« first day (193 days earlier)      last day (2208 days later) » 

8:07 AM
@Sathya I got another interesting answer, though it would require me to create a CVS file, so that's a bit overkill I reckon
1
A: How can I improve my INSERT statement performance?

f00the fastest way to insert 4 million rows (16MB of data) would be to use load data infile - http://dev.mysql.com/doc/refman/5.0/en/load-data.html so if possible generate a csv file then use load data infile.. hope this helps :) EDIT So I took one of your original data files rolloff.dat and wro...

 
so the main performance hit is python building the query to send to MySQL, not running the query?
 
yeah I think so
but the top answer takes away that problem
I can live with 6 seconds
for now
 
6 seconds for how much data?
 
making a CVS file surely will last longer
reading a ASCII file, parsing it, stripping it of zeros and building the query then inserting it :-)
 
6 seconds doesn't sound so bad then I suppose
 
8:16 AM
no and I would make the user point to a folder, start parsing all the data
then when they 'commit' build the query and insert
perhaps I would even just start doing that already right after they've pointed to the folder (well they might have to click Ok to confirm the right folder first :P)
I still need to tackle that darn animation thing though
my SO question 'looks' answered, but Ignacio's answer doesn't 'solve' it
4
Q: How can I speed up an animation?

Ivo FlipseI'm trying to create a Matplotlib animation of my paw data, where you can see the pressure distribution on the entire pressure plate over time (256x64 sensors for 250 frames). I found a working example on Matplotlib's own site and managed to get it working on my own data. However the 'animation'...

3
Q: A call to arms for the Gaming blog!

Ivo FlipseI'm looking for help from users who want to help breath new life into the Gaming blog! On Super User we got officially blessed by Jeff and I honestly believe that Gaming should be able to deserve one as well. On Super User we mainly started out with Questions of the Week, which help create a ste...

And I'm looking forward to your Gaming diary ;P
 
@IvoFlipse yeah, background loading could be a handy fix, in fact it's handy for anything hard-drive related
@IvoFlipse that's unfortunate, I'd assumed that was done with :/
 
yeah I guess I should worry about less performant computers :P
no I never solved it
 
and gaming are starting an SE blog as well, good good :)
if you wanted my game diary you should of pestered me before finishing DA2 ;P
 
started reading books, in the hope to decipher the glumpy code
well then I expect a review ;P
 
lol
how do I rank on gaming now nowadays....?
ooo, top of second page
 
8:25 AM
woehoe ;)
 
is the idea is to get traffic/blogs etc flowing on blogoverflow before making it official?
 
yeah, or rather to show we can maintain it without needing top down approval or guidance
 
f/e, it makes sense to have a "beta" of the blogs to start with
 
indeed
but Gaming is so much easier than Super User, there's a lot more 'fun' content :P

 Gaming Blog Editor room

Wherein we leverage the leak power from our lazers to do this ...
 
@IvoFlipse lol, well that is kind of the point
 
8:28 AM
indeed it is
I suggested some questions to be converted, now to find minions to write them :)
because my next target it Fitness :P
 
lol
conquering the community one site at a time? ;P
 
hehe something like that :P
 
well, I might be able to lend a hand - but I don't think I'm really part of the community on gaming; just someone who turns up and asks/answers occasionally...
 
well I stopped visiting the room, because it was even worse than Root Access :P
 
lol
 
8:33 AM
now I've left that for now too, because I just can't get things done
 
I've only visited the room once IIRC, to ask about migration of a question to them :/
@IvoFlipse so you end up in here instead ;P
 
yeah, but at least here I can ask for feedback on my work
 
f/e
my efforts to stripe my old HDD's have produced some potentially interesting results
games seem to load faster, but "normal" apps slower...
 
strange :S
 
yeah, I'm thinking it's an access time issue
or that ReadyBoost is speeding up the normal apps when not striped
started gathering as much timing data as I can, will switch RB off at some point and retest, see if it is making a difference or if things are just odd
 
8:37 AM
ReadyBoost doesn't work when it is striped?
 
no, but it caches commonly used files I believe - files I've just copied to a striped drive won't be commonly used
 
Any kind of testing can be interesting to write about, if only if the results are 'odd' or unexpected
ah, that makes sense
 
who knows, it's speculation at the moment, I'll measure it without RB soonish and see what happens :)
speaking of odd, I may have fixed my gfx-card-not-starting-up-when-the-computer-boots problem
by changing my DVD drive
 
that is odd
my problems 'seem' gone now too
haven't plugged in that usb headset anymore though
any reason why the DVD might conflict?
 
