[jdev] A rapidxml fork for XMPP

[Alexander Holler]

Am 05.12.2013 07:15, schrieb mat henshall:

> extremely fast and efficient on small devices. BTW the trick to avoid the
> tag comparison is to convert the various tags and namespaces to hash values
> as they come in during parsing and then only compare on the hash value. We

Hmm, I would have assumed most servers already do convert the tags and
namespaces to something else to not have to handle strings afterwards. I
don't know If there is a big difference to calc a hash or to compare the
strings once might make a big a difference in the overall picture (if 
they are sorted by frequency, comparing strings might be faster than 
calculating the hash and then comparing unsorted hashes).

In fact I never cared much about the necessary time needed to do so nor 
did I care much about the time necessary to parse the XML. My question 
always was, why should these steps be necessary at all. They are an 
unnecessary complication.

Maybe someone should develop an textural represention of the
ipv6-headers for ipv7 if plaintext is that comfortable and accepted as
it seems to be. (Just kidding) ;)

Regards,

Alexander Holler

Btw, there was a very interesting talk about hash-flooding at the 29c3. 
Video and papers are online available.