fluid-work IRC Logs-2013-02-01

[00:42:30 CST(-0600)] <thealphanerd> Bosmon: thanks for the suggestions

[00:43:02 CST(-0600)] <Bosmon> NERDD!

[00:43:12 CST(-0600)] <thealphanerd> wanna see my homework?

[00:43:35 CST(-0600)] <thealphanerd> http://cl.ly/1T1R10170K0F

[00:47:03 CST(-0600)] <Bosmon> Wow

[00:47:05 CST(-0600)] <Bosmon> Nice colours

[00:47:55 CST(-0600)] <thealphanerd> thanks

[00:48:58 CST(-0600)] <Bosmon> Funky transfer function

[00:49:24 CST(-0600)] <thealphanerd> all julius's deal

[00:49:29 CST(-0600)] <thealphanerd> umm

[00:49:41 CST(-0600)] <thealphanerd> not sure how I meant to phrase it

[00:49:50 CST(-0600)] <thealphanerd> it was designed by the professor

[00:49:51 CST(-0600)] <Bosmon> Yes

[00:49:54 CST(-0600)] <thealphanerd> I just plugged and chugged

[00:49:59 CST(-0600)] <Bosmon> I didn't realise that was the transfer function of a delay....

[00:50:05 CST(-0600)] <Bosmon> Pretty bizarre

[00:50:08 CST(-0600)] <thealphanerd> yup

[00:50:14 CST(-0600)] <thealphanerd> both the cases are bizarre cases

[00:54:25 CST(-0600)] <Bosmon> I guess it must be different in discrete time

[00:54:48 CST(-0600)] <Bosmon> In continuous time the transfer function of a delay is e^{-kt}

[00:55:06 CST(-0600)] <thealphanerd> hmmm

[00:55:12 CST(-0600)] <thealphanerd> I like your use of latex

[00:55:16 CST(-0600)] <thealphanerd> LaTeX

[00:56:46 CST(-0600)] <Bosmon> Oh yes, that's right

[00:57:07 CST(-0600)] <Bosmon> The Z-transform just uses polynomials, not exponentials

[00:57:21 CST(-0600)] <Bosmon> So it makes sense all the transfer functions get whacked in the same way

[00:57:35 CST(-0600)] <thealphanerd> we used state space models last week

[00:57:38 CST(-0600)] <thealphanerd> that was really cool

[00:57:55 CST(-0600)] <thealphanerd> where your transfer functions could get to a super high order and compilated

[00:58:00 CST(-0600)] <Bosmon> In both cases the transform of a delay just takes the form of one element of the transform

[00:58:01 CST(-0600)] <thealphanerd> you still had a second order state space model

[00:58:08 CST(-0600)] <Bosmon> http://web.abo.fi/fak/mnf/mate/kurser/transfunc/transfunc.pdf

[00:58:15 CST(-0600)] <Bosmon> Section 1.6 of this shows the difference

[00:58:25 CST(-0600)] <Bosmon> In discrete time you transform with z^{-k}

[00:58:46 CST(-0600)] <Bosmon> Whereas in continuous time you transfrom with e^{-kt}

[00:59:25 CST(-0600)] <Bosmon> Aha

[00:59:30 CST(-0600)] <Bosmon> Section 2.6 is even more interesting

[00:59:43 CST(-0600)] <Bosmon> It explains that the Fourier series can be derived by evaluating the Z-transform on the unit circle

[00:59:54 CST(-0600)] <Bosmon> There are so many of these bloody transforms

[01:00:02 CST(-0600)] <Bosmon> And it's very useful to see how they can be related

[01:00:49 CST(-0600)] <Bosmon> What's fascinating is that complex variable theory will tell you how this seemingly impossible thing might be true

[01:01:06 CST(-0600)] <Bosmon> Since you would think if you're just looking at something on the unit circle, you've lost essentially all the information in the function

[01:02:07 CST(-0600)] <Bosmon> But as it turns out, there are very few kinds of things that complex functions can be, and for the most part just knowing what something does on the unit circle should tell you everything there is to know about it

[01:03:48 CST(-0600)] <Bosmon> The other thing that this exercise will tell you, is that you'll have essentially no luck approximating any kind of delay filter by means of one of your standard low-order rational filters, like the biquad

