Semantic Web Interest Group IRC Chat Logs for 2003-07-06

08:39:31 <danbri>http://rdfweb.org/people/danbri/2001/08/rpmfind/

08:39:32 <dc_rdfig> A: http://rdfweb.org/people/danbri/2001/08/rpmfind/ from danbri

08:39:44 <danbri> A:|Proposed fixes to RPMFind's RDF syntax

08:39:45 <dc_rdfig> Titled item A.

08:40:29 <danbri> A:I did this a couple years ago, talking to Daniel Veillard. I don't think anything happened to use it. Not sure RPMFind still uses this format at all now...

08:40:29 <dc_rdfig> Added comment A1.

08:55:53 <w1k1|afk> w1k1|afk is now known as w1k1

09:18:27 <danbri>http://www.lawver.net/archive/002785.html

09:18:27 <dc_rdfig> B: http://www.lawver.net/archive/002785.html from danbri

09:18:36 <danbri> B:|On the new AOL blogging tool

09:18:36 <dc_rdfig> Titled item B.

09:18:51 <danbri> B:" There are some really cool features in the product at launch. RSS support will be in 1.0, along with a bunch of other stuff that I?m not going to tell you about. Ok, I?ll tell you one? you can send an IM to a bot and have it post to your blog with rich text support and other cool stuff (like add titles, etc). That really blew a couple people away, even though I know bloggerbot kind of did the same thing in a limited (as I remember, fla

09:18:52 <danbri> ey). How cool is that?"

09:18:52 <dc_rdfig> Added comment B1.

09:19:06 <danbri> B1:" There are some really cool features in the product at launch. RSS support will be in 1.0, along with a bunch of other stuff that I?m not going to tell you about. Ok, I?ll tell you one? you can send an IM to a bot and have it post to your blog with rich text support and other cool stuff (like add titles, etc). "

09:19:06 <dc_rdfig> Replaced comment B1.

09:19:11 <danbri> B:"That really blew a couple people away, even though I know bloggerbot kind of did the same thing in a limited (as I remember, flakey). How cool is that?"

09:19:12 <dc_rdfig> Added comment B2.

09:19:18 <danbri> B:Chumping goes mainstream...

09:19:18 <dc_rdfig> Added comment B3.

09:27:51 <danbri> sbp`? any chance you could release your parser? http://infomesh.net/2003/rdfparser/ promises it last week...

09:27:57 <danbri> morning mortenf

09:28:06 <mortenf> morning danbri

09:30:03 <danbri> is http://bitsko.slc.ut.us/2003/06/foaf-check/rdf.py really a complete rdf parser? it seems... so... tiny...

09:30:12 <danbri> after all the fuss people make about our syntax ;)

10:43:50 <sbp`> danbri: release the parser: the one that bitsko released is broken

10:44:03 <sbp`> i.e. it doesn't work in some versions of Python

10:44:19 <sbp`> so I'm waiting on a response from him about it, but I'm inclined to just release anyway

10:44:22 <sbp`> but I want him to update his /rdf.py

10:44:32 <sbp`> and yes, it is a complete parser

10:44:36 <sbp`> (AFAIK)

10:45:22 <danbri> complete: nice work! :)

10:45:49 <sbp`> thank you

10:46:16 <JibberJim> Can you port it to javascript please...

10:46:19 <sbp`> hehe

13:32:51 <ericP> anybody know if the WG ever proclaimed whether

13:33:08 <ericP> @prefix a http://example.com/a

13:33:17 <ericP> @prefix ab http://example.com/ab

13:33:38 <ericP> @prefix owl ...

13:34:01 <ericP> a:bc owl:sameAs ab:c .

13:34:54 <ericP> well, i stated it was true, but really, i'm wondering if, given the a and ab namespaces above, a:bc is the same URL as ab:c

13:34:58 <ericP> anybody?

13:35:03 <ericP> beuller?

13:35:16 <mortenf> must be, since they're both http://example.com/abc

13:36:59 <ericP> so that would imply that an implementation need not keep track of the namespaces used to create nodes.

13:37:22 <mortenf> yeah, would also sometimes be impossible

13:37:29 <ericP> instead, just leave it up to an algorithm at the end

13:38:03 <ericP> impossible, yeah, xpointers are a pain in that regard

13:38:29 <mortenf> and use of duplicate prefixes - or none at all, i.e. rdf:type

13:57:07 * danbri catches up

13:58:09 <danbri> ericP, I think this is purely a matter of encoding syntax... RDF/XML uses XML namespaces... N3 has its own version for abbreviating, etc.

