Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2013-03-06

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

All times shown according to UTC.

Time Nick Message
00:21 MeierM joined #pdl
01:55 tckb joined #pdl
01:56 tckb hi fellow mongers!
01:56 tckb anyonse alive here?
02:11 tckb joined #pdl
02:13 jberger wow, here and gone
02:30 tckb joined #pdl
02:31 tckb hey I have troubled installing PLPlot
02:31 tckb any help ?
02:33 jberger what platform are you on?
02:33 jberger PLPlot can be hard to install on some platforms
02:33 tckb Im on MacOSX 10.7.4
02:33 tckb yeah
02:33 jberger yeah, I think thats one of them
02:33 * jberger looks
02:34 tckb Here is my Perl -V output
02:34 tckb http://scsys.co.uk:8002/233760
02:34 tckb and the error http://scsys.co.uk:8002/233759
02:35 tckb any clue how to fix it
02:35 tckb guys at #perl@freenode suggested a downgrade install to 32bit
02:36 tckb I'm pretty new to CPAN and perl
02:36 jberger I can't give you any specific advice, I use linux and don't use PLPlot
02:36 jberger it looks like its complaining that certain functionality didn't get built
02:37 tckb yea thats the troublee
02:37 jberger is there some reason you are using PLPlot?
02:37 tckb there was a recommendation to install it as 32Bit
02:37 jberger could you use on of the other plotting libraries?
02:37 tckb I'm trying to find my way around CPAN
02:38 jberger *one
02:38 tckb actually, I m searching for a decent plotting library in Perl
02:38 tckb and PLPlot seems to be a suitable candidate
02:38 jberger the recent work for plotting in PDL has been on PDL::Graphics::Gnuplot
02:38 tckb any suggestions?
02:38 jberger and PDL::Graphics::Prima
02:39 jberger the first being a connection to the time-honored Gnuplot program
02:39 jberger and the second being a brand-new cross-platform library built specifically for PDL
02:39 tckb oh, I haven't checked Prima yet
02:40 jberger for print-quality plots i would use P::G::Gnuplot
02:40 tckb hows it
02:40 jberger for interactive use Prima
02:40 tckb in terms of speed and useabiliy
02:40 jberger Prima is promising, but its very new and a little rough on the edges
02:40 jberger very fast
02:40 jberger decently usable, especially with the ::Simple layer
02:41 jberger PDL::Graphics::Prima::Simple that is
02:41 tckb cool, gimme a min .. I'l have a look
02:41 jberger the problem is that there aren't great ways of exporting the Prima plots yet
02:41 tckb oh
02:41 jberger for that the gnuplot module is great because gnuplot makes great plots
02:41 tckb exporting to a png/jpeg ?
02:42 jberger easy-peasy with PDL::Graphics::Gnuplot
02:42 tckb yeah I'm familiar with gnuplit
02:42 jberger then I recommend that
02:42 tckb but, not quite sure how its used in Perl
02:42 jberger are you familiar with PDL? (just gauging your knowledge base)
02:43 tckb not really, I'm a matlab/c/java programmer and I am very new Perl
02:44 jberger PDL is rather similar to matlab in concept
02:44 tckb today would be day-1 of perl ;)
02:44 jberger oh ok
02:44 tckb eah
02:44 jberger PDL makes Perl objects representing C-like number arrays
02:45 tckb yeah, I was looking for something to Matlab - Yet comparable to C and free/open source
02:45 jberger we would say, you're in the right place :-)
02:45 tckb indeed!
02:45 jberger http://pdl.perl.org/?docs=MATLAB&title=PDL::MATLAB
02:45 jberger here is a tutorial aimed at Matlab users
02:46 jberger hopefully that gets you started
02:47 tckb I was actually looking for it and accidentally stumbled across the performance comparison of C/Java/Perl .. that served as a kickstart I needed
02:47 tckb Yeah, I've read that tutorial
02:47 jberger also this general tutorial http://pdl.perl.org/?docs=QuickStart&title=PDL::QuickStart
02:48 jberger ok
02:48 jberger cool
02:48 tckb cool!
02:48 tckb Any examples / screenshots of PDL::Graphics::Prima
02:48 tckb ?
02:49 jberger the examples in ::Simple documentation should work
02:49 jberger not sure if there are any screenshots
02:49 tckb just to get an idea of it
02:49 jberger https://metacpan.org/module/PDL::Graphics::Prima::Simple
02:50 jberger I could take a quick screenshot :-{
02:50 jberger oops
02:50 jberger :-P
02:50 tckb :P
02:50 tckb I am waiting … :)
02:52 jberger http://www.youtube.com/watch?feature=player_detailpage&v=YnW1hTVQYfA#t=867s
02:52 jberger thats quicker
02:52 jberger :-)
02:52 jberger from my YAPC::NA talk last summer
02:52 tckb watching ..
02:53 jberger that reminds me
02:53 jberger http://www.youtube.com/watch?feature=player_embedded&v=IE-vnnRWiOg
02:53 jberger that's David's talk about PDL::Graphics::Prima
02:53 tckb any good yet solid fast track tutorial on Perl ?
