Hello,
Sorry for the change in behaviour, perhaps I should write somewhere that
'f' must be symmetric. I could do so in #19390, for it actually cleans the
constructor code and their documentation.
So should the "edge-detection" function be required to return true for any
> pair of vertices joined by an edge, or should there be an edge constructed
> between any pair of vertices for which the edge-detection function returns
> true?
>
The best for me is to request that the function should be symmetric (as it
is required in theory anyway: a graph is the result of a symmetric binary
relation), and that anything else is "undefined behaviour". This way
building a graph calls f for a total of n(n-1)/2 times (if there are no
loops), and not twice that amount.
Furthermore, I wonder where that change came from. It "may" come from the
graph code but really I do not see how: the constructor was patched heavily
recently, but that's not in the beta yet. itertools.combinations always
respected the ordering you give it: if 0 comes before 1 in the list, it
will give you (0,1) (thus the -1 in your code) and not (1,0). I don't think
that [0..8] ever returned [8,7,6,....,0] either. Weird.
I'd argue the former requires some help with the documentation, while the
> latter would be a change to the code/behavior.
>
I will add a "symmetric" somewhere in the doc. The amount of of guessing
that the graph constructor does is already beyond sanity. Just a funny
example I showed to a colleague yesterday:
{1 : [2,3], 2: [3, 4]} # we all think of the same graph, even if 2 does not
name 1 as its neighbors
{1 : [2,3], 2: [3, 4]} # with multiedges=True. Same result ? What do you
expect now ?
{1 : [2,3], 2: [3, 4], 3: [1]} # with multiedges=True. And now ?
{1 : [2,2], 2: [1]} # with multiedges=True. And now ? How many edges
between 1 and 2 ?
More headaches in the two comments starting
at: http://trac.sagemath.org/ticket/19385#comment:17
Have fuuuuuuuuuuun,
Nathann
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
