2007-05-16T00:06:24 <flowhase> re
2007-05-16T00:06:32 <flowhase> sorry for being gone :/
2007-05-16T00:06:49 <flowhase> i had kind of a nervous breakdown
2007-05-16T00:07:11 <flowhase> erm s/for being/i were
2007-05-16T00:07:30 <flowhase> i had to get sorted the last few days and i am back on track
2007-05-16T00:14:14 <xorAxAx> have you written down the discussion results? :)
2007-05-16T00:15:55 <flowhase> xorAxAx: some minutes ago :>
2007-05-16T00:16:00 <xorAxAx> cool :)
2007-05-16T00:16:23 <flowhase> i thought i better do this first and rejoin irc afterwards ;>
2007-05-16T00:17:00 <xorAxAx> hehe
2007-05-16T00:19:57 <ThomasWaldmann> flowhase: "one of the standard themes" that should be "modern" as the other two are less well maintained or have known problems
2007-05-16T00:20:19 <flowhase> ThomasWaldmann: well, i thought so :>
2007-05-16T00:20:30 <flowhase> it's also the most beautiful imo
2007-05-16T00:20:34 <ThomasWaldmann> (and welcome back :)
2007-05-16T00:24:12 * flowhase is currently trying to get more organized
2007-05-16T00:24:38 <ThomasWaldmann> flowhase: maybe add some ideas about how you want to test the theme output
2007-05-16T00:24:51 <flowhase> i now have a in/out-panel in my room, many todo-postits on it, rated after importance
2007-05-16T00:25:10 <flowhase> hmm
2007-05-16T00:25:48 <flowhase> actually that is a good question ... i don't really know how to test that (apart from browser/curl)
2007-05-16T00:26:06 <ThomasWaldmann> or at least WHAT you need to test
2007-05-16T00:26:50 <flowhase> hmm
2007-05-16T00:27:31 <flowhase> on a high level i think it is the html 4.01 compliance
2007-05-16T00:27:46 <ThomasWaldmann> "Make more use of the "C" in CSS (cascading)" < very good idea
2007-05-16T00:27:53 <flowhase> on a very low level it's that everything important is rendered for sure
2007-05-16T00:27:59 <flowhase> hehe
2007-05-16T00:28:24 <flowhase> ffesti suggested that it would be terrific to have something like that :>
2007-05-16T00:29:09 <flowhase> and i am kind of used to reordering css, i like exploring css possibilities
2007-05-16T00:29:11 <ThomasWaldmann> i mean rather the non-immediately obvious things like browser compatibility, rtl support, different screen sizes, ...
2007-05-16T00:30:02 <ThomasWaldmann> (modern copes rather good with that, but you maybe won't produce identical output, just more or less equivalent output)
2007-05-16T00:30:20 <flowhase> yeah, i am pretty sure it will not be identical
2007-05-16T00:30:56 <flowhase> hmm
2007-05-16T00:30:57 <ThomasWaldmann> (and that's good, I guess :) just try to make sure no regressions happen.
2007-05-16T00:31:20 <ThomasWaldmann> another rather annoying thing to check is MSIE
2007-05-16T00:31:35 <flowhase> i will try the ie4linux package
2007-05-16T00:31:58 <flowhase> i also have a winxp install on my desktop box
2007-05-16T00:31:59 <ThomasWaldmann> i doesnt need to be perfect, but at least usable somehow
2007-05-16T00:32:12 <flowhase> (though i am mostly mobile around town and not in my room)
2007-05-16T00:32:28 <ThomasWaldmann> ie4l or vmware are maybe good for testing
2007-05-16T00:34:32 <flowhase> hmm, does linux vmware run on amd64?
2007-05-16T00:35:11 <ThomasWaldmann> 64bit mode?
2007-05-16T00:35:22 <flowhase> yes
2007-05-16T00:36:00 <flowhase> it needs kernel modules and those were binary iirc
2007-05-16T00:36:25 <ThomasWaldmann> hmm, i think i didnt try that yet, but looking at the target market, vmware server maybe SHOULD run
2007-05-16T00:37:08 <flowhase> ah .. totally forgot that this one is available for free too now
2007-05-16T00:37:35 <flowhase> well, i will find a way to manage the IE testing
2007-05-16T00:40:12 <ThomasWaldmann> http://blog.derjohn.de/snipsnap/space/start/2007-03-25/1
2007-05-16T00:44:50 <flowhase> argl, snipsnap
2007-05-16T00:44:50 <dreimark> flowhase: it does
2007-05-16T00:45:27 <ThomasWaldmann> flowhase: hit, loss, neglectable. what do you mean by "hit"? win?
