Perl 6 - the future is here, just unevenly distributed

IRC log for #rosettacode, 2011-02-08

| Channels | #rosettacode index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:21 FireFly left #rosettacode
00:38 slavik1 left #rosettacode
00:54 Mathnerd314 joined #rosettacode
01:24 kpreid left #rosettacode
01:27 kpreid joined #rosettacode
02:50 SoniaKeys joined #rosettacode
03:34 mwn3d_phone shortcircuit: another vandalism attack cleaned up
06:06 SoniaKeys left #rosettacode
11:09 FireFly joined #rosettacode
11:43 FireFly left #rosettacode
11:46 BenBE left #rosettacode
11:50 BenBE joined #rosettacode
12:12 kpreid left #rosettacode
12:25 parsleyfirefly joined #rosettacode
12:42 kpreid joined #rosettacode
13:50 FireFly joined #rosettacode
14:07 parsleyfirefly1 joined #rosettacode
14:11 kpreid left #rosettacode
14:33 slavik1 joined #rosettacode
14:34 kpreid joined #rosettacode
14:35 slavik1 left #rosettacode
14:42 shortcircuit Doing a presentation on Rosetta Code in front of a local Java Users' Group this month.
14:42 shortcircuit Just sent them the abstract and bio. Wasn't sure what to do for a bio, so I sent them my Geek Code.
14:43 parsleyfirefly left #rosettacode
14:44 parsleyfirefly1 is now known as parsleyfirefly
14:46 mwn3d_phone shortcircuit: what exactly will you talk about? The whole site? The java stuff on the site?
14:49 shortcircuit mwn3d_phone: Usually, I lead off by distributing a set of language-nonspecific URLs and a set of language-specific URLs. Then I give a brief overview of the site organization, and answer the questions of people who are browsing the links on their laptops.
14:53 mwn3d_phone Cool
14:59 slavik1 joined #rosettacode
15:42 kpreid left #rosettacode
15:56 wolfram_ joined #rosettacode
15:56 shortcircuit wolfram_: The folks in here should be able to offer feedback on your FFT task
15:58 wolfram_ shortcircuit: Ok, will try
16:01 opticron ooh...fft
16:01 opticron fun
16:02 wolfram_ Created a new task at http://rosettacode.org/wiki/FFT. It's my first Rosettacode edit, so someone might have a look if I did alright. If things look good, maybe the draft status can/should be removed!?
16:02 fedaykin "FFT. - Rosetta Code"
16:06 opticron looks ok.  I would probably specify the wave as an input array for work to be performed on
16:08 opticron also, it appears that perl6 has new wizardry that I didn't know about
16:08 opticron not that I've been following perl6 at all
16:09 wolfram_ You mean, specify the input wave in the formulation of the task (as opposed to: in the perl6 solution?)
16:09 MigoMipo joined #rosettacode
16:11 opticron yes, so people have consistent input and consistent results to check against
16:11 shortcircuit Would it help for me to host a datafile that can be parsed and manipulated?
16:13 shortcircuit I could provide a single-channel 8-bit raw PCM audio file.
16:15 shortcircuit linear-encoded. Or whatever encoding anyone cares to provide the sox output string to transcode to.
16:17 opticron heh
16:17 opticron that could be fun
16:17 wolfram_ @shortcircuit: You mean for FFT analysis? Sounds (!) like overkill to me. If you do so, I might be tempted to append my perl/Tk (perl5) code for graphical spectrum analysis :-)
16:17 lambdabot Unknown command, try @list
16:19 wolfram_ I mean, in the 16 samples I use, I could even create a multitone. That should be sufficient to check the validity of the code.
16:21 shortcircuit wolfram_: I know P6 handles JSON well. Can you provide a JSON export of your input wave as part of the task description?
16:21 shortcircuit The utility comes from being able to compare outputs for correctness.
16:21 shortcircuit (If it's only 16 samples, then, yeah, a PCM file is way overkill)
16:23 wolfram_ shortcircuit: but I know neither P6 nor JSON very well. I know the FFT but for the P6 required for my code I had to learn heavily from #perl6
16:24 wolfram_ But I was certain, that a very compct P6 FFT should be possible...
16:25 shortcircuit wolfram_: I think you can just use .json on your array object to get the .json form of it. So something like "print obj.json;" should work.