@IvoFlipse shrug but it'd started making unpleasant clicking sounds before POST (with no disk inserted)
 
8:46 AM
This Glumpy code is ridiculously fast
I just need to ffs rip out the formulas and insert my data instead! :@
rant off
 
lol
should be easier than writing all the code yourself thou...
 
indeed
well one problem I just spotted is that I need to make sure I have float32's, so I need to convert my arrays
> Z = numpy.random.random((32,32)).astype(numpy.float32)
 
whoa
 
though it seems it would just be a matter of using that last part I reckon
 
clicking on your flag weights provides a history - when did that happen?
 
8:49 AM
last night?
 
nice
 
it doesn't show what has been done with my flags though :P
 
@IvoFlipse doesn't it? it does for me...?
@IvoFlipse 22 pages? wow...
 
no it only shows you what the moderators did to the post, not whether they found the flag valid
 
ah, good point
well, that's probably for the best, right? people are less likely to argue?
 
8:53 AM
I'm not sure if it's public
can't see it incognito, I was expecting it to be a mod feature perhaps
 
well flag weight is private, so it makes sense for this to be as well
 
@DMA57361 oo nice
 
9:51 AM
 
I don't get what's so interesting about Instagram
 
oops
 
but then again, I don't have any friends to share the pictures with :P
 
:)
@IvoFlipse add me, add me :P
 
don't have an iphone either :P
and the ipad doesn't take pictures :P
 
9:53 AM
well use picplz :)
 
I have it
never even tried it
 
I don't have pics to share plz :P
This code still feels like magic to me
they use a decorator on window
which get's called on_idle, so I guess that's after every draw
it has a variable dt, that's not being declared nor calculated anywhere :S
 
10:08 AM
@IvoFlipse you should start taking so that you can share plz :p
@IvoFlipse it has the same effect on me :P
 
I read the chapter on decorators, but it was so boring I kept falling asleep :P
I reckon the @window.event creates temporary variables like dt and items
items is a list
 
what does the windows.event() function do?
 
that consists of a glumpy.image.Image object and matplotlib.image.AxesImage
 
because this indicates that:
@window.event
def on_idle(dt):
is the same as
def on_idle(dt):
    #[function body]
on_idle = window.event( on_idle(dt) )
I'm guess it's just a posh way of assigning that function to run when the window triggers an idle event
 
decorators, if I understood correctly, allows for classes to inherit from functions and to do some fancy calling
the reason to do so, was that else sometimes a class/function would be used at the other end of a script, making it less obvious what is being done
either way, the problem is that I don't know how to replace the 'updating' of image.data[...] with my own :\
 
10:30 AM
ah I see, yeah, that's a bit odd...
 
> for image, axis, alpha in items:
the strange thing is, I think that only 2 things come out of there :S
I haven't seen an alpha :S
> ([[<glumpy.image.Image object at 0x059DE490>, <matplotlib.image.AxesImage object at 0x059DE3B0>, 1.0]], <type 'list'>)
that was print(items, type(item))
oh wait I see, 1.0 at the end
@window.event
def on_idle(dt):
    global Z
    for image, axis, alpha in items:
        for frames in range(248):
            image.data[...] = Z[:,:,frames]
            #print(sum(sum(Z[:,:,frames])))
            image.update()
    window.draw()
this works or rather, doesn't crash and Z[:,:,frames] is a different frame all the time
but the window just shows the last image :\
 
@IvoFlipse as in it doesn't show an animation, just the last frame?
 
indeed
 
they increment X and Y each loop:
X += numpy.pi/15.
Y += numpy.pi/20.
 
You'd almost suspect image.update() not to work
 
10:42 AM
do you need to increment a frame counter each loop?
 
oh wait
I should pass an iterator or something
that on_idle asks for the next frame
 
what about just doing
frame = 0

@window.event
def on_idle(dt):
    global Z, frame

    frame += 1

    for image, axis, alpha in items:
        image.data[...] = Z[:,:,frame]
        #print(sum(sum(Z[:,:,frame])))
        image.update()
    window.draw()
 
frame = 0

@window.event
def on_idle(dt):
    global Z, frame

    frame +=1
    for image, axis, alpha in items:
        image.data[...] = Z[:,:,frame]
        #print(sum(sum(Z[:,:,frames])))
        image.update()
    window.draw()

window.mainloop()
you mean that? :P
 
that's what I just said ;P
 
I had exactly typed what you just said :P
lol
 
10:44 AM
great minds think alike and all that jazz ;D
 
works
 
but, does it work?
 
and pretty fast
 
awesome
 
interpolation looks horrible
let's put it on cubic
still ugly :P
 
10:45 AM
lol
but, the animation is working?
 
yup
takes about 5 seconds to loop through the whole thing
let's see if I can turn on the FPS again
52.98 fps, 58.55 fps
nice :D
bilinear interpolation: NICEEEEEE
 
what FPS were the recordings made in?
 
how many frames a second does the pressure plate record?
 
ah 125 Hz
 
10:53 AM
and that is looking very good
 
indeed
I'm very pleased
Time to answer my own question
 
indeed
 
Guess this is no longer correct :P
I am Harry f*cking Potter :P
Sadly, I can't make the boilerplate code magically appear
 
rofl
 
0
A: How can I speed up an animation?

Ivo FlipseI found Joe Kington's answer that mentioned using Glumpy instead. At first I couldn't get it to work on my own data, but with some help on chat we managed to figure out how to adapt one of the Matplotlib examples that come with Glumpy to work on my data. import numpy, glumpy from glumpy.pylab im...

