00:01 <DocScrutinizer05> I'm sure a DamnSmallLinux will run happily as DHCP/NAT on that thing, even with its 128MB RAM

00:03 <DocScrutinizer05> just the ethernet is crap, damn

00:04 <DocScrutinizer05> maybe 2 PCI cards will do

00:05 <DocScrutinizer05> roh: how's about your milling skills and opportunities?

00:37 <wpwrak> DocScrutinizer05: considering to go from "case buyer" to "case maker" ?

00:37 <DocScrutinizer05> kinda

00:37 <DocScrutinizer05> spacer layer

00:37 <DocScrutinizer05> 3mm ABS

00:37 <DocScrutinizer05> should be "easy"

00:39 <wpwrak> constant height ? or with variations along the Z axis ?

01:49 <roh> DocScrutinizer05: depends... rather a question of mounting it

01:50 <roh> dunno about abs.. never milled it. what about delrin? (POM)

02:12 <DocScrutinizer05> whatever works

02:12 <DocScrutinizer05> wpwrak: needs some 3D modelling as well

02:14 <roh> if you have any cad data, send a mail, and i'll see whats possible or if it needs more work

02:23 <wpwrak> yeah, i can generate STL now ;-) not entirely sure if it's correct - in meshlab, the colors don't look right, but that could be just meshlab having a problem

02:24 <wpwrak> this one: http://downloads.qi-hardware.com/people/werner/anelok/tmp/top-mesh.png

02:24 <wpwrak> http://downloads.qi-hardware.com/people/werner/anelok/tmp/top-mesh.stl.bz2

02:25 <wpwrak> this is something i can run on my mill. just need to find time to go buy some more acrylic - this needs 7 mm and i only have 5 mm at home

04:21 <roh> wpwrak: / DocScrutinizer05 upload to camgeeks.de - recieve a quote ;)

04:23 <roh> .s

09:06 <wpwrak> roh: your 3D prints look surprisingly smooth. they didn't come out of the machine like this, did they ?

10:48 <roh> wpwrak: most of them did.. which picture exactly do you mean?

12:47 <wpwrak> both the white holder on the pole and the "OH". that's really without surface treatment ?

14:18 <roh> wpwrak: basically yes

14:18 <roh> i'm not sure these pics were printed here, but there is so much going on that its hard to keep up

14:19 <roh> but yes, thats both fdm stuff without chemical rework

14:19 <roh> usually you use a cutter and remove some supportstruts which need to be made to build overhangs and such stuff like stringing, but the surfaces get quite nice

14:20 <roh> depends a lot on layer thickness setting s etc

15:50 <piyush_prl> sir plz help me out

15:51 <piyush_prl> today my pen driver got wet in water

15:51 <piyush_prl> what to do

16:00 <kyak> make it dry

16:40 <DocScrutinizer05> wtf is a pen driver?

16:59 <DocScrutinizer05> pen drive == flash stick?

17:00 <DocScrutinizer05> we'll prolly never know. (s)He's gone

17:12 <wpwrak> penetration drive ? maybe some kind of drilling machine ?

17:14 <DocScrutinizer05> I already wondered if that's been a carefully crafted social bot standard opening

18:55 <wpwrak> just had an idea for improving random number generation: since we have nearly infinite storage (in relation to what we actually need), let the user add a file with random bits. then use bits from that file.

18:56 <larsc> rand() { return 4; /*choosen by fair diceroll */ }

18:56 <wpwrak> (and mix them into the RNG we already use. so even if that one is - unknown to us - is compromised/weak, we can still get decent entropy)

18:57 <wpwrak> well, worst case is that the user gives us a string of zeroes or such. in which case we're not worse off than with using the existing RNG.

18:58 <larsc> how hard is it to integrate a real hw rng?

18:58 <wpwrak> there's one in the transceiver

18:58 <larsc> but it's too slow?

18:59 <wpwrak> it's about 1 Mbps. that should be sufficient, i guess.

19:00 <wpwrak> the question is whether it's truly random. there's always the risk that a RNG is biased, be it by defect, by design flaw, or by NSA.

19:01 <eintopf> you need a webcam and a lavalamp

19:01 <eintopf> for a hw random generator

19:01 <roh> does that really matter on such a soc?

19:01 <roh> with no real protetion against hw level attacs

19:01 <roh> +c

19:02 <wpwrak> you could still attack things other than the physical device. e.g., RF or backups of your password data

19:03 <roh> jup. so at some point it makes no sense making one part of the crypto toolchain perfect, as long as the other parts aren't on the same level

19:03 <wpwrak> you can also harden physical security. e.g., by adding a seal between top and bottom shell.

19:04 <roh> i guess as long as the rng is working well and has no known or measureable defects one would need to worry about physical security measure before going further

19:04 <wpwrak> attackers may not have uniform preferences for their attacks

19:05 <wpwrak> and you could always wear your anelok on your body, making it difficult to mess with it. leave it in a safe when you go to the gym :)

19:08 <wpwrak> (attackers) e.g., a pickpocket in the subway will find a physical attack easy but will probably despair on the crypto. on the other hand, the autistic blackhat somewhere in his mother's basement in utah is very unlikely to even want to leave the house, let alone travel to berlin to steak your anelok

19:09 <wpwrak> but he may have all the resources needed to break your weak IVs

19:09 <wpwrak> (plus the knowledge that goes with it)

19:13 <wpwrak> what doesn't make sense is to add weak obstacles in paths where attackers will already have overcome much stronger obstacles of a similar nature. so don't rot13 your twofish-encrypted conspiracy memoirs

19:15 <wpwrak> that is, sometimes even such a thing could help, if the attack is not against elements in the crypto chain but the entire chain. e.g., if the attack is based on guessed plaintext, a rot13 may shift you out of the profile the attacker assumes

19:16 <wpwrak> of course, once they find out that you were using rot13, it's trivial for them to generate another set of vectors that will fit this time

19:17 <roh> well.. if you got any files for us to print or try mill...

19:20 <wpwrak> for now, i'll try to make do with what i can make at home. i expect this to be good enough for prototypes quantities. i can even run small batches of a few dozen units, if needed. my mill doesn't mind working at night ;-)

19:21 <wpwrak> of course, there are numerous limitations to that approach. but after that, i need to look into higher-volume processes anyway.

19:28 <roh> heh. i just got spammend by http://www.dgmtmold.com/ some weeks ago

19:28 <roh> how fitting

19:33 <wpwrak> yeah, that's the sort i'll eventually want :)