Front page | perl.perl5.porters |
Postings from August 2000
Jeffrey Friedl wrote: > Ilya Zakharevich <ilya@math.ohio-state.edu> wrote: > |> You know too much. In this whole discussion you refuse to *unlearn* > |> your knowledge of NFA and backtracking. > -snip > > Even if I unlearn backtracking, it won't remove the fact that the semantics > of Perl regular expressions have always worked this way, that perl regular > expressions are defined to work this way, and that people expect them to > work this way. Changing the basic semantics would break a *lot* of code. > > Maybe what I don't understand is that there's some way to redefine the > semantics as radically as you seem to want to, yet without breaking code. > If that's the case, then I'm very much wanting to understand it. Please > help me to understand it. i am biased, since i learned re's from reading jeffrey's book. when my re's do not do what i expect them to, i analyze them thinking in terms of backtracing, step by step, the way i believe the regex engine is working. if someone has a technique which is easier to understand and works as well, i would be happy to see it. then, lets try it out on several test cases to make sure it holds up. just about anything works on a simple case, but only the best techniques work when things get complicated. so my vote, if i have one, is to leave backtracing in the docs. when a better, tested solution is available, we can replace backtracing in favor of it. > > > If that's the case, indeed, I'll try very hard to shed preconceived > restrictions, but frankly, Ilya, it would be much easier if there was less > hostality in your repleies. I don't know if you feel somehow "threatened" > because I'm fishing in your pond, or you just don't like me because I wrote > a book that has a lot of pages about backtracking, or what, but sometimes > it's hard to see through the hostility to the point you're trying to make. > I have a pretty thick skin, but since I've always shown you respect and > don't feel I've ever given you any reason to not to show the same for me, > I've been pretty shocked at the immediate and raw distain you've shown to > me since I came back to p5p. (I still look forward to the technical > information in your notes -- it's just the attitude that they're wrapped in Ilya appears very bright when it comes to math and re's, very weak when it comes to dealing with people and garnering support for this ideas. he could probably contribute 10 times as much if he learned to explain his ideas in a positive fashion instead of a negative one. my advice, just pick through his postings to find the kernels of useful information, and ignore the commentary and negativity. please do not let him disuade you on your task of bettering perl's re's and their associated documentation. -- ___cliff rayman___cliff@genwax.com___http://www.genwax.com/