Inidicator Arrow convention for Sort Order  
Author Message
Roedy Green





PostPosted: 2006-2-20 21:31:00 Top

java-programmer, Inidicator Arrow convention for Sort Order Which way should an arrow point to indicate an ascending sort or an
ascending list follows?

For an ascending sort the bigger numbers are at the bottom.

So no matter which way you do it, you can misinterpret.

Is there some other less ambiguous convetion to use?
--
Canadian Mind Products, Roedy Green.
http://mindprod.com Java custom programming, consulting and coaching.
 
Roedy Green





PostPosted: 2006-2-20 21:42:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order On Mon, 20 Feb 2006 13:31:04 GMT, Roedy Green
<email***@***.com> wrote, quoted or
indirectly quoted someone who said :

>Which way should an arrow point to indicate an ascending sort or an
>ascending list follows?
>
>For an ascending sort the bigger numbers are at the bottom.
>
>So no matter which way you do it, you can misinterpret.
>
>Is there some other less ambiguous convetion to use?

There are other sources of ambiguity:

is in the arrow telling you how the sort is NOW or how you want it to
be if you click it?

Yet another source of ambiguity. When you write elements you start at
the top and work down. When you look at a pile, you think of starting
at the bottom and working up.

Does the arrow represent the direction you read to see ever bigger
entries or does it represent the order the computer went though the
list when it was writing the page top to bottom?

SAP uses a red triangle array pointing up means the data are sorted
ascending NOW.

here are two other conventions from Oliver Biot.


I would propose to use an unambiguous convention, maybe
with 3 triangles of increasing size pointing left to right
for natural sort order, and the vertically flipped image
for the reverse order. This assumes a left-to-right writing
convention, of course :)

Another possibility is to write a->z (z<-a) for text columns
and 0->9 (9<-0) for numeric columns. We could do a similar
thing for date/time columns.


Because there are so many ways to interpret this, perhaps trying to
indicate direction will inevitable confuse folk who presume the other
convention. perhaps a colour code.
--
Canadian Mind Products, Roedy Green.
http://mindprod.com Java custom programming, consulting and coaching.
 
Thomas Weidenfeller





PostPosted: 2006-2-20 21:44:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order Roedy Green wrote:
> Which way should an arrow point to indicate an ascending sort or an
> ascending list follows?

Down.

/Thomas
--
The comp.lang.java.gui FAQ:
ftp://ftp.cs.uu.nl/pub/NEWS.ANSWERS/computer-lang/java/gui/faq
http://www.uni-giessen.de/faq/archiv/computer-lang.java.gui.faq/
 
 
Thomas Weidenfeller





PostPosted: 2006-2-20 21:56:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order Roedy Green wrote:
> is in the arrow telling you how the sort is NOW or how you want it to
> be if you click it?

How it is now.

> SAP uses a red triangle array pointing up means the data are sorted
> ascending NOW.

Then different SAP applications do different things. E.g. just a few
hours ago I used one which doesn't indicate any sorting direction at all.

> I would propose to use an unambiguous convention, maybe
> with 3 triangles of increasing size pointing left to right
> for natural sort order, and the vertically flipped image
> for the reverse order. This assumes a left-to-right writing
> convention, of course :)

IMHO you are looking for problems where there aren't any.

/Thomas

--
The comp.lang.java.gui FAQ:
ftp://ftp.cs.uu.nl/pub/NEWS.ANSWERS/computer-lang/java/gui/faq
http://www.uni-giessen.de/faq/archiv/computer-lang.java.gui.faq/
 
 
Roedy Green





PostPosted: 2006-2-20 22:34:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order On Mon, 20 Feb 2006 14:55:37 +0100, Thomas Weidenfeller
<email***@***.com> wrote, quoted or indirectly quoted someone
who said :

>> I would propose to use an unambiguous convention, maybe
>> with 3 triangles of increasing size pointing left to right
>> for natural sort order, and the vertically flipped image
>> for the reverse order. This assumes a left-to-right writing
>> convention, of course :)
>
>IMHO you are looking for problems where there aren't any.