[01:04:01 CST(-0600)] <Bosmon> Unless, I guess, miraculously the delay happens to be exactly 1 or 2 : P

[01:04:36 CST(-0600)] <thealphanerd> hmmm

[01:07:25 CST(-0600)] <Bosmon> Although I guess for any delay, there will be ONE best biquad filter that tries to approximate it

[01:07:30 CST(-0600)] <Bosmon> But I bet it would sound really terrible

[01:07:36 CST(-0600)] <Bosmon> That would be a pretty interesting exercise actually

[01:07:41 CST(-0600)] <Bosmon> Might lead to some interesting sound effects

[01:08:03 CST(-0600)] <Bosmon> Probably sounds like WHALES or something......

[01:11:18 CST(-0600)] <thealphanerd> lol

[01:11:24 CST(-0600)] <thealphanerd> sorry for the one word answers

[01:11:29 CST(-0600)] <Bosmon> I guess the quality of approximation would be much better as the delay period becomes much less than the sample period

[01:11:36 CST(-0600)] <thealphanerd> I'm trying to implement the second transfer function using c++ and stk

[08:43:49 CST(-0600)] <jhung> anastasiac: is "JSONcc" the proper term for the transcript format? Are there alternate names for it?

[08:45:02 CST(-0600)] <anastasiac> jhung, I'm not 100% sure. checking...

[08:45:39 CST(-0600)] <anastasiac> jhung, why do you ask?

[08:46:27 CST(-0600)] <jhung> Just curious to see if there is a friendlier name for the format.

[08:47:48 CST(-0600)] <anastasiac> ah, well, not that I know of You could try "JSON Transcripts" perhaps?

[08:48:39 CST(-0600)] <jhung> Thanks!

[10:18:19 CST(-0600)] <jhung> anastasiac: with Amara captions and transcripts, the language is already specified, correct? The user does not need to input this from our plugin/

[10:19:09 CST(-0600)] <anastasiac> well, jhung, they should probably let us know; otherwise we'd have to programmatically parse the url to figure out the language code...

[10:21:35 CST(-0600)] <jhung> Automatic is better since the user may not even know what language the caption or transcript is.

[10:23:52 CST(-0600)] <anastasiac> jhung, the user would have to know which language they're specifying, afaik, because we need to know the language in order to retrieve the caption from Amara

[10:25:21 CST(-0600)] <jhung> anastasiac, but if they were given an URL and no other details...

[10:25:46 CST(-0600)] <anastasiac> the url contains the language code

[10:26:50 CST(-0600)] <jhung> But does the user understand how to parse the language code? Also, what if the URL is long and the language code is somewhere down the string. They may not bother to even look at it.

[10:29:05 CST(-0600)] <jhung> I guess the point I'm getting at is that if there's a fool-proof way of automating some of the form-filling grind, we should attempt to implement it. For now I'll put a language selector, but hopefully some day in the future it will be automatic.

[10:29:13 CST(-0600)] <jhung> Maybe we'll have hover boards by then.

[10:41:02 CST(-0600)] <jhung> cindyli: your update didn't come through.

[10:41:29 CST(-0600)] <cindyli> i'll put up here. thanks for letting me know, jhernandez

[10:41:42 CST(-0600)] <cindyli> sorry, jhernandez, i meant "jhung"

[10:41:45 CST(-0600)] <cindyli>

[10:41:55 CST(-0600)] <jhung>

[10:43:57 CST(-0600)] <cindyli> fluid-everyone: my standup update - I am looking at a failure transcript integration test in IE8. Also starting UIO refactoring with uiEnhancer component

[11:25:38 CST(-0600)] <Justin_o> cindyli: so it seems like there is a timing issue

[11:25:53 CST(-0600)] <Justin_o> if i put break points in the test will pass.

[11:26:39 CST(-0600)] <cindyli> ah, i see. thanks, Justin_o, setTimeout maybe

[11:27:33 CST(-0600)] <Justin_o> cindyli: i added in another console log and it appears that the the onLoadedMetadata event is firing after the the listener for onVideoAndTranscriptsLoaded is fired

[11:27:52 CST(-0600)] <Justin_o> but it could be related to the ordering of events

[11:28:06 CST(-0600)] <cindyli> interesting

[11:30:31 CST(-0600)] <Justin_o> cindyli: now i removed all the break points and it works ;(

[11:31:18 CST(-0600)] <cindyli> haha. how could that happen. refresh and might fail again. same as what happens with build site, Justin_o

[11:31:33 CST(-0600)] <Justin_o> cindyli: tried a bunch of refreshes..

[11:31:44 CST(-0600)] <Justin_o> i might need to get someone else to test it for me i guess

[11:31:52 CST(-0600)] <cindyli> good idea

[11:38:59 CST(-0600)] <Justin_o> cindyli: anastasiac gave it a try.. she had it failing again.. so it looks like it was passing due to some caching.. so with a clear cache.. if you run all the tests.. it hangs. But if you rerun the hanging test by itself it will pass.. then going back to all the tests works

[11:40:50 CST(-0600)] <cindyli> thanks for giving another try, justin. Conclusoin - failure, right?

[11:43:10 CST(-0600)] <Justin_o> cindyli: yes.. strange.. i started commenting out tests.. so it just has to do with the one i have now for using the mammals video with the local transcripts.

[11:43:37 CST(-0600)] <cindyli> ok

[11:49:02 CST(-0600)] <Justin_o> cindyli: setTimeout of at 0 for calling click works ;(

[11:49:41 CST(-0600)] <cindyli> Justin_o: where did u set timeout

[11:49:57 CST(-0600)] <Justin_o> cindyli: the call to anElement.clic()

[11:50:20 CST(-0600)]

[11:50:39 CST(-0600)] <Justin_o> cindyli: so i'm guessing there is a slight timing issue about when the video is actually loaded maybe

[11:50:49 CST(-0600)] <Justin_o> perhaps the element is fully hooked up at this point

[11:51:35 CST(-0600)] <cindyli> Justin_o: isn't set timeout to 0 an equivalent of no timeout

[11:52:55 CST(-0600)] <Justin_o> cindyli: should be

[11:53:34 CST(-0600)] <cindyli> don't quite understand why setting timeout 0 works

[11:57:40 CST(-0600)] <cindyli> Justin_o: how do you wanna fix the onReady events? an aggregate event of "onReady" and "onTranscriptLoaded"?

[11:58:13 CST(-0600)] <Justin_o> cindyli: an aggregate of something.. not sure what the pieces are all just yet, but wanted to double check that we weren't' missing code changes

[12:01:56 CST(-0600)] <cindyli> Justin_o: i don't think it makes sense to include "onTranscriptLoaded" into "onReady". should we try "onCreate"

[12:02:42 CST(-0600)] <Justin_o> cindyli: i'm trying to find out when onTranscriptLoaded gets called now.. but i think definitely we'd want onCreate and afterRender

[12:04:52 CST(-0600)] <cindyli> Justin_o: onTranscriptLoaded gets called when language is chosen and the transcript for that language is fully loaded. loading transcripts is async

[12:05:17 CST(-0600)] <Justin_o> cindyli: i guess the question really is what does it mean for the transcripts component to be ready

[12:05:19 CST(-0600)] <Justin_o> ?

[12:07:28 CST(-0600)] <Justin_o> michelled, anastasiac, cindyli: do you have an opinion on what it means for the transcripts component to be ready

[12:07:43 CST(-0600)] <anastasiac>

[12:08:21 CST(-0600)] <anastasiac> Justin_o, perhaps that might depend (in part) on who wants to know, and why?

[12:09:29 CST(-0600)] <Justin_o> anastasiac: well anyone listening to the onReady event … i know that's pretty vague.. but i guess i mean.. when someone listens to that event.. what should they expect the transcript component to look like at that moment

[12:10:18 CST(-0600)] <anastasiac> ok, so if we don't really know who's listening or why, it might be prudent to not say 'ready' until pretty much everything is actually quite ready

[12:10:35 CST(-0600)] <anastasiac> I'd say that's after the transcript itself has been loaded, click handlers set up, etc

[12:11:08 CST(-0600)] <anastasiac> one question arises: if there's a failure loading the transcript, and we've detected that and displayed an appropriate message, is the component 'ready'?

[12:11:18 CST(-0600)] <anastasiac> it's pretty much done everything it can do...

