dsandler.org/wp

From abm:

Amusing ad in a local newspaper:

“Alta Mesa Memorial Park
lot #192 1800/obo Moved!! Would like to sell”

Does anyone else find this as funny as I do?

I guess everyone really is moving out of this area.

You know, of all the people I know, only you read the burial plot
classifieds.

Ice halos tonight.

More pain: the source file I’m working on is 5206 lines long. The function
I’m debugging is 574 lines. (All the source files and all the functions are
like this.)

Ughhhh — integer size issues. Some idiot is putting 0x0001000 into my
32-bit integer, thinking he’s writing 1 (0x0001) into a 16 bit slot.

So it seems that Dan Bernstein (author
of qmail, the mail server used at
dsandler.org) has developed an
exponential improvement in number-factoring circuits. This has caused me
to go read up on related papers of
his, as well as differential
cryptanalysis (more links.

(Tidbit for those unfamiliar: When IBM was developing DES a while back, the
NSA stopped by and asked them to remove certain key possibilities from DES.
IBM complied, and for twenty years it was generally assumed that the NSA had
added a backdoor to DES — in other words, “please don’t use these
keys, because we can’t crack them; use these other keys instead, which we
can break at any time.” Twenty years later (early 1990s), a new kind
of cryptanalysis was discovered that is particularly potent on certain
regions of the DES keyspace — specifically those keys the NSA said to
avoid. So it turns out that (a) the NSA was strengthening DES by
disallowing use of keys that would be proven weak by later research … and
(b) that the NSA had differential cryptanalysis 20 years before anyone
else.)

Needless to say, this math is way beyond me. But it’s really
interesting.

Mmm, mini-muffins!

My manager invoked the 90-90
rule in business communication today. Keen.