02:53 jberger David == run4flat == PDL::Graphics::Prima developer
02:53 jberger hmmmm
02:53 jberger well the Modern Perl Book is the new go-to
02:54 jberger let me get you that link
02:54 tckb "fast track" ==> video tutorials should be great! ;)
02:54 jberger but first, here are a few more videos: http://perl4science.github.com/videos/
02:54 tckb ah, thats what I meant
02:55 tckb thanks!!
02:55 jberger http://www.onyxneon.com/books/modern_perl/index.html
02:55 tckb I have a lot of experience with OOP
02:55 tckb in Java
02:55 tckb and am looking something parallel in Perl
02:55 jberger Perl's OOP is a little different, and PDL's is perhaps even a little more different
02:55 tckb I've heard Moose does that ?
02:56 tckb yeah
02:56 jberger yep, thats the new go-to for OOP in Perl
02:56 jberger PDL is much older than that
02:56 jberger and its object is more deeply tied to C of course
02:56 tckb hmm…
02:56 tckb yeah
02:56 jberger but there has been recent work on parallelizing
02:57 jberger there have been several recent threads here: https://groups.google.com/forum/#!forum/the-quantified-onion
02:57 jberger including MCE (many core engine) + PDL
02:57 tckb I'm actually a student researcher with deep interest in Audio analysis but unfortunately I haven't found any comprehensive libs/modules un Perl
02:59 jberger also David's recent PDL::Parallel::threads https://metacpan.org/module/PDL::Parallel::threads
02:59 jberger I have heard (!) other people using PDL
02:59 jberger for audio
02:59 jberger not sure there are many modules for it
03:01 tckb Nah, actually PDL offers a lot of tools for computations in  Linear Algebra
03:02 tckb kinda like BLAS and LAPACK
03:02 tckb for c/c++
03:02 jberger There was this: http://darwintunes.org/
03:02 jberger code at: https://github.com/bobular/DarwinTunes
03:04 jberger actually I think PDL uses some of those internally
03:05 tckb yeah, i'd expect that
03:05 tckb looks interesting will give it a look
03:05 jberger enjoy
03:06 jberger and for the record, I know some people have gotten PLPlot to work
03:06 jberger I just can't help you there
03:08 tckb haha
03:08 tckb PLPlot looks more 'beautiful
03:08 tckb but code looks pretty 'unreadable' to me
03:08 tckb just saying
03:10 tckb thanks for the help!!!
03:11 tckb I'll be back with more troubles later !
03:11 tckb ;)
03:14 tckb left #pdl
14:42 run4flat jberger++
14:42 jberger thanks
14:51 run4flat btw, PLplot is a perfectly fine plotting library
14:51 run4flat plenty fast for point and line plotting
14:52 run4flat I used to use it
14:52 run4flat My problem was that I couldn't see how to make it interactive without be extremely cludgy
14:52 run4flat and the nontrivial dependence on CMake made me a bit wary
14:53 run4flat Knowing what I know now, I bet I could create a Prima plotting device for PLplot, and we could get very nice interactive plotting using it
14:53 run4flat My only gripe would be that you must specify the axis min and max. I absolutely detest that
14:54 run4flat Again, having worked out the algorithm for automatic scaling, I could probably cook up something for PLplot and submit a patch to allow autoscaling
14:55 run4flat but that would require a big investment of my time into the library
14:55 run4flat and I think it's more fun to write my own with Perl, instead. :-)
14:55 jberger I gotcha
14:56 jberger but if the guy can't get it installed, and I don't know the tricks to installing it, its time to move him to something else
14:56 run4flat sure
14:57 jberger a day 1 perler, gotta make sure he doesn't have too many barriers to entry :-)
14:57 run4flat This is precisely the problem with PLplot and PDL, that Doug doesn't have the time to advocate for it
14:58 run4flat it seems that Zentara doesn't, either, as (s)he has been quiet lately, too
15:02 jberger :-( http://www.cpantesters.org/distro/A/Acme-Ford-Prefect.html
15:02 jberger no mac reports yet
15:02 jberger but fails from Free/OpenBSD
15:02 jberger is a bad sign
15:05 chm joined #pdl
15:08 chm hi all
15:09 jberger chm o/
15:10 chm ping run4flat
15:10 chm hi joel
15:11 chm It was nice to see more interest re PDL today.  I guess another reason for David to update the docs for PGP/PGPS
15:11 chm How's the writing coming along?
15:11 jberger slow and steady
15:11 jberger but coming
15:11 chm ETA?
15:12 jberger I'm trying to schedule the defense for the last week in april
15:12 chm Good luck
15:12 chm I see they are having another perl conference here in DC-Baltimore area.  I'm thinking to give a PDL presentation re 2.006
15:15 jberger nice
15:15 jberger when is that?
15:15 chm Last part of April, I believe.
15:15 jberger I've been looking for an excuse to visit my sister in arlington
15:15 jberger hmmmm
15:16 chm David presented at last year's where I got to meet him in person.
15:16 jberger thats not probably going to work then :-)
15:16 chm The more the merrier.