 
11:05 AM
woo, I get linked :)
 
I almost considered letting you answer
but I need to answer more questions on SO, else I'll never get more rep :P
 
and to be fair, you did the majority of the work - I just helped unroll a loop
 
well this room has proven it's worth ones more ;)
 
is the 58FPS something you might have direct control over? or is it just coincidence that it's close to the common 60Hz refresh rate of screens?
 
I think it's on purpose
that way they can optimize for performance
 
11:11 AM
fair enough, presumably it's fast enough for your needs thou?
 
for sure it is :)
Now I need to start working on some of the GUI elements...
I wonder how hard that's going to be
 
:/
 
some nice tabbed + Ribbon interface :P
 
will that still be using wxPython? or something different to go with glumpy?
 
oh glumpy can be embedded
it's just for the drawing
 
11:16 AM
excellent
 
actually it let's matplotlib draw, then updates it blazingly fast
I'm going to wade through wxpython's example
I need a database-ish screen
some database fields and buttons, a tree with some results perhaps
120 fps @ 5k pixels resolution
 
getting an error... but presumably a very fast glumpy animation?
 
heck no: Epic RED camera's :P news.ycombinator.com/item?id=2381076
 
oic
 
all the new movies are being shot with such things, I believe Peter Jackson ordered a bunch of The Hobbit
 
why do I have to sign in for that link...?
 
11:59 AM
@DMA57361 that's amazon's cloud streaming thingy
 
who's what now?
 
think of it as dropbox by amazon @DMA57361
 
fair enough
hmmm.... maybe I should pay more attention to these things? but sometimes it's just too much effort :)
 
yeah sorry I gave the link after logging in, so it's not their 'frontpage'
 
@DMA57361 well it was announced overnight @DMA57361 caught it just before I left for work
you should have a glimpse at techmeme.com - it'll give you a general snapshot of what's hot
 
12:10 PM
I already checked Hacker News and Techcrunch :P
not much left :P
 
chucked? /// nev mind
 
12:30 PM
oh boy even a simply GUI with some buttons can be a 1000 lines of code :O
 
yep
painful, isn't it? :(
 
pretty much
 
tbh that's the main reason I've never got in to GUI building as yet
 
well I'll just first try and get it working with a bare minimum of features
hello @JonasHeidelberg and @Icode4food
oh my @JonasHeidelberg is really a freshman for Stack Exchange :)
@DMA57361 shall I use such an awful sidebar (its in the middle now :P) to navigate through my app?
 
sure, why not :P
 
12:40 PM
unless I can fit all my required functionality onto one toolbar
 
indeed, but a ribbon helps with that by being several toolbars
 
you mean because it's so wide?
 
no, because the tabs at the top change the ribbon in to a different bar
 
ah yes, but then I still need to supply some UI element to switch 'modes'
 
ah yes, I recall now - we had a discussion about nested ribbons at one point didn't we :)
hmmm.... maybe that side bar's looking helpful ;P
 
12:44 PM
yeah, hence I thought a non so fugly side bar would be useful
hehe the option it has are quite useful actually
 
or, could the first "block" of commands in the ribbon be your mode switcher - and just duplicate that block on to each ribbon section?
 
you can make it pretty narrow and let it expand when you need it to
hmmm that's an interesting option too
put this inside the ribbon :P
or at the bottom of the window, haha
 
lol
 
only 230 lines of code
 
@IvoFlipse cos most of it is hidden in library code?
 
12:49 PM
might be
I should figure out how this demo works, because it runs all the other demos
 
ooo, yeah, that'd be handy
is this stuff all in the demos that came with the wxPython install?
 
especially because it's almost like a framework for running my own code
I think you have to download this separately
the wxpython demo for windows
 
@IvoFlipse f/e, either way I think I'll get hold of them, they look interesting
 
as soon as I have some code you can mess with that as well
 
hurrah, I like having things to break ;P
 
1:01 PM
painful!
 
not something you're going to base your design on then?
seriously thou, how many top/tool/side bars do you need? :/
 
ill do my very best not to
I hope as little as possible
I'd prefer not to use any pop ups
but with your suggestions for the navigation through the Ribbon
I could have the screen itself change context with the 'tabs' of the ribbon
like Add to database, Search, Edit, Manage
I hate how I need to add all these options that are just tedious to add :P
 
Hello @Ivo :-). Freshman is the right word I guess... and I only opened the chat (background tab) to find out what "Fake Programmers" stood for ;-)
 
hehe, freshman CS?
 
