First, in my EE work, jumps are old fashioned. Instead, crossed lines means
no connection, whereas a T with a dot means a connection. Full crosses with a
dot are ambiguous:
|
-------|-------- no connection
|
-------o-------- connection
|
|
|
|
-------o------- ambiguous
|
|
Second, autorouting is a tough problem. There is a general field of graph
layout, with ideas like springs, swimlanes, etc. "dot" is a candidate. Also,
it is akin to path-plannning (as in robotics), with treatment of constraints,
A* algorithm, et al. So it is a worthy challenge for someone, sometime. I
just want Dia to handle the normal stuff for now. And to get my hands on the
python scripting stuff (which is where anything like autorouting should be
done).
> Old-Return-Path: <[EMAIL PROTECTED]>
> Sender: [EMAIL PROTECTED]
> From: Lars Clausen <[EMAIL PROTECTED]>
> Date: 14 Jan 2001 08:35:46 -0600
> User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7
> Content-Type: text/plain; charset=iso-8859-1
> Resent-From: [EMAIL PROTECTED]
> Reply-To: [EMAIL PROTECTED]
> X-Mailing-List: <[EMAIL PROTECTED]> archive/latest/3189
> X-Loop: [EMAIL PROTECTED]
> Precedence: list
> Resent-Sender: [EMAIL PROTECTED]
> X-MIME-Autoconverted: from quoted-printable to 8bit by krim.seanet.com id GAA15170
> X-UIDL: 924f39cb3239e53417beb1eb58cec7d8
>
> On Sun, 14 Jan 2001, [EMAIL PROTECTED] wrote:
>
> > Fair enough, I shall wait. If I were in the position to lend some code,
> > I certainly would!
> >
> > From a futures perspective, I would vote against smart-routing of lines
> > as implemented in Visio. It has been my experience that the smart router
> > is more of a bother than a help; I'd much rather place the lines myself
> > and not come back to find that the tool as rearranged them for me. This
> > is particularly annoying when one makes a slight move to one object and
> > the router then disturbs a number of unrelated connections. If you do
> > build such a router, perhaps it could be an option that could be enabled
> > or disabled by the user.
>
> I would certainly want a lot of manual control over such a system. I
> imagine something like this: With Autorouting turned on, Dia finds an
> "optimal" route when a line is created. After that, the line is normal.
> Then you can choose to re-Autoroute them later. Also, there are probably a
> bunch of parameters to the routing algorithm that you could twiddle with.
> Even if we do have 'continuously autorouted' objects, there should probably
> be a treshold of change before they are re-routed. Thus it wouldn't
> reroute until the current route 'looks' really bad. But until we have a
> routing algorithm, we can't really say much about this.
>
> -Lars
>
> --
> Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause) | Hardgrim of Numenor
> "I do not agree with a word that you say, but I | Retainer of Sir Kegg
> will defend to the death your right to say it." | of Westfield
> --Evelyn Beatrice Hall paraphrasing Voltaire | Chaos Berserker of Khorne
>
>
--
Harry George
[EMAIL PROTECTED]