13:58:33 <danbri> By the time it gets parsed into triples, its all normalised

14:00:14 <danbri> I think there's a general sense that for any namespace uri which names a vocabulary, it is dumb / innacurate to claim that there is another namespace uri that is a left-stemmed substring of it (hmm or vice versa)

14:00:29 <danbri> ...even though all the syntax level stuff wouldn't notice the mistake

14:01:13 <danbri> Nothing anyway says it is wrong, its just a confusing sort of way to deploy namespaces so people don't do it

14:09:30 <tav> tav is now known as tav|offline

14:11:00 <ericP> sonever have a namespace that ends in chars that could be in the localName

14:11:10 <ericP> s/sonever/so never/

14:12:21 <ericP> there is no determination of what is at a namespace, but if there was, we would need to institute the above rule

14:12:51 <ericP> otherwise, someone might look for the namespace http://example.com/ or http://example.com/a or http://example.com/ab

14:24:36 <danbri> they could be three different namespaces. urn:x-nstest:example.com/ urn:x-nstest:example.com/a urn:x-nstest:example.com/ab could all be used for XML namepaces, and as containers for collections of RDF vocabulary terms...

14:24:48 <danbri> ...it would just be annoying and confusing

14:25:18 <danbri> ...and urn:x-nstest:example.com/ as a namespace couldn't contain something called 'a' or 'ab'

14:25:40 <danbri> there is some precdent there: the RDF M+S namespace contains both RDF vocab plus also XML-only constructs (rdf:Description etc)

14:26:17 <danbri> there is a half-hearted notion of hierarchy for some URI schemes, but it was never really explained clearly...

14:26:32 <danbri> ...except in terms of syntactic machiner, '../../' etc

16:28:47 <mdupont>http://gnufans.net/intrspctr.pl?RdfApi

16:28:48 <dc_rdfig> C: http://gnufans.net/intrspctr.pl?RdfApi from mdupont

16:29:02 <mdupont> C:| Musings about a c# rdf api

16:29:02 <dc_rdfig> Titled item C.

16:43:09 <mdupont> hey dajobe

16:43:13 <mdupont> hows things going?

16:43:23 <mdupont> can you review C: for me ? :)

17:32:36 <mdupont>http://demo.dotgnu.org/useradmin.dgmx

17:32:37 <dc_rdfig> D: http://demo.dotgnu.org/useradmin.dgmx from mdupont

17:32:53 <mdupont> D:| Mike duponts attemp at a dotgnu foaf webservice

17:32:53 <dc_rdfig> Titled item D.

17:33:10 <mdupont> D:|Mike DuPont's attempt at a dotgnu foaf xmlrpc webservice

17:33:11 <dc_rdfig> Titled item D.

17:33:33 <mdupont> D:currently it is only an interface

17:33:33 <dc_rdfig> Added comment D1.

17:33:46 <mea_cu|pa> mea_cu|pa is now known as mea_culpa

17:36:39 <danbri> D:Interesting. I like the idea of exposing getters and setters for person records this way (though I'd use SOAP instead of XML-RPC). I am concerned you'll have a maintainance nightmare on your hands, if you have to shadow each FOAF property with a web service method name. How about pushing the property into a parameter?

17:36:40 <dc_rdfig> Added comment D2.

17:37:17 <danbri> D:Ie. I might call setRepeatableProperty( http://xmlns.com/foaf/0.1/weblog, some-value)

17:37:18 <dc_rdfig> Added comment D3.

17:37:53 <danbri> D:Or setFunctionalProperty( http://xmlns.com/foaf/0.1/dateOfBirth, some-value) for properties that can only take a single value

17:37:53 <dc_rdfig> Added comment D4.

17:38:41 <mdupont> danbri: i am using xmlrpc, because dotgnu supports in nativly, check this : http://demo.dotgnu.org/~mdupont/useradmin.cs

17:38:44 <danbri> D:This would also allow non-FOAF properties to be set, in theory at least. All of this being for the simple/flat case of a string or URI-valued property. For fancier models, it gets fiddly...

17:38:44 <dc_rdfig> Added comment D5.

