# IRC log for #marpa, 2015-09-20

All times shown according to UTC.

Time Nick Message
01:13 ronsavage joined #marpa
01:14 ronsavage jdurand: Re http://irclog.perlgeek.de/m​arpa/2015-09-19#i_11246334. I have noted it, but it'll only appear on my website Marpa page when I fix some illegal utf8 in the XML file which I use to generate the website :-(.
04:19 idiosyncrat joined #marpa
04:19 idiosyncrat left #marpa
04:41 idiosyncrat joined #marpa
04:41 idiosyncrat jdurand: nice
05:03 idiosyncrat left #marpa
07:21 ronsavage joined #marpa
08:57 lwa joined #marpa
09:00 jdurand The IRI counterpart uploaded to CPAN: https://github.com/jddurand/MarpaX-RFC-RFC3987
14:09 ceridwen joined #marpa
18:30 jdurand_ joined #marpa
20:02 ernimril_ joined #marpa
20:38 idiosyncrat_ joined #marpa
20:50 idiosyncrat_ The difference between my Marpa theory papers and tutorials, is that instead of examples I give proofs.
20:50 idiosyncrat_ Which certainly makes tutorials preferrable for a lot of purposes!
20:50 idiosyncrat_ A few words about proofs:
20:51 idiosyncrat_ 1.)  In reading a math paper, even experts usually skip the proofs on 1st reading.
20:53 idiosyncrat_ 2.) A proof is not necessarily this complex thing, though it certainly can be.  Tarski defines a proof as "some consideration intended to show the truth of a statement".  That's all a proof is.
20:54 idiosyncrat_ 3.) The proofs are statements as to *why* I believe something to be true, and why you should.
20:55 idiosyncrat_ Doing proofs have a number of advantageous effects:
20:56 idiosyncrat_ With parser, there is no way to test every possible kind of combination of syntax.  To establish that your algorithm works in general, testing is *not* adequate -- no test suite can be good enough.
20:57 idiosyncrat_ Proofs are important when I consider changes to Marpa -- will this change break things, I ask? ...
20:57 idiosyncrat_ First thing I do when considering a parse engine change is re-read my proofs.
20:58 idiosyncrat_ The proofs are *why* I believe Marpa works, and any change that invalidates the old proofs beyond repair is too dangerous to implement.
20:59 idiosyncrat_ btw, restating 3.) above:
21:00 idiosyncrat_ 3.) The proofs are statements as to *why* I believe something to be true, and why I believe that you should
22:09 MerittClarksdale joined #marpa
22:12 MerittClarksdale Greetings
22:13 MerittClarksdale What are your views regarding Clang?
22:18 idiosyncrat_ I think clang & LLVM sound like great ideas.
22:40 ronsavage joined #marpa
23:19 MerittClarksdale Does their parser design meet your approval
23:57 ronsavage I've updated Marpa's homepage with links to J Durand's 2 new Marpa-based modules (URI, IRI). Also, check the module review page (http://savage.net.au/Module-reviews.html) for Buddy Burden's new article on date processing modules.