[antlr-interest] Decision DFA w/ 51 Nodes & 133 Arcs — Should I Be Worried?

[Randall R Schulz]

Hi,

I'm concerned about what appears to me to be an outsized DFA for a 
particular production in a grammar I'm working on.

As the subject says, this DFA has 51 nodes and 133 arcs (and two accept 
states).

Is this a sign of a potential problem? Or do I just have an invalid 
intuition about typical complexity and efficiency for LL(k) grammar 
DFAs with backtracking?

Should I be looking at elaborating the grammar to reduce the complexity 
of the DFA generated by this production?

Conversely, if I were to turn off backtracking and manually address the 
ambiguity, would I just end up producing a lot of rules of comparable 
complexity to this DFA?

I went systematically through the grammar viewing the decision DFA for 
each production, and all the others are miniscule by comparison.


Thanks.


Randall Schulz