17:39:21 <mdupont> cool

17:39:35 <mdupont> i dont have all the the uml yet

17:39:43 <mdupont> i mean owl

17:39:55 <mdupont> i just extracted the rdftypes from the foaf

17:40:07 <danbri> mdupont, if you can generate it from the namespace, i guess that's not so bad...

17:40:12 <mdupont> yes,

17:40:15 <mdupont> that is the idea,

17:40:25 <mdupont> right now i am using this script :

17:40:45 <danbri> I have been meaning to take a look at the vocab to partition it more clearly into simple/flat stuff and things that need nested substructure

17:40:58 <danbri> 'knows' etc being in latter, as it typically references a bnode

17:41:33 <danbri> so naive question... how does the server know which person record to edit?

17:41:39 <mdupont> demo.dotgnu.org/~mdupont/makeclass.pl

17:41:42 <danbri> can you key into its person list in multiple ways?

17:41:53 <danbri> eg. homepage vs mailbox vs sha1sum'd mbox etc

17:41:54 <mdupont> and right now the implementation is based on the idea of a stringset

17:41:56 <mdupont> perl attribute

17:42:06 <mdupont> so you can add and delete strings from the set

17:42:21 * danbri nods

17:42:23 <mdupont> danbri: good question,

17:42:31 <mdupont> that is the idea of the user manager

17:42:38 <mdupont> the user manager can add and remove users

17:42:53 <mdupont> the user interface is only good for one user object per session

17:43:04 <mdupont> so, you would have a new method to delete

17:43:07 <mdupont> or a copy method

17:43:19 <danbri> do you have an implementation of the server side of this?

17:43:23 <mdupont> and basically, i image some session

17:43:26 <mdupont> no not at all

17:43:37 <mdupont> just wrote that, and am very very impressed how easy it is

17:43:48 <mdupont> check this makeclass.pl

17:44:00 <danbri> i looked :)

17:44:03 <mdupont> that creates the methods that create the entire xml source

17:44:25 <mdupont> the thing is that you can also get an account on demos.dotgnu.org and use it yourself:)

17:44:39 <mdupont> i am going to use this interface for people to request accounts

17:44:43 <mdupont> by submitting thier foaf

17:44:58 <mdupont> that will be a foaf at a time webservice later

17:45:09 <mdupont> with support for checking the signing of the foafs

17:45:10 <danbri> that's interesting... there's been a fair bit of discussion of foaf and access control in weblog land lately

17:45:28 <danbri> also bitsko's recent work w/ foafcheck (all linked from recent rdfweb.org articles someplace)

17:45:35 <mdupont> and registering users, and attaching webservices to the faofs

17:45:52 <mdupont> i think we can add foaf attributes to the dognu web attributes for more semantic webservices

17:46:08 <danbri> Have you seen the approach AndyS uses in Joseki, http://www.joseki.org/protocol.html for remote access to RDF stores?

17:46:10 <mdupont> like allowing a whole set of rdf attributes to be attached to webservices and webmethods

17:46:16 * mdupont looks

17:46:17 <danbri> I wonder if something similar might work for this app...

17:46:33 <danbri> ie use a generic RDF protocol rather than a foaf one, for extensibility

17:46:39 <mdupont> i want to support rdfstorages even in the remote via dotgnu.rdf

17:46:47 <danbri> yup

17:46:52 <mdupont> the idea is to build this all on dotgnu.rdf

17:47:05 <mdupont> and use various rdf data servers around

17:47:19 <mdupont> they could even communicate over xmlrpc.... scary thought

17:47:28 <danbri> :)

17:47:35 <mdupont> do you want an account on demos.dotgnu.org?

17:47:40 * danbri tries to find why Gnome doesn't work for him

17:47:49 <danbri> Sure! though I'm not sure what I could do with it yet...

17:48:01 <mdupont> danbri: i know that you have much experiences and idea about webservices and rdf

17:48:05 * danbri has pnet etc installed here, but forgets the helloworld basics of compiling, running...

17:48:13 <mdupont> this is just a spontenious idea

17:48:28 <mdupont> well, i am writing an article about how to make webservices on demo.dotgnu.org

17:48:32 * danbri has no experience of happy integration of RDF and SemWeb stack, just yet...

17:48:39 <danbri> some ideas tho ;)

17:48:43 <_joshua> i hate instant messaging.

