« first day (155 days earlier)      last day (3926 days later) » 

00:26
@voretaq7 case in point about Tuesdays: yesterday I accomplished virtually nothing. Today I sped up the code I've been working on for the last month by 6-7x.
00:58
my github pushes suggest wednesdays are my most productive day of the week
though that probably correlates with Wed being the day I meet with my Ph.D. advisor and need to have something to show
@casey my weekly status meeting is on Tuesday, which means I pretty regularly accomplish something significant on Mondays.
01:19
@BretCopeