Nov. 24th, 2005

piranha: red origami crane (Default)
i'm pulling this from the comments because i was unclear enough to not want to continue there.

several things:

i know from reading right-wing blogs that the things bill named are all possible motivations for continuing to support this sitting president. while the approval ratings for bush set me thinking about this, it's not really the focus for what i've been pondering.

frex, somebody who voted for X because zie can't conceive of a person from the other party doing enough for the country should theoretically answer differently to "is X doing well" as opposed to "do you support X". i can believe a president is making lots of bad decisions, but yet could still be a better choice than all the others; when i am asked whether i'm satisfied with zir decisions i'll still say "no". but lots of people don't seem to do that. it's kinda hard to follow though, what with polls all wording questions in different ways.

i am interested in this phenomenon beyond the specific instance of bush (or nixon, who sunk a bit lower than bush now, but still had more than a quarter of the voters behind him) -- such as what people reported in the comments to james' entry: that roughly 25% of people voted for those whom they knew nothing about. approval ratings in the US apparently never sink much below 20%, but they do so in canada and other countries -- what's the difference?

snippy suggested there was money for polls (there is? where? this is the first i've ever heard of it.) and asked whether it'd be so hard to question people about their actual motivations. i do think it'd be pretty hard; writing good polls is more difficult than it might at first appear. i learned about that when evaluating prototype test experiences at ETS -- *gah*; that was excruciating. i sometimes think it'd be more informative to just ask people open-ended questions and codify them afterwards, but it's not like that is easy. and open-ended questions discriminate against people who don't express themselves well verbally -- and those might be precisely the sorts of people i'd want to hear from. hilzoy wrote a long post about clear thinking which i still want to mine (but not link to because it's again more bush-centred than what i'm pondering).

to really do this well, one has to conduct a controlled experiment from start to finish -- run a fictitious candidate, let people vote, and question the voters afterwards as to their motivations. it'd be a nice sociology project; but way too much effort for a regular person. and i think this sort of thing has to have been done several times over; political scientists would want to know about this in a big way. i continue to hope for links from the lazyweb. :)
piranha: red origami crane (Default)
with fireHOL, a language to express firewalling rules. it produces a stateful firewall. (it keeps track of the events in any particular sequence of interaction, which means it knows which packets are legitimate for any specific connection state, and can make intelligent decisions about which to keep and which to reject).

it is easy to write. it is even easy to read. look how damn easy it is:

    interface eth0 homelan
    policy accept
    
    interface eth1 internet
    protection strong
    server http accept
    server ssh accept src trusted.example.com
    client all accept
    
    router homelan2internet inface eth0 outface eth1
    masquerade
    route all accept

my actual firewall is of course more complicated, but it didn't take more than 30 minutes to learn the language and write out a scipt (fireHOL uses bash scripting). and it would have worked perfectly right away were not telus blocking standard bittorrent ports, the bastards. but at least that became obvious pretty much right away, while before using fireHOL i wasn't certain whether i was handling the forwarding of the bittorrent traffic correctly (as it turns out, i wasn't).

so now i am happy. i have a nice, new, clean firewall script that i actually understand backwards and forwards.

Profile

piranha: red origami crane (Default)
renaissance poisson

July 2015

S M T W T F S
   123 4
567891011
12131415161718
19202122232425
262728293031 

Most Popular Tags

Expand Cut Tags

No cut tags