16:27 wolfram_ No luck: Method 'json' not found for invocant of class 'Range' / Method 'json' not found for invocant of class 'Array'
16:28 shortcircuit rakudo: my @seq = ^16; print @seq.json();
16:28 p6eval rakudo cad076: OUTPUT«Method 'json' not found for invocant of class 'Array'␤  in main program body at line 22:/tmp/u97JAywkLv␤»
16:28 shortcircuit Hm. Weird. I could have sworn I saw code like that being bandied about in #perl6.
16:29 shortcircuit Might be good to poke the folks in #perl6 about that.
16:31 wolfram_ Got to leave. Anyway the input and output are both attached to my perl6 code. If someone has a good format to put that into the task, go ahead. I might have another look tomorrow.
16:32 wolfram_ left #rosettacode
16:42 kpreid joined #rosettacode
17:22 BenBE left #rosettacode
17:22 BenBE joined #rosettacode
17:46 kpreid left #rosettacode
17:57 mwn3d_phone I'm seeing double carrot quotes in the perl 6 example for [[FFT]]
17:57 fedaykin http://rosettacode.org/wiki/FFT
17:57 kpreid joined #rosettacode
17:57 mwn3d_phone Are those supposed to be there?
17:57 shortcircuit @odd  »*=«
17:57 lambdabot Maybe you meant: bid do id todo
17:57 * shortcircuit doesn't know.
17:59 opticron woah
18:00 TimToady yes, Perl 6 hyperoperators use some latin-1 chars, though there are ascii workarounds
18:00 TimToady that's just a vectorized *=
18:01 TimToady the technical term for those characters is "naughty French quotes"  :)
18:01 TimToady the ascii versions are called "Texas quotes" because they're bigger
18:02 TimToady so you can also say >>*=<< if you don't mind ugly
18:11 _ilbot joined #rosettacode
18:11 Topic for #rosettacode is now Rosetta Code: Coding in (inf - 1) languages ; http://irclog.perlgeek.de/rosettacode/today ; Pop a tab and stick around.
18:43 shortcircuit FYI: I'm setting $wgGroupSettings['*']['upload'] = false, and $wgGroupSettings['autoconfirmed']['upload'] = true.
18:44 shortcircuit Just having a user account will be insufficient, initially, to upload. It'll have to reach the 'autoconfirmed' point first. (Which is the same threshhold that drops CAPTCHAs.)
18:51 mwn3d_phone1 joined #rosettacode
18:53 BenBE shortcircuit What's the point for autoconfirm?
18:55 shortcircuit BenBE: After N seconds, an account is autoconfirmed. In the case of broad, automated spam attacks, it increases the window credentials need to be remembered over on the part of the spammer, increasing their operational costs.
18:55 shortcircuit It's not a particularly sophisticated counter, but it's a speedbump.
18:55 mwn3d_phone left #rosettacode
18:56 shortcircuit Currently, it's at 24 hours, but I might extend or modify that somehow.
18:56 mwn3d_phone1 is now known as mwn3d_phone
18:57 BenBE shortcircuit Well, even if I were an attacker I'd just launch my attack in two waves then ;-) Storage isn't that much an issue.
18:58 BenBE The first wave to generate loads of accounts and the second one to use those accounts for spamming.
18:58 shortcircuit BenBE: It's not an effective counter to a targeted attack. Anything short of individual management of privileges can be easily solved with a few lines of code.
18:58 shortcircuit BenBE: Now, though, you'd have to write those lines of code.
18:58 BenBE And even if you don't think of it: I can remember hundred thousands of accounts with only one seed and one small function simular to a rainbow table.
18:59 shortcircuit BenBE: Sure. And when this mechanism becomes ineffective, there will be a different mechanism to append or replace.
18:59 BenBE shortcircuit I'm sure there will be attack tools available already that implement this "delayed attack" mechanism.
19:00 shortcircuit BenBE: I'm sure there are. However, simply putting the skipcaptcha on autoconfirm massively reduced our spam while reducing the burden to human users.
19:01 BenBE I don't argue about that ^^
19:01 shortcircuit It's a balance of site convenience vs forcing attackers to be more sophisticated. As I said, anything short of individual management of privileges can be defeated with a few lines of code.
19:02 shortcircuit Right now, anon uploads of images are a problem, and text spam occasionally gets past initial captchas. So there will be a delay before image uploads are possible for new user accounts.
19:03 shortcircuit Image spam is much riskier for me than text spam, because image spam isn't someone throwing data at my community, it's using me as an image host.
19:03 shortcircuit That's going to cost more bandwidth than text spam will. (Not that I'm running low on bandwidth, I just don't want to be the image host for an email spam campaign.)
