New dissect 3.0.7 fixes the problems reported earlier.
(*/ % #)&.:*: i. 3 3 3
turned out to reveal all sorts of problems.
I have added tooltips, with some dinky initial text. I would appreciate
ideas on what the tooltip should say to be useful.
If anyone wants to experiment with colors, just edit the source, looking
for SHAPECOLORS. That's the table of colors used to distinguish
selection levels. The same colors are used for data (in DATACOLORS),
except that the main data background (the zeroth color) is forced to white.
You may change SHAPECOLORS and reload, but you will need to change the
font size to see the new colors.
Bob Therriault suggested that the colors rely more on luminance and less
on chrominance. I started along that path, but on my monitor green=255
displays the same as green=201, i. e. I have some whacked-out gamma
curve, and I just gave up and went with saturated colors.
I fondly hope that users will be able to guide me into making the
interface friendlier.
Henry Rich
On 8/8/2014 8:20 PM, Raul Miller wrote:
Ok... colors...
I tend to favor deeper, richer colors (though of course contrast is also
important).
Colors alone are inadequate though - I think additional information should
be made available on hover (mouse over an element after a short delay) -
perhaps as a tooltip or in some reserved area of the display (either choice
can be annoying, but mostly if you want to use that part of the screen for
something else).
Note that position and font attributes can also be used. And, of course,
both foreground and background colors can be used.
But honestly, that's all meaningless. To get colors right, I would want to
experiment with them until I found a set that I liked. And, for that, I
guess I should have in mind some examples that show all the different
possibilities.
Does anyone have a representative set of test sentences to consider??
Thanks,
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