2007-05-16T00:45:55 <flowhase> yes
2007-05-16T00:45:59 <flowhase> damnit
2007-05-16T00:46:06 <flowhase> now that you mention it :>
2007-05-16T02:19:33 <ThomasWaldmann> http://neil.fraser.name/software/mobwrite/
2007-05-16T11:25:48 <dreimark> ThomasWaldmann: SystemInfo is broken by your last change in 1.6 t.www.de if you are not logged in
2007-05-16T11:27:22 <dreimark> and if you are logged in
2007-05-16T14:56:32 <dreimark> oin
2007-05-16T15:01:57 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2066:6d1d337c5a76 1.6/MoinMoin/macro/SystemInfo.py: SystemInfo: fixed isSpiderAgent
2007-05-16T18:33:21 <dreimark> bbl
2007-05-16T21:09:12 <grzywacz> moin
2007-05-16T22:01:26 <neagulm> hello
2007-05-16T22:05:14 <dreimark> moin
2007-05-16T22:08:17 <dreimark> neagulm: I forwarded a bunch of mails from skip to you
2007-05-16T22:08:40 <neagulm> dreimark, am I'm reading them...
2007-05-16T22:21:22 <neagulm> dreimark, What is your opinion about Skip's proposal ?
2007-05-16T22:23:10 <neagulm> i have to leave #moin-dev for 30 seconds for switching computers. brb
2007-05-16T22:26:16 <neagulm> dreimark, i'm back
2007-05-16T22:27:21 <dreimark> neagulm: I've invited skip for a few minutes
2007-05-16T22:27:50 <neagulm> dreimark, i just received the email :)
2007-05-16T22:28:17 <dreimark> so he has no your email address in his adressbook too :)
2007-05-16T22:28:19 <dreimark> now
2007-05-16T22:30:04 <dreimark> neagulm: if they go on with changes into their codebase to make it easier to use it with others application its a important step
2007-05-16T22:31:16 <starshine> neagulm: if nobody mentioned it yet, red/green probably isn't a good color pair for bad/good identification
2007-05-16T22:31:37 <starshine> since r/g colorblindness is the most common kind of colorblindness
2007-05-16T22:32:02 <dreimark> starshine: its more important for men as womans :)
2007-05-16T22:33:26 <neagulm> dreimark, yes. But i don't know if the approach sugested by Skip is the right one for a wiki. I say that because we could use a normal wiki page for storing the suspect pages/revisions. Or some other "wikian" method
2007-05-16T22:35:35 <neagulm> dreimark, but in the end is does not really matter how we store them. By using SpamBayes to take care about this problem I have less to work :) So it would be ok for me.
2007-05-16T22:36:19 <dreimark> neagulm: I do prefer to use the wiki for that too. But we have to take care on the project spambayes too. interconnection is pretty fine
2007-05-16T22:37:26 <dreimark> may be we can have it configurable. He asked me to forward the mails so I think he's intereseted in your opinion too ;)
2007-05-16T22:38:25 <dreimark> There is currently no timetable when that changes are published
2007-05-16T22:46:28 <neagulm> dreimark, i'm thinking about this approach.
2007-05-16T22:48:29 <dreimark> starshine: any ideas about icons instead of only colors?
2007-05-16T22:49:47 <dreimark> e.g. increasing darth vader
2007-05-16T22:49:55 <dreimark> for spam
2007-05-16T22:50:36 <starshine> how many icons total?
2007-05-16T22:51:16 <starshine> you could have a paper, a crumpled paper, a burning crumbled paper, a radioactive trashbin with paper in it..
2007-05-16T22:51:57 <neagulm> :)
2007-05-16T22:52:44 <starshine> and a black skull with glowing eyes.
2007-05-16T22:53:08 <starshine> make its edge "shiny" so someone with dark theme can still see it.
2007-05-16T22:56:22 <dreimark> neagulm: how much?
2007-05-16T22:56:56 <skipm> neagulm, dreimark: hi
2007-05-16T22:57:01 <neagulm> dreimark, "how much" ?