19:04 BenBE I do just the same on my site: I implement Anti-Spam stuff in a way which was hopefully not used before.
19:04 BenBE And that way making it hard to use pre-existing tools for spamming.
19:05 BenBE Adding a simple math captcha for my blog got me from 20-50 spam comments a day down to about 1 comment a week ...
19:05 shortcircuit I've been pondering multi-CAPTCHA verifications for certain scenarios.
19:05 shortcircuit Anon-edits with links, account creation, password failures...
19:06 shortcircuit I'm pretty sure they're using low-wage workers at places with a low cost of living for cracking the captchas at this point, so the problem becomes increasing their particular time cost.
19:07 shortcircuit Linking it with a DNSBL to control the triggering of multi-captcha vs single-captcha would reduce the overhead for legit users.
19:07 BenBE The captcha in my blog has a non-standard announcement text before it, is not clear plain text of the question AND uses non-standard field names that should tell an human attacker that I'm sure about that I don't tolerate spam.
19:08 shortcircuit I'm comfortable delegating to reCAPTCHA; they can mix up and adapt their mechanisms to increase spammer operational cost without my frequent attention and involvement.
19:08 shortcircuit The less time I have to spend on server maintenance, the better.
19:09 BenBE Well, Do you knoww how to break image captchass effectively? Write a small "Sex Image Viewer", distribute it to hundrets of people willing to watch them and present them with the Captchas you scrape - telling the user he got the captcha wrong ;-)
19:12 shortcircuit BenBE: That's an old one, yeah.
19:12 shortcircuit It's jsut that spam isn't often a problem on the site, and when it gets to be a problem, it usually only takes a small tweak to make it manageable again.
19:13 shortcircuit I outsource captcha management precisely so I don't have to deal with it frequently.
19:15 BenBE shortcircuit But still works reasonably well ;-) Social Engineeering FTW ^^
19:16 BenBE I just presented my 2ct on the spam topic ...
21:11 mwn3d_phone Guy who works at IBM in another channel: "'does this normally coredump?' Yeah....that means it worked"
21:12 BenBE That's absolutely correct ;-)
21:12 BenBE He was probably asking about the debugging functions of the kernel ;-)
21:16 slavik1 I hate Java
21:17 mwn3d_phone slavik1: ok I'll bite. What did it do to you?
21:17 slavik1 cause stupid developers from appearing
21:17 slavik1 gah!
21:17 slavik1 cause stupid developers to appear
21:30 kpreid left #rosettacode
22:21 MigoMipo left #rosettacode
22:21 shortcircuit slavik1: Don't get me started on Macs. ;)
22:21 slavik1 shortcircuit: that, too
22:21 slavik1 hippies suck!
22:21 slavik1 and I want my meego phone!!!
22:22 mwn3d_phone shortcircuit: when is your talk to the java devs?
22:22 mwn3d_phone And where are they?
22:24 slavik1 shortcircuit: tell them they suck
22:28 shortcircuit http://gr-jug.org/
22:28 fedaykin "GR-JUG"
22:29 shortcircuit Third Thu.
22:29 shortcircuit I was going to do it last month, but a late Thursday at work shifted around my after-work committment scheduling.
22:30 shortcircuit mwn3d_phone: Why, you thinking road trip? :)
22:30 kpreid joined #rosettacode
22:31 mwn3d_phone Heh I don't think so. Just anticipating an influx of java examples.
22:33 mwn3d_phone Also I don't even like the 7 hour drive to my parents house. Id be thinking plane trip if anything.
22:34 shortcircuit mwn3d_phone: The train is a very enjoyable means of transportation.
22:34 shortcircuit Also, from NY, you're not looking at losing cell signal along the way.
22:35 shortcircuit You'd hit GRR at about 11-ish in the evening. I could pick you up and lend you a couch.
22:35 mwn3d_phone I take it to work (commuter rail not amtrak). Longer in time than driving but much more enjoyable
22:43 parsleyfirefly left #rosettacode
23:18 Mathnerd314 left #rosettacode
23:26 FireFly left #rosettacode
23:29 Mathnerd314 joined #rosettacode
23:42 TimToady phenny: ask wolfram_ /me is kinda wondering why you used evens and odds instead of the complex numbers that Perl 6 has built in...
23:42 TimToady phenny: tell wolfram_ /me is kinda wondering why you used evens and odds instead of the complex numbers that Perl 6 has built in...
23:42 TimToady phenny?
23:43 TimToady duh, wrong channel, sorry

| Channels | #rosettacode index | Today | | Search | Google Search | Plain-Text | summary