17:48:45 <danbri> article: sounds useful

17:48:56 <mdupont> with the foaf service as the example

17:49:04 <danbri> hi joshua. thanks for sharing. does irc count as IM?

17:49:04 <mdupont> it will have a phpgw front end

17:49:08 <mdupont> and a perl client

17:49:18 <danbri> phpgw?

17:49:23 <mdupont> php group ware

17:49:24 <_joshua> I'm not sure

17:49:29 <mdupont> is part of the dotgnu vision

17:49:34 * danbri hates RSI

17:49:35 <_joshua> i'm trying to make all the instant messaging systems i use available on my danger.

17:49:46 <danbri> danger?

17:49:49 <_joshua> it's very frustrating, especially when people use more than one system

17:49:51 <danbri> a new gadget?

17:49:53 <_joshua> WearableGizmo

17:49:54 <danbri> yeah...

17:49:56 <danbri> ah

17:50:04 <mdupont> it will access the foaf service via xmlrpc

17:50:11 <mdupont> and present a html gui to the user

17:50:18 <_joshua> it speaks AIM, POP3 and HTTP

17:50:29 <mdupont> _joshua: sounds like you need jabber

17:50:31 <danbri> no imap...

17:50:42 <_joshua> so i can convert IRC to AIM no problem, but I am having problems with gale and zephyr

17:50:53 <_joshua> Jabber doesn't seem like the right answer

17:50:57 <_joshua> no, no IMAP. which blows.

17:51:15 <mdupont> _joshua: jabber is the answer to all questions imho

17:51:52 <_joshua> are you serious?

17:51:57 <_joshua> i can't connect to jabber with this

17:52:11 <_joshua> and i can't connect to gale or zephyr with jabber either

17:52:13 <mdupont> :(

17:52:18 <_joshua> so it's adding another piece

17:52:26 <mdupont> ok, you know better than me :(

17:53:17 <_joshua> i am either going to be forced to use AIM or email as the gateway to the device

17:53:42 <_joshua> also if i use the POP mail it only checks it every 10 minutes or so. AIM is instantaneous more or less, and so is the @tmail.com account

17:53:56 <mdupont> danbri: your right, i need to have some type of object selector

17:53:57 <mdupont> :)

17:55:13 <danbri> yeah, for object selection either (i) a URI (ii) a pair of an unambiguous property name and its value... should do. So a strcture which encompased both those options is a pretty handy thing to have in an rdf utility api

17:56:09 <mdupont> an uri

17:56:20 <mdupont> the uri of the foaf

17:56:46 <danbri> well for people, it would almost always be (ii), but i was thinking generic net api...

17:56:54 <danbri> joseki, TAP etc approach

18:01:16 <mdupont> hey JibberJim

18:02:00 <JibberJim> hey

18:02:37 <mdupont> so, i am very happy with svg

18:02:43 <mdupont> thanks for your support

18:29:59 <mdupont> hey AaronSw

18:30:05 <AaronSw> hi

18:30:33 <mdupont> D:is now updated to take the uri of the foaf to update as the first parameter of the faof webservice

18:30:34 <dc_rdfig> Added comment D6.

18:40:10 <mdupont> danbri do you have a ssh2 public key?

18:40:31 <mdupont>http://rdfweb.org/people/danbri/danbri-pubkey.txt

18:40:31 <dc_rdfig> E: http://rdfweb.org/people/danbri/danbri-pubkey.txt from mdupont

18:40:34 <mdupont> is that it?

18:40:42 <mdupont> oops

18:41:00 <mdupont> E:|danbris public key, sorry, please delete

18:41:00 <dc_rdfig> Titled item E.

18:42:02 <danbri> thats my pgp key...

18:49:40 <AaronSw> someone is spamming all the W3C mailing lists..

18:49:56 <danbri> more so than usual? :(

18:49:59 <AaronSw> I guess they found a way around the aa system

18:50:08 <danbri> got an example url?

18:51:43 <AaronSw> Oh, never mind, they're just spoofing the envelope sender to be various W3C mailing lists. odd.

18:52:53 <danbri> yeah, that's a pita

18:53:27 <danbri> but we shouldn't have w3c lists in the accepted poster lists (and I don't believe we do, by default unless they've been added by hand for some reason)

18:53:50 <danbri> ah maybe all @w3.org addresses excused from being on the aa list? I forget...