15:16 Molaf__ joined #pdl
15:18 chm joel: Looping back to the PDL OO discussion earlier
15:18 chm I'm pretty excited about the possibility of making PDL-3.000 a Moo
15:19 chm based implementation.  That adds a lot of flexibility going forward without
15:19 chm much code penalty.  We can pull in Moose for the "big guns" when needed.
15:19 chm My thought is that we have some C-level hooks to support method modifiers and
15:20 chm such so that we could get flexibility and high performance when needed.
15:20 chm For example, I think an "around" modifier plus Role composition could be
15:21 chm used to generate or compile PDL code to use a GPU or other HPC environment
15:21 jberger is there a good way to attach the data to the Moo-based object
15:21 chm Same way as in PDL, the C level makes methods and properties.  Some are done at the perl/Moo level.  Others in compiled C/C++/... for performance.
15:22 chm The current PDL OO framework is just the original hash based objects with some magic to allow the C stuff to be used for "the fast stuff"
15:24 chm Another plus of converting from hash to Moo[se] is to regularize the objects used through out and to allow perl prototyping of new features that could be migrated to C as needed for performance.
15:24 chm It is amazing what can be cooked up with a few method modifiers and roles.  :-)
15:25 chm At the moment, I'm focusing on wrapping up PDL-2.006.
15:25 chm David/run4flat and Craig are getting a KISS plotting default to go with the release and PDL Book
15:26 chm I'm waiting with bated breath for something to come in.
15:26 chm I was hoping for a release this Friday, but I'm not sure where we stand.
15:27 chm PDL-2.004_995 is testing *very* well so once the graphics are in place, it will be a sweet release.
15:32 run4flat chm, morning
15:33 run4flat working, working
15:33 chm hi, snowing here so I'm at home for the day.
15:33 run4flat nice!
15:33 run4flat well, good news: I got an offer from Dickinson!
15:33 run4flat :-)
15:33 chm How are things going PGPS/PGS-wise?
15:33 chm Congratulations!
15:33 run4flat eh, sorta coming
15:34 * jberger <3 method modifiers and roles
15:34 run4flat you'll shake your finger at me, though:
15:34 run4flat I worked out a new binning framework
15:34 run4flat and it's basically working
15:34 run4flat So, I rewrote hist_plot to use it
15:34 run4flat and *now* I'll work on the simple stuff
15:34 chm That's ok.  I've resigned myself that the release won't be happening on Fri.
15:35 chm I would prefer that you get the pdldoc support and the simple hist_plot in as
15:35 chm usability and clear docs and searchable docs would help the visibility of PGP/PGPS
15:35 chm Just be sure to let me know when you have something to test
15:35 chm so I can plan the formal release.
15:36 run4flat sure
15:36 chm I don't want to see gnuplot support get all the credit.  :-)
15:36 run4flat it's so difficult to coordinate all of my efforts. :-)
15:36 run4flat haha
15:36 run4flat on the one hand, I want to add features
15:36 run4flat on the other hand, I want to add docs
15:36 run4flat and, I still feel bad that I don't actually have any tests
15:37 run4flat :-/
15:37 run4flat so, I think I have histogram stuff in a much better place
15:37 run4flat but it's not well documented
15:37 chm If you were to release the current code with a simple hist plot, the PGS support, and =head2 + add_module, that would be great for 2.006
15:38 chm We could refine and improve/update everything afterwards.  I *really* want PGP/PGPS to be useable with PDL-2.006
15:38 run4flat Well, first and foremost, I need to have something on CPAN for which a PGS works
15:38 chm True
15:38 run4flat *then* I can work during the following week(s) on improving docs access
15:39 run4flat so, hist plotting works
15:39 run4flat I need to cut a quick CPAN release
15:39 chm That sounds like a better plan: just PGS support and hist_plot for a release this week to support PDL-2.006 next week-ish?
15:39 run4flat and then I need to do the PGS
15:39 chm Docs post PDL-2.006
15:39 run4flat hopefully by Friday
15:39 run4flat I think it's possible
15:40 run4flat I mean, the documentation for the quick release to CPAN will be a bit disorganized
15:40 run4flat but it will at least *work*
15:40 chm Any documentation is better than none.  Searchability helps alot.  At least to find stuff to read...
15:41 run4flat yes, I know, but first thing's first: PGS support
15:41 run4flat that way PDL can get its release out the door
15:41 chm Excellent.  Keep me posted.  I've got to get back to some things here.  Will check back later.  o/
15:41 run4flat \o
15:45 jberger PGS++
15:45 jberger wow, step away for a minute and the backlog gets huge
15:46 jberger and then chm leaves
15:46 jberger :-)
15:46 jberger oh well
22:57 Meiermann joined #pdl
23:54 opkick joined #pdl
23:54 opkick [pdl01] run4flat pushed 2 new commits to master: http://git.io/5koVaQ
23:54 opkick pdl/master 690b491 Chris Marshall: Add $PDL::PREFIX_RE to perldl and pdl2...
23:54 opkick pdl/master c600415 David Mertens: Fixed pdl-from-string doc error in QuickStart
23:54 opkick left #pdl

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