2007-05-16T22:57:05 <neagulm> skipm, hello
2007-05-16T22:57:31 <skipm> neagulm, dreimark: i have about 30 minutes then have to catch a train
2007-05-16T22:57:57 <dreimark> starshine: minimum is three icons because one has to be undecided
2007-05-16T22:58:04 <dreimark> hi skipm
2007-05-16T22:58:46 <dreimark> skipm: one of the most important questions is when will this new interface be available ?
2007-05-16T22:58:54 <starshine> should there be any for scores-good?
2007-05-16T22:59:05 <dreimark> starshine: yes
2007-05-16T22:59:06 <starshine> paper with a pretty thing on it
2007-05-16T22:59:24 <starshine> scroll with a lil gold seal
2007-05-16T22:59:50 <dreimark> starshine: cool
2007-05-16T22:59:52 <starshine> for a total of 2 positive, one just-a-paper, and 4 grades of bad
2007-05-16T23:00:21 <neagulm> skipm, dreimark what is your opinion about checking differences beetwen revisions ?
2007-05-16T23:00:40 <dreimark> neagulm: that is very important.
2007-05-16T23:01:04 <skipm> dreimark: ah the new core server with plugins? i took the pop3 proxy and ripped out the pop3 stuff last night. Now I need to write a plugin for web apps
2007-05-16T23:01:39 <skipm> dreimark, neagulm: given that you'd maybe be client #1 or #2 how would you find it easiest to communicate with the plugin?
2007-05-16T23:03:13 <neagulm> skipm, in what context ?
2007-05-16T23:04:09 <dreimark> starshine: sounds good
2007-05-16T23:04:45 <dreimark> skipm: you are asking about the interface ?
2007-05-16T23:05:29 <skipm> dreimark: maybe i misunderstood your use of the word "interface".
2007-05-16T23:07:53 <skipm> dreimark: i thought you were asking about my idea for a standalone server with a (for example) XML-RPC or POP3 plugin to talk with
2007-05-16T23:09:10 <dreimark> skipm: ahh my fault, sorry
2007-05-16T23:12:19 <dreimark> skipm:xml-rpc we do use for several other implemetantions already
2007-05-16T23:12:48 <dreimark> some parts from wikisync could be used for that too
2007-05-16T23:12:52 <skipm> yes, i figured it would be adequate for the task and make such a standalone classifier language-neutral
2007-05-16T23:13:57 <dreimark> skipm: I do believe that is an important change for your project and if the interface is standardized its easier for us
2007-05-16T23:14:03 <dreimark> to use updates
2007-05-16T23:14:05 <dreimark> from sb
2007-05-16T23:14:57 <neagulm> skipm, dreimark i agree!
2007-05-16T23:15:08 <dreimark> sorry interface is not the right word
2007-05-16T23:15:11 <skipm> dreimark, neagulm: my motivation for trying this approach is that i ran into exactly the same problem when trying to integrate sb w/ roundup - i didn't know my ass from a hole in the ground when it came to the roundup code
2007-05-16T23:16:48 <dreimark> hehe
2007-05-16T23:18:33 <skipm> dreimark, neagulm: then there are about a bazillion other web sites and/or web apps which could use decent spam detection. it's kind of tough to integrate with all of them.
2007-05-16T23:18:54 <dreimark> skipm: can you give an idea how long it will take ?
2007-05-16T23:19:31 <dreimark> skipm: yeah that why it is an imporatant wanted and necessary change
2007-05-16T23:20:30 <xorAxAx> i dont see how xml-rpc would be useful here in the first place :)
2007-05-16T23:20:30 <dreimark> I would prefer it available during soc
2007-05-16T23:20:42 <xorAxAx> because of data storage concerns
2007-05-16T23:20:46 <skipm> dreimark: good question. there's a crunch to get this going for the new bugs.python.org (roundup tracker). I'm working on it as I can. At least for an hour or so on monday mornings at Tech Coffee (nerds hack open source and drink cooffee) but i'm also trying to squeeze in a few hours the rest of the week
2007-05-16T23:21:37 <skipm> dreimark: i would guess it will be a couple weeks before i have something i feel comfortable with.
2007-05-16T23:22:19 <dreimark> xorAxAx: any other ideas ?
2007-05-16T23:22:33 <xorAxAx> dreimark: how about a clean pythonic direct interface?
2007-05-16T23:22:34 <skipm> dreimark, neagulm: if you could provide some requirements for your needs, it would help direct my efforts. maybe we can carry on this discussion via email? i kind of have to go catch my train
2007-05-16T23:22:42 <neagulm> xorAxAx, skipm, dreimark it is not required to store/send the actual data. hash functions can be used to transform the original message. Tokens are still tokens
2007-05-16T23:23:05 <xorAxAx> neagulm: can you explain the idea of having some remote interface in this context?
2007-05-16T23:23:25 <skipm> neagulm: hash functions? you lost me.
2007-05-16T23:23:35 <dreimark> skipm, neagulm can you please create a page on MM about that idea too
2007-05-16T23:24:23 <skipm> dreimark, neagulm: is it okay to paste in bad ascii art here?
2007-05-16T23:24:49 <skipm> dreimark: sure. subpage of SkipMontanaro okay?
2007-05-16T23:24:59 <dreimark> yep
2007-05-16T23:25:31 <dreimark> bad ascii art - lets see, currently no opinion
2007-05-16T23:25:43 <skipm> dreimark, neagulm: i really have to book out of here. neagulm - send me (skip@pobox.com) an email so i have yours (bad ascii art can wait)...
2007-05-16T23:25:57 <neagulm> skipm, I was refering about xorAxAx concerns about data storage. The classifier should not be trained with the actual feature vector vect=[v1, v2, v3, ..., vn] but with [hsh(v) for v in vect] :)
2007-05-16T23:26:20 <xorAxAx> neagulm: i was rather thinking about use cases :)
2007-05-16T23:26:38 <xorAxAx> now add fuzzy hashes and you have written some backend library for services like razor
2007-05-16T23:27:53 * grzywacz is waiting for coffee to kick in before doing some creative stuff ;-S
2007-05-16T23:28:31 <xorAxAx> good evening my student
2007-05-16T23:28:39 <grzywacz> xorAxAx, did you happen to find anything essential missing from my proposal of xmlrpc calls?
2007-05-16T23:28:44 <grzywacz> xorAxAx, evening my hmre.
2007-05-16T23:28:55 <xorAxAx> let me reread your proposal
2007-05-16T23:29:02 <grzywacz> Ok.
2007-05-16T23:29:06 <grzywacz> ;-)
2007-05-16T23:29:35 <neagulm> xorAxAx, I think that we are "out of sync"...
2007-05-16T23:29:51 <neagulm> xorAxAx, I think that i worked to much today
2007-05-16T23:29:56 <xorAxAx> neagulm: well, explain to me which use cases there will be for the user :)
2007-05-16T23:30:17 <xorAxAx> i understood that you want to do remote classification by transmitting hashes
2007-05-16T23:30:33 <xorAxAx> grzywacz: did you consider the ideas we exchanged in the query?
2007-05-16T23:30:57 <grzywacz> xorAxAx, let me check the log
2007-05-16T23:31:16 <xorAxAx> grzywacz: anonymous authors have a "username" as well - their ip/hostname
2007-05-16T23:31:25 <grzywacz> Well, the opposite direction is obviously still missing.
2007-05-16T23:31:48 <grzywacz> xorAxAx, hm, good. These may be handy as well.
2007-05-16T23:32:26 <neagulm> xorAxAx, for the user nothing special, he will work normaly as with a local classifier
2007-05-16T23:32:36 <xorAxAx> yes, the other direction, grzywacz
2007-05-16T23:32:40 <grzywacz> The last time I launched eclipse to poke around the code to identify places for hooks my system has literally blown up because of OOM. ;-)
2007-05-16T23:32:54 <xorAxAx> grzywacz: see the storage soc2007 page for metadata you may want to transmit
2007-05-16T23:33:18 <grzywacz> Ok, checking.
2007-05-16T23:33:27 <xorAxAx> neagulm: well, i dont mean how he would use it, but why he would use it :)
2007-05-16T23:34:22 <ThomasWaldmann> moin
2007-05-16T23:34:27 <grzywacz> moin ThomasWaldmann
2007-05-16T23:35:03 <neagulm> xorAxAx, Eg. the classifier could be shared between more wikis
2007-05-16T23:35:08 <neagulm> moin ThomasWaldmann
2007-05-16T23:35:20 <xorAxAx> neagulm: well, you mean the actual corpii/states, right?
2007-05-16T23:35:30 <xorAxAx> the classifier itself doesnt need to be "shared" in the first place
2007-05-16T23:38:15 <neagulm> xorAxAx, yes. not shared as sharing his state but as all wiki's using the same instance of the classifier
2007-05-16T23:38:32 <neagulm> xorAxAx, dreimark, I think that I'm to tired
2007-05-16T23:38:41 <xorAxAx> neagulm: doesnt the latter imply the former?
2007-05-16T23:38:52 <xorAxAx> if it does not, i dont see the idea at all
2007-05-16T23:39:01 <xorAxAx> and want to read some wiki page about it
2007-05-16T23:39:39 <dreimark> neagulm: you have +1 to us
2007-05-16T23:40:26 <dreimark> and I am a bit tired too
2007-05-16T23:43:19 <neagulm> xorAxAx, this english language... What I want to say: A classifier is running on machine A. Wiki W1 is using XML-RPC (or anything else) to ask A about a page. Wiki W2 is using the same method to query A. No computation is done localy. The classification ocours on A.
2007-05-16T23:43:50 <xorAxAx> neagulm: why would it ask it about the page?
2007-05-16T23:44:17 <neagulm> xorAxAx, send it the page in a secure manner
2007-05-16T23:44:43 <xorAxAx> hmm?
2007-05-16T23:45:54 <dreimark> xorAxAx: currently we do something quite similiar with BC
2007-05-16T23:46:02 <xorAxAx> BC?
2007-05-16T23:46:08 <dreimark> Badcontent
2007-05-16T23:46:11 <dreimark> cC
2007-05-16T23:46:25 <xorAxAx> ugh
2007-05-16T23:46:43 <neagulm> xorAxAx, dreimark I'm switching computer again
2007-05-16T23:46:44 <neagulm> brb
2007-05-16T23:46:45 <grzywacz> ...
2007-05-16T23:46:45 <neagulm> :)
2007-05-16T23:46:52 <dreimark> that was fast
2007-05-16T23:46:56 <dreimark> :)
2007-05-16T23:47:04 <xorAxAx> well, the main point here is the idea of data secrecy
2007-05-16T23:47:25 <xorAxAx> i am not sure if i want to submit hashes of all my n-grams/words (128 bit per word?) to a remote wiki
2007-05-16T23:47:38 <grzywacz> I must buy more ram before coding starts, otherwise it's going to be a total disaster. -_-;
2007-05-16T23:47:52 <xorAxAx> and even then, use cases should be drafted first, on a wiki page
2007-05-16T23:47:54 <dreimark> grzywacz: how much do you have?
2007-05-16T23:48:00 <grzywacz> 512...
2007-05-16T23:48:10 <dreimark> I've 384 here
2007-05-16T23:48:18 <grzywacz> firefox+xchat+eclipse+psi+firefox = total death
2007-05-16T23:48:37 <grzywacz> Maybe there's a hardware/kernel as well, but it would be swapping like crazy anyway.
2007-05-16T23:48:48 <grzywacz> *hardware/kernel problem
2007-05-16T23:49:02 <grzywacz> Oh, forgot to mention amarok and kde...
2007-05-16T23:49:50 <dreimark> currently, freeciv, konversation, konqueror(6), firefox, kde, eclipse
2007-05-16T23:50:30 <xorAxAx> 31316 alexande 18 0 333m 93m 3840 S 12.9 19.0 71:37.04 irssi clear/ reset
2007-05-16T23:50:34 <xorAxAx> won
2007-05-16T23:51:07 <grzywacz> wtf
2007-05-16T23:51:08 <grzywacz> ;]
2007-05-16T23:51:34 <grzywacz> dreimark, mind that I have a 64-bit system and in some cases it can almost double the mem usage...
2007-05-16T23:52:08 <grzywacz> Maybe I should switch back to my beloved windowmaker...
2007-05-16T23:53:39 <ThomasWaldmann> ram (esp ddr2) is dirt cheap currently
2007-05-16T23:54:33 <grzywacz> Unfortunately I have ddr 1 and a strange motherboard with single/double-sided memory issues. ;-)
2007-05-16T23:55:49 <dreimark> grzywacz: the office system is extremly fast against that here, but most of my code was written on a slower and less mem machine.
MoinMoin: MoinMoinChat/Logs/moin-dev/2007-05-16 (last edited 2007-10-29 19:10:18 by localhost)