18:54:15 <AaronSw> No, I mean they're spamming me directly but pretending to have come from the W3C list

18:55:24 <danbri> ah... ok

18:56:11 <danbri> I started signing all my mails again, but Graham Klyne said it made them unreadable in his client...

18:56:55 <AaronSw> I've found spambayes to be really effective

18:58:49 <danbri> I get too much spam in my unknown-senders folder using spamassassin

19:07:04 * DanC wanders by...

19:07:43 <DanC> email forgery is pure evil

19:08:04 <DanC> nothing has evoked the vigilante in me as strongly as forged email

19:08:40 <danbri> I have been wondering about viability of the foaf/whitelist sharing approach, since forgery is on the rise

19:09:11 <mortenf> if it's coupled with signatures, it should still work

19:09:32 <danbri> I was thinking forgery is making it a doomed approach. But curently I think it paints spammers into a very unenviable corner...

19:09:49 <DanC> what corner is that?

19:10:23 <DanC> mortenf, do all your collaborators know how to sign their mail?

19:10:42 <mortenf> oh no...

19:11:03 <danbri> having to fake the from: fields of ordinary trusted email users...

19:11:07 <mortenf> but i don't think one approach will catch everything.

19:11:08 <danbri> makes their deeds more clearly evil

19:11:15 <danbri> and hopefully more illegal

19:12:08 <DanC> umm... the people who do most of the spamming don't seem to have any regard for laws

19:12:48 <danbri> true

19:13:01 <_joshua> i was wondering: what if there was an email backchannel

19:13:06 <DanC> I'd like to join a pool of domains that put up an "I promise not to let spam come out of my domain" bond

19:13:42 <DanC> you put in $10. if somebody can proove to a panel of your peers that you let spam come out of your domain, you lose it.

19:13:49 <_joshua> i.e. i get an email message from JoeBob@yahoo.com. why can't i go back and ask yahoo "did joebob@yahoo.com email joshua@burri.to?"

19:13:59 <_joshua> danc: what if people fake email coming from you?

19:14:39 <DanC> they have to prove that I was responsible somehow... either by running an open relay or some such.

19:14:40 <_joshua> "hey yahoo, is ip addr x.y.z.q authorized to send email for you?" etc

19:14:50 <_joshua> that doesn't really solve anything, though

19:15:02 <_joshua> people who don't actually spam aren't the problem

19:15:18 <DanC> I'd refuse mail from anybody who's not in the pool.

19:15:36 <DanC> the "is ip addr x.y.z.q authorized" is basically how the RMX thingy works. Have you read about it?

19:15:37 <_joshua> how do you know it's really from them?

19:15:48 <DanC> RMX

19:15:54 <_joshua> yes. i suggested something like that approach elsewhere a few years ago, and was shot down

19:16:03 <_joshua> "what if i am travelling and i want to send email?"

19:16:22 <DanC> yes, travelling is one of the obstacles to RMX.

19:16:29 <_joshua> smtp-auth?

19:16:37 <DanC> but SMTP/SSL is more widely deployed now... i.e. you can tunnel back to your ISP or whatever

19:16:39 <_joshua> i dunno. i think the problem is actually in the structure of email itself

19:16:48 <_joshua> but RMX is a good first step

19:17:01 <_joshua> then spammers wouldn't be able to fake Froms

19:17:19 <DanC> yeah... forgery undermines so many other mechanisms.

19:17:19 <_joshua> i'm wondering if a reverse query is workable. "did you really send email?"

19:18:41 <_joshua> i send you mail. you get mail that's from me@mine to you@yours. then in a separate conversation, yours asks mine if me@mine really sent mail to you@yours

19:19:24 <_joshua> a spammer could set up a host to answer these questions, but then we'd have ahostname and an ip address to ban, and they wouldn't be able to use proxies and whatever

19:19:54 <JibberJim> How would you identify the machine to ask that question to?

19:20:17 <_joshua> whatever's to the right of the @

19:20:27 <_joshua> perhaps this would be a part of SMTP, so MXing for that domain

19:20:37 * DanC futzes around with jabber... wonders if jabber service is federated ala IRC... and how the two interoperate

19:21:08 <_joshua> if it's a backup system it might say "i recognize that user but i don't know if it was sent. ask later" so you get a transient failure ala normal mail

19:21:20 <JibberJim> Hmm, surely we have more than 1 mail server in the MX records, and in any case why should the server I send with, be the server I recieve mail with?