1:10 PM
@DMA57361 I guess if I separate the UI code from the code that acts upon it, it wouldn't be a lot more work, it just feel tedious
Computer Science
 
no, no
I didn't mean it that literally
I'm in Biophysics
 
hehe, oeh interesting
 
trying not to get addicted to S[OFU]
 
What kind of things do you do mostly?
 
still doing okay so far ;-)
 
1:11 PM
@IvoFlipse true, but separating them might make life easier in the long run
 
So far :P
 
my topic is adaptive optics in microscopy
computer-wise I had to play with LabView for a while, now I'm back to mostly Matlab
 
Labview doesn't really have a strong community on SO, they all sticked to their old school Lava or NI forums
Matlab? I switched to Python, because I didn't want to burden my employer with the license fee :P
So I assume you work at a university or still studying?
 
I'm hoping not to need to go back to Labview :-p
yes, still uni
??? what did I do?
"perform action?"
 
what do you mean @JonasHeidelberg ?
 
1:15 PM
I typed "yes, still uni (newline) P h D (without the spaces)"
and it seemed that the three letters P h D were somehow interpreted?
the chat system told me "You can perform this action again in 1 seconds"
 
rate limiting @JonasHeidelberg
 
oh, I was using too many newlines?
 
not necessarily too many new lines
too many messages in a timeframe
 
right, I'm off, see you all later...
 
okay .... I'll type slower ;-)
 
1:17 PM
ttyl @DMA57361 :)
 
Matlab... the licenses were present when I arrived, I took over some legacy code... never broke out of the habit. But I've heard before that Python would be the way to go... currently its the inertia which is keeping me ;-)
 
if it works for you, don't let me stop you :)
 
:-)
I already noticed that I "Jonas" answering Matlab questions on SO is taken and picked a longer name... even though I like the idea that nickname strings don't need to be unique :-)
 
but surely not @JonasHeidelberg right? ;)
 
No, there is even only one other user who has elevated Heidelberg into his username ;-)
(@IvoFlipse: do I need to address you directly whenever I talk to you? Does that make live easier?)
 
1:28 PM
naah, I mainly focus on this chatroom anyway, so you don't have to constantly @-address me :)
 
:-)
So the "Fake" in the room name is to not scare newbies away?
 
rather to point out I'm a newbie myself :P
 
1:46 PM
so what can I expect with "adaptive optics in microscopy"?
and hi @PaulBrown
 
lots of users in here today, why I wonder
 
I guess we're advertised on the frontpage of SU
 
@Sathya, yes I came here through some frontpage link
 
hmm
 
Adaptive Optics in Microscopy is already quite established for people "looking you into the eye" (= eye doctors); for other samples it is a hot topic, but only starting to show results
Na Ji's work in Neurobiology is quite exciting in my eyes
 
1:50 PM
I'm not familiar with Na Ji's work ;-)
 
you would have surprised me otherwise ;-)
She (as many people) wants to understand how the brain works
so she is looking a neurons inside a rat brain
using adaptive optics she gets better data
so she'll probably be able to look deeper, which is important to understand the brain as a whole
When people don't know adaptive optics, I always say it is like adding contact lenses: make good visibility under bad conditions
 
so you look into the brain through the eyes?
 
(in her (Na Ji's) case, "bad conditions" means to look through several other cell layers)
if you look in through the eyes, you can't see further than the retina. This is interesting, and eye doctors do this all the time. But if you want to understand the cortex (the big outer part of the brain) you won't see it through the eyes
so you are stuck with animal experiments :-/
meaning that you do a small operation to open the brain
or at least make the skull thin enough to become transparent
 
hoeray for mice :P
 
yeah :-)... and zebrafish, and fruit flies, and ... ;-)
 
1:56 PM
hmm yeah that might be easier
 
okay, I'll start to get my second SU question off the ground
maybe the community can help me avoid to reinstall my WinXP box...
 
heading home, later fellows @IvoFlipse & welcome to SU @JonasHeidelberg
 
by @Sathya :)
who knows :)
though for tech support Root Access is the place to be :)
 
2:29 PM
Hmm, I just posted it as a regular question... how appropriate is it to refer to regular questions in the chat?
 
@JonasHeidelberg it's perfectly fine...
thought I suggest a move over to root as @IvoFlipse suggested
This room is mainly for wannabe programmers :P
 
okay, thanks :-)
cu there
 
2:53 PM
Sorry I was on Skype with a phd-er who is doing a study after how researchers switch to developers (or something of the kind)
 

« first day (193 days earlier)      last day (2208 days later) »