Re: [graphviz-interest] forcing distinct/more ranks?
"Emden R. Gansner" <[email protected]>
| Newsgroups | gmane.comp.video.graphviz |
|---|---|
| Message-ID | <[email protected]> |
On 2/11/13 5:21 AM, paul womack wrote:
> I thought of this, but didn't want to force "white space" in years where there
> are no data-based nodes.
You only need to generate ranks for years for which there is real data, as in the shells example.
> Ah - perhaps "rank = same" overrides.
Yes, that what I meant by attributes that turn off the top-down constraint. The constraint attribute and
clusters in conjunction with the old ranking function are other ways that this constraint can be
overridden.
>
> (indeed, the presence of those invisible edges is most odd)
Normally, A->B positions B lower than A, but if A and B are forced onto the same rank, A->B turns into
a left-right constraint. In the shells example, however, I agree that these additional constraints
are probably
unnecessary.
Emden
_______________________________________________
[email protected]
http://lists.research.att.com/mailman/listinfo/graphviz-interest