[12:11:25 CST(-0600)] <anastasiac> but it can't actually carry out its function

[12:13:24 CST(-0600)] <Justin_o> anastasiac: yes.. is that case the same as if you provided no transcripts?

[12:13:39 CST(-0600)] <cindyli> anastasiac, Justin_o, right, if there's a failure loading the transcript or even the transcript has not yet been provided. i lean towards that transcript component ready and the transcript loaded are separate

[12:13:40 CST(-0600)] <anastasiac> ah, right

[12:13:42 CST(-0600)] <Justin_o> i suppose we could also have the case where some transcripts load and others fail

[12:14:21 CST(-0600)] <anastasiac> so one possible meaning of 'ready' could be "I've done everything I can do, there's nothing left to wait for, nothing more is going to happen"

[12:14:31 CST(-0600)] <anastasiac> can't tell from this if there are transcripts or not

[12:15:03 CST(-0600)] <anastasiac> if we combine that ready with error events, perhaps that might be sufficient?

[12:18:17 CST(-0600)] <Justin_o> i'm looking through the code a bit more now.. when we fetch from amara, do we have no way of knowing if there is an error

[12:19:15 CST(-0600)] <anastasiac> Justin_o, I believe that amara does not fire any error codes, or give any such indication. IIRC, in my error-handling branch, I put in a timeout and double-checked something after the timeout

[12:20:43 CST(-0600)] <Justin_o> anastasiac: right.. i remember you mentioning that now

[16:08:33 CST(-0600)] <avtar> fluid-everyone: we're about to switch over to a new dns server at the data centre here in toronto

[16:09:19 CST(-0600)] <avtar> you may experience some connectivity issues with the fluid and idrc sites for the next 10-15 mins

[16:29:59 CST(-0600)] <avtar> Bosmon:

[16:30:05 CST(-0600)] <avtar> could you try jira again?

[16:30:11 CST(-0600)] <Bosmon> Sure

[16:30:42 CST(-0600)] <Bosmon> Got a bad first request, but the rest were ok

[16:30:49 CST(-0600)] <Bosmon> Was this some kind of reverse DNS problem?

[16:32:32 CST(-0600)] <Bosmon> I'm mostly getting time_starttransfer at 0.2

[16:32:40 CST(-0600)] <Bosmon> But I got one 0.4 and one 0.8

[16:32:43 CST(-0600)] <Bosmon> Nothing dreadful yet

[16:45:30 CST(-0600)] <avtar> Bosmon: sorry, i messaged you and then walked away from my desk

[16:45:38 CST(-0600)] <avtar> so disrespectful

[16:45:40 CST(-0600)] <avtar> :/

[16:45:53 CST(-0600)] <Bosmon> I do it all the time

[16:45:59 CST(-0600)] <Bosmon> Especially if I'm having a bath

[16:49:37 CST(-0600)] <avtar> Bosmon: bind on this vm is performing considerably better

[16:50:16 CST(-0600)] <Bosmon> That's cool

[16:50:21 CST(-0600)] <Bosmon> Any idea what the problem was?

[16:52:35 CST(-0600)] <avtar> the last vm was experiencing i/o locking between bind receiving requests and rsyslog shipping its logs

[16:53:44 CST(-0600)] <avtar> btw

[16:54:04 CST(-0600)] <avtar> that pingdom page defaults to testing from a location in the netherlands

[16:54:23 CST(-0600)] <avtar> i normally choose the nyc option

[16:55:26 CST(-0600)] <avtar> there isn't much we can do about latency across the pond with having everything hosted here in toronto

[17:08:23 CST(-0600)] <Bosmon> Thanks, avtar

[17:09:33 CST(-0600)] <Bosmon> Somehow choosing the NYC option is ignored : P

[17:13:32 CST(-0600)] <avtar> weird...

[18:06:18 CST(-0600)] <thealphanerd> Bosmon: eyo

[18:35:36 CST(-0600)] <Bosmon> THER ALLPHER.... NERDD!

[18:36:15 CST(-0600)] <thealphanerd> hows it going?

[18:36:23 CST(-0600)] <thealphanerd> http://www.smbc-comics.com/?id=2874

[18:37:02 CST(-0600)] <Bosmon> It's not bad....

[18:37:14 CST(-0600)] <Bosmon> Today I have implemented..... THER GLOBBALLY GINGGER... *WORLDD*!