19:21:33 <deltab> DanC: it's like e-mail

19:22:00 <DanC> "Jabber's architecture is modeled after that of e-mail" -- http://www.jabber.org/about/techover.html?PHPSESSID=d01d0133fb35ecf0f50c0cdbbc156bdb#distributednetwork

19:22:03 <_joshua> I guess the fundamental question is, "can we continue to have systems working exactly like they currently do AND solve the spam issue" and i think the answer is no

19:22:11 <DanC> PHPSESSID? sigh.

19:22:22 <_joshua> i think there will have to be fundamental changes to the current mechanisms

19:22:39 <DanC> yes, we have to change the economics of email at least a little bit

19:22:40 <_joshua> despite having a From: line, all email is actually of anonymous origin

19:22:50 <DanC> hence my idea of a $10 ante

19:22:53 <_joshua> I'm not sure it's an economics thing, actually. Some change has to be made

19:22:59 <danbri> . http://www.jabber.org/about/techover.html#distributednetwork work ok

19:23:26 <_joshua> If I can force spam to always come from accountable hosts or domains

19:23:42 <_joshua> then people could still send you mail, but you'd be able to set up blacklists that aren't dodgeable

19:24:09 <_joshua> right now they can just use proxies or whatever to get around IP and other blacklists

19:24:19 <_joshua> since there's no accountability at the originating side

19:24:20 <DanC> proxies?

19:24:36 <_joshua> yeah. lots of spammers use open relays and misconfigured proxies to send mail

19:24:46 <_joshua> lots of those home network sharing proxies, for example

19:25:13 * JibberJim was pleased to see an attack on my home ADSL box looking for open proxy, by the ADSL supplier.

19:25:25 <_joshua> My ISP does that as well

19:25:34 <_joshua> Good network hygiene

19:25:55 <_joshua> anyway i wonder if we're just thinking about the problem wrong

19:26:56 <_joshua> ooh. so if you have an smtp callback like i described earlier, spammers could always say "yes i sent that" so their spam could go through

19:27:20 <_joshua> so when you call back you say something like "did i-just-made-this-up@yours try to send an email to joshua@burri.to"

19:27:46 <_joshua> and the spammer server says "yep! i did" and you know it's lying, so you can just shithole that server immediately, and everything in the future for there

19:28:38 <DanC> hmm... so jabber.org is only slightly less centralized than AIM or yahoo messenger, yes?

19:28:55 <DanC> i.e. it's no less centralized; it's just a wee bit more open

19:29:00 <mdupont> DanC: anyone can setup a server

19:29:01 <deltab> jabber.org : jabber :: hotmail.com : mail

19:29:09 <mdupont> the server source code is open

19:29:15 <mdupont> the protocol is open

19:29:21 <mdupont> many clients are not open

19:29:23 <danbri> i have a jabber account on qmacro's server

19:29:28 <danbri> my jabber id includes his domain name

19:29:32 <mdupont> ok

19:29:41 <mdupont> i am mdupont@jabber.org iirc

19:29:48 <DanC> but I don't really want a danc@jabber.org account unless I want to trust their server

19:30:02 <_joshua> you can set up your own server

19:30:05 <_joshua> gale works in this way too

19:30:11 <DanC> gale?

19:30:24 <danbri> i am happy with that btw as other metadata can get from my identity otherwise described to my jabber id... and if my jabber id changes, it shouldn't be too expensive for folks to find my new id

19:30:26 <_joshua> another instant messaging system. not all XML-ariffic like jabber

19:30:37 <mdupont> DanC: setup your own server and only let people with ssh in

19:30:42 <DanC> I started setting up a jabber server in w3.org, but it went nutso and filled up /var with gigabytes of log messages.

19:31:00 <_joshua> i run a gale server so i'm galeable at joshua@burri.to which is my normal mail address

19:31:14 <mdupont> DanC: ooops!

19:31:28 <danbri> i looked into gale at one point, but have forgotten whatever i learned

19:31:37 <_joshua> it's crypto-heavy

19:32:01 <mdupont> i dont know too much about how to do this stuff, but chillywilly from dotgnu has his own server, as well as theo

19:32:03 <DanC> "Gale", a secure, distributed, open-source chat system

19:32:10 <DanC> -- http://www.mail-archive.com/cryptography@c2.net/msg02739.html

19:32:40 <_joshua> ah, that's ryan. he's the sealand guy

19:32:57 <_joshua> there's some stuff that's wrong in that

19:33:25 <_joshua> it's not based on email addresses, as he says. but they can be similar

19:34:14 <DanC> "the public [IRC] networks are not really suited for professional use"

19:34:41 <_joshua> Generally they aren't.

