Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2013-12-06

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

All times shown according to UTC.

Time Nick Message
12:57 run4flat joined #pdl
12:57 run4flat Ugh, the Benders seem to have forgotten me again. :-(
12:57 run4flat anybody care to share some oper love?
13:27 Mithaldu Bender1, trust run4flat
13:27 Bender1 But I don't trust you in #pdl, Mithaldu
13:27 Mithaldu Bender2, trust run4flat
13:27 Bender2 But I don't trust you in #pdl, Mithaldu
13:27 Mithaldu :(
13:27 run4flat argh!
13:27 Mithaldu oh wait, i don't have op either
13:27 run4flat yup
13:27 run4flat :-)
13:27 run4flat hmm... lungching should be up
13:27 run4flat lungching, ^^^
13:27 run4flat Mithaldu, thanks for trying!
13:32 Mithaldu haha, sure, even if it was a bit derpy
13:34 vicash Bender1, trust Bender2
13:34 Bender1 But I don't trust you in #pdl, vicash
13:34 vicash ha.. just trying
13:37 lungching run4flat: you are now op :)
13:37 Mithaldu lungching: can you try telling bender to trust people? :)
13:37 lungching Bender1: trust run4flat
13:37 Bender1 lungching, I already trust run4flat
13:38 lungching Bender1: trust Mithaldu
13:38 Bender1 OK, lungching
13:38 run4flat Mithaldu, I think that the Benders take note when someone they trust /oper's someone
13:38 run4flat let's make sure this worked
13:38 run4flat left #pdl
13:38 Mithaldu \o/ thanks :)
13:38 run4flat joined #pdl
13:38 run4flat ok, seems to be working, essentially
13:38 run4flat anybody else want ops?
13:38 run4flat vicash?
13:38 run4flat sivoais?
13:39 Mithaldu sivoais does so much i think you need to push it on him
13:39 run4flat Bender1, trust sivoais
13:39 Bender1 OK, run4flat
13:40 run4flat agreed, Mithaldu
13:40 run4flat preaction?
13:48 vicash run4flat: no i dont want ops
13:48 vicash but thanks !
13:49 run4flat n/p :-)
13:56 vicash run4flat: i have a question. let's say i plot 3 hours worth of data where each data point is sampled at 100ms giving about 100K points .. i use PDL::Graphics::Prima to plot and things look great, speed is good etc. the x-axis represents time and the ticks are in minutes . I can zoom in as much as i want until i dont see anything.. what i want to do is restrict the zoom to the lowest data point resolution which is of 100ms. how do i do that ?
13:57 run4flat you would tie into one of the Axis events
13:57 * run4flat checks the Axis docs
13:57 vicash i looked at your code a few months ago and the zoom was extrapolated based on the points around it..
13:58 run4flat The specific event would be the ChangeBounds event, for the x-axis
13:58 vicash ok i will look it up . thanks
13:59 vicash so can i use this ChangeBounds event to restrict both zoom-in and zoom-outs ?
13:59 vicash because if i zoom out i can make the plot look like a dot
13:59 run4flat yes, I believe so
14:00 vicash ok i will try it out today
14:00 run4flat in the x-axis spec, you would say something like this:
14:00 run4flat x => { onChangeBounds => sub { ... code ... } }
14:01 run4flat But, ChangeBounds isn't called with any useful information
14:01 run4flat so you would have to query if the $max - $min is too small, or large, and then reset them
14:02 vicash ok.. i will play around and see
14:02 run4flat yep, bug me if you have questions
14:02 run4flat BTW, that sort of thing is a feature I've been thinking about adding to the axes
14:03 run4flat though how to do it in a way that gracefully handles both log and linear scales is still up in the air
14:04 vicash log scales... i never thought of using those things
14:04 run4flat right, so for now we'll keep it simple and focus on the ChangeBounds notification. :-)
14:04 vicash ok, i will play around and see if i need to edit code, then maybe i will send you a pull request
14:05 vicash if ChangeBounds solves my problem that will be a quick fix
14:06 * run4flat nods
14:15 run4flat I just looked into the event propagation for ChangeBounds. ChangeBounds will always emit a Paint notification to the parent widget. However, your custom ChangeBounds callback code will be executed before the underlying plot's Paint callback gets issued.
14:15 run4flat This means you have the opportunity to clean up the axis bounds before the paint gets issued.
14:16 run4flat it also means that any zoom out too far or zoom in too far will *always* repaint the plot
14:16 run4flat ok, that's enough blather from me. :-)
15:28 vicash thanks.
16:57 webart I wonder if PDL is "statistical" enough for someone to use it in an academic publication  http://www.jstatsoft.org ... then get tenure ... then kaching!
16:59 webart There's PDLStats  of course but I think plain old PDL's mathematical mojo in analyzing imagery and large data sets (does analyzing photos of distant galaxies count as stats?)  should make it eligible :)
21:02 vicash left #pdl
21:03 run4flat webart, I think that one could make an argument for publishing stuff in jstatsoft
21:03 run4flat whether the editors would buy that argument, on the other hand, is another question
21:04 run4flat And for reproducible code repositories, I would use stratopan myself
21:06 sivoais There was an article in 2004:
21:07 sivoais Giovanni Baiocchi. Using Perl for Statistics: Data Processing and Statistical Computing. <http://www.jstatsoft.org/v11/i01>
21:12 webart nice :-)  I must've used "perl" in their search tool  ... anyway the more the better!!
21:12 webart we need more perling profs with tenure and large research budgets :)
21:14 sivoais Web articles would be best, imo.
21:58 run4flat all of the above
21:58 run4flat anyway, I'm outa here
21:58 run4flat have a good weekend!
21:58 run4flat o/

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