to the fact that that is mathematically accurate if
the checkbox is treated as a boolean value, the notion of important
things first corresponds, generally to a descending order; e.g., having
the heuristically "heaviest" items at the top, or ordering by the newest
date first.

Notwithstanding the above, I do agree it is the case that, some of the
time, there is not an intrinsically obvious ascending sort order.  But
I'm very skeptical of what I perceive to be your claim: that in such
cases there may be an intrinsically obvious (or at least reasonably
apparent) "natural" or "logical" ordering.

Fortunately, such cases are not the most common.  Most of the time, the
field that one is interested in sorting on has a readily apparent
alphanumeric nature; and thus a similarly obvious ascending sort order.
Just because there exists a checkbox field in the row doesn't mean a
user is *likely* to want to sort on that field; and that does diminish
the importance of the sort indicator associated with the field.  In
fact, I suggest that if there truly is no apparent ascending sort order,
one should consider omitting the directional indicator altogether.

-- 
Braden McDaniel <braden endoframe com>



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