19:34:47 <_joshua> takeover kiddies and whatever

19:35:00 <_joshua> #perl on EFnet was an ongoing battle

19:35:17 <_joshua> I just made a simple gateway to/from Gale to AIM so I could talk to friends on gale via my Danger

19:35:28 <_joshua> and I wrote a nice little IRC gateway too

19:35:33 <DanC> hmm

19:35:56 * DanC wonders how to reconcile an "anybody can join and talk to everybody" structure like #rdfig with public key encryption

19:36:13 <_joshua> It also has public instances

19:36:18 <_joshua> er, "categories"

19:36:37 <_joshua> anyone can subscribe to anything. for example, you could subscribe to joshua@burri.to

19:36:48 <_joshua> however, you'd just get the stuff encrypted with my public key, so it'd be garbage

19:36:54 <dngor> It also has a #gale channel on this very network, but we're all asleep.

19:36:59 <_joshua> public categories have keys that can be fetched

19:37:01 <dngor> (and have been for days)

19:37:18 <_joshua> I don't get the point of that channel :)

19:37:36 <DanC> if #gale had a corresponding category in the gale world, it might make sense

19:37:39 <_joshua> My wife is out shopping with the person who wrote fugu, right this very second

19:37:47 <_joshua> they left me alone so i could play with the danger

19:37:59 <dngor> As long as irony's heart beats and lungs do their lungy things, we shall idle on #gale!

19:38:21 <_joshua> DanC: Gale is more long-winded; you can have multiline messages. so it doesn't map to irc perfectly

19:38:35 <dngor> Unless tansaku's a bot. Them I'm going to feel very embarassed.

19:38:57 <_joshua> no, i've seen him talk

19:42:45 <DanC> TODO: (a) set up my pgp key for use in evolution. (b) spam-protect dm93.org so I can start giving out my email address there (c) clean up my sidekick notes (d) try to talk to iChatAV from linux

19:42:45 <dc_rdfig> Label TODO not found.

19:46:13 <DanC> oh.. random question: can I run multiple DNS servers on a debian box that has mutiple IP addresses? i.e. are the UDP ports kept separate?

19:48:26 <_joshua> of course, so long as you don't bind 0.0.0.0 but instead the specific address

19:50:50 <DanC> and does djbdns bind to a specific address?

19:51:51 <AaronSw> you can set the address djbdns binds to in /service/djbdns/env/IP

19:52:14 <AaronSw> err s/djbdns/dnscache|tinydns/

19:52:26 <DanC> ok, thx

19:54:26 <AaronSw> gr

19:54:41 * DanC made a note "Shareholder democracy

19:54:41 <DanC> 13 Apr kc star

19:54:41 <DanC> The market Floyd norris, writer for ny times" ...

19:54:51 <DanC> ah... good... http://www.kansascity.com/mld/kansascity/business/personal_finance/5562720.htm

19:55:04 <DanC> that was easy to find with google...

19:55:18 <DanC> odd... I think I tried to find it from kansascity.com earlier and failed. go figure.

19:59:45 <DanC> logger_2, pointer?

19:59:45 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-07-06#T19-59-45

20:03:01 * DanC jots a few notes about gale, psi, jabber at http://dm93.org/z2001/RemotePresence

20:04:08 <DanC> AaronSw, do you use iPhoto? I've been noodling on an HTTP server in python that serves up iPhoto's data

20:07:05 * DanC is once again frustrated that he can't use sql update on his quicken records... really wants to switch to something like saCASH

20:12:15 <AaronSw> yeah, I use iPhoto occasionally

20:12:22 <AaronSw> (why do I keep getting disconnected?)

20:14:52 <mortenf> " AaronSw has quit (Read error: 104 (Connection reset by peer)) " - don't know :(

20:32:24 <mortenf> btw aaron, the rdf at http://creativecommons.org/licenses/index.rdf made redland barf, it doesn't declare the default ns - i thought you might want to fix that some day.

20:33:20 <AaronSw> Oops! It should be fixed soon.

23:54:40 <DanC-AIM> Great day at the pool