On Tue, Nov 23, 2010 at 2:58 PM, Ivica Ico Bukvic
<ico@vt.edu> wrote:
> Ok, this is looking familiar-- have a look at tidy-test.pd, select
> each group of objects one at a time and do 'tidy up':
> 1. Works as it should.
> 2. Objects are below the y-threshold, so they (understandably)
> collapse.
> 3. Works as it should.
> 4. Objects spaced 30 px apart vertically seem arbitrarily to no
> longer get tidy.
> 5. Horizontal row of objects gets tidy (with arbitrarily large space
> in between any two objects). Works as it should-- columns of objects
> should behave the same way.
>
> Now select all and tidy. Notice that #3 now gives different results.
>
> -Jonathan
Interesting... It seems to me tidy would work better if it simply
checked first if it had objects that are vertically or horizontally
closer to its x,y position and then adjust according to them only,
rather than keeping a HIST of values it may be able to potentially
adjust to. This would of course require change to the algorithm's
fundamentals. Another thought would be to have the whole thing keep
running with ever-increasing thresholds until something is adjusted or
some ridiculous threshold is reached.