Personally, I have trouble with all apps with such arrows. I always
end up studying the data to figure out the convention every ruddy
time.

In one app I wrote, I used an arrow up to the right to suggest data
are sorted in ascending order now.

It probably should be a LAF thing and something configurable so that
users could get all their apps to be consistent. I wonder if different
cultures have different "obvious" ascending/descending conventions.

--
Canadian Mind Products, Roedy Green.
http://mindprod.com Java custom programming, consulting and coaching.
 
 
Roedy Green





PostPosted: 2006-2-20 22:46:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order On Mon, 20 Feb 2006 14:43:37 +0100, Thomas Weidenfeller
<email***@***.com> wrote, quoted or indirectly quoted someone
who said :

>Down.

I know of 5 associations for up and 3 for down. Then of course you can
reverse that if you think of the icon as an operator rather than an
indicator.

You run into similar ambiguity problems with toggling buttons. Are
they labelled with the current state or the new possible state?

to me a button is a command, not an indicator. You should use
checkboxes or radio buttons for when you want to indicate current
state as well as allow commands.
--
Canadian Mind Products, Roedy Green.
http://mindprod.com Java custom programming, consulting and coaching.
 
 
Oliver Wong





PostPosted: 2006-2-21 0:17:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order
"Roedy Green" <email***@***.com> wrote in
message news:email***@***.com...
> Which way should an arrow point to indicate an ascending sort or an
> ascending list follows?
>
> For an ascending sort the bigger numbers are at the bottom.
>
> So no matter which way you do it, you can misinterpret.
>
> Is there some other less ambiguous convetion to use?

Use a parallelogram for which one of the parallel edge is so short, it
becomes a triangle. A Parallelogram shaped like V has a "big edge" at the
top and a "small edge" at the bottom. This indicates that big values are at
the top, and small values are at the bottom. You just need to define the
concept of "big" and "small" for various types of data. 9 is bigger than 1.
January 2000 is bigger than January 1900. Z is bigger than A. Etc.

The inverse sort is ^, with a small edge at the top, and a big edge at
the bottom.

- Oliver

 
 
Oliver Wong





PostPosted: 2006-2-21 0:22:00 Top

java-programmer >> Inidicator Arrow convention for Sort Order
"Roedy Green" <email***@***.com> wrote in
message news:email***@***.com...
>
> There are other sources of ambiguity:
>
> is in the arrow telling you how the sort is NOW or how you want it to
> be if you click it?

A button should be labelled such that what the label says becomes true
once the button is clicked. A label should be labelled such that what the
label says is true now.

I think most people consider table headers to be labels which react when
clicked-upon. So the table headers should show what the sorting is now.

>
> Yet another source of ambiguity. When you write elements you start at
> the top and work down. When you look at a pile, you think of starting
> at the bottom and working up.
>
> Does the arrow represent the direction you read to see ever bigger
> entries or does it represent the order the computer went though the
> list when it was writing the page top to bottom?

[...]
> here are two other conventions from Oliver Biot.
>
>
> I would propose to use an unambiguous convention, maybe
> with 3 triangles of increasing size pointing left to right
> for natural sort order, and the vertically flipped image
> for the reverse order. This assumes a left-to-right writing
> convention, of course :)
>
> Another possibility is to write a->z (z<-a) for text columns
> and 0->9 (9<-0) for numeric columns. We could do a similar
> thing for date/time columns.
>
>
> Because there are so many ways to interpret this, perhaps trying to
> indicate direction will inevitable confuse folk who presume the other
> convention. perhaps a colour code.

See my parallelogram post elsewhere in this thread. It makes no
assumptions about reading direction conventions or anything like that. It is
a pictogram which represents the state of the sort order. That is, I'm
confident that an extraterrestial alien (assuming they have the mathematical
concepts of geometry and ordering of elements), encountering this pictogram,
could accurately guess at its meaning, after interacting with the program
for a bit.

- Oliver