Re: self-loops in graphs

[Michael Piotrowski]

> Jeff,

> [lots of helpful stuff]
> Next, the margins look strange but I'd have to fiddle with it
> to see why they worked at all.  The summary at the end of the
> diagrams chapter shows exactly how circles fit over things.

Following Uwe's suggestion, I arrived at
            
            @Circle margin {0.2f} @Circle margin {0.4f}
            
which looks much better.  I'm sorry, but I can't find the information you're 
referring too (I'm probably too tired or too stupid, or both).

> He also wrote:
> 
>   The way I coded the start state is clumsy, ...
> 
> I can see two clumsinesses.  First, the invisible node that the
> arrow starts from.  I think this is basically inevitable; every
> arrow has to start from somewhere.  But it could perhaps be
> packaged up in a definition such as
> 
>   extend @DiagramPrint @Diag
>   def @HereIs right label {
>       label:: @Node outlinestyle { noline } margin { 0c }
>   }
>
> So that you can write
> 
>   @HereIs Start

Sounds good, too.  I'll try this and compare it with Uwe's approach.
 
> anywhere you like.  The other clumsiness is using @DP and || for the
> layout of the nodes.  As the chapter says, node layout is @Diag's
> weak point, but I do think using @Tab is a better option here, since
> it gives you some kind of structure that things can be added to.
> There is an example of the use of @Tab in diagrams in the @Diag chapter.

Yup, I've read this, and I had already planned to use @Tab for more serious 
stuff (see, I do read the manual :-)

Thank you very much for your time and support.  I hope I'm not bothering you 
too much and that I'll eventually learn the ropes of Lout (heck, I just had to 
learn a very ideosyncratic language for writing Left-Associative Grammars on 
which I'll be examined on Monday--I should be able to learn Lout, too :-).  
It's definitly worthwhile, I think.

Thanks again

-- 
Michael Piotrowski                         <address@hidden>
                              <http://www.linguistik.uni-erlangen.de/~mxp/>
Department of Computational Linguistics --- University of Erlangen, Germany