1

I am in the middle of writing something exposing various interactions between various properties that have certains objects in some theory in mathematics. I would like to summerize this in a graphe, and I would like it to look like this :

enter image description here

I started to try to do this (without the ellipses...) with the package xypic (that I already use for my diagrams) thanks to the following example stated in the documentation of the package :

\entrymodifiers={++[o][F-]}
\SelectTips{cm}{}
\xymatrix @-1pc
{
    *\txt{start} \ar[r]
    & 0 \ar@(r,u)[]^b \ar[r]_a
    & 1 \ar[r]^b \ar@(r,d)[]_a
    & 2 \ar[r]^b
    \ar ‘dr_l[l] ‘_ur[l] _(.2)a [l]
    &*++[o][F=]{3}
    \ar ‘ur^l[lll]‘^dr[lll]^b [lll]
    \ar ‘dr_l[ll] ‘_ur[ll] [ll]
}

but I am really not happy with it.

So I wonder if a package devoted to the design of intricate graphs as the one in the graphical example does exist, or at least something apporaching. If not, what would tex-gurus advise in my case ?

I must precise that I would like to avoid using tikz (except if there a dedicated "module" of it answering to my question, but I don't think so) nor any kind of really graphism-oriented latex package if possible, and if the eventual other solution produces diagrams of quality compared to those that tikz would produce.

15
  • 1
    Best strategy for such a diagram? Graphviz, definitely. In addition, its output can be integrated in LaTeX documents. Have a look to Drawing relationships between elements of a database. Mar 7 '15 at 17:53
  • 1
    Not that I'd especially recommend it for this but what is supposed to be poor about the quality of TiKZ diagrams? [I think it would be very annoying to use it for this but, if you did, I can't see why you should get a high quality result. It just wouldn't be a very efficient way to produce it unless you were just very familiar with TiKZ and so used it rather than learn something new.]
    – cfr
    Mar 7 '15 at 18:21
  • 1
    Finally, after investigating the "market" a bit more, I opted for pstricks. It could appear complicated, but it's not that much complicated, and I will be able to work in greater generality with it, so that it is perfect.
    – Olorin
    Mar 7 '15 at 22:35
  • 5
    I'm voting to close this question as off-topic because it was solved in the comments by the OP using a different graphics package. Mar 24 '15 at 12:15
  • 1
    Well, one can enter whatever he/she wants as a close reason. The alternative is for you to self-answer your question, because without that it won't be useful to future visitors, hence why I voted to close. Mar 24 '15 at 12:20

Browse other questions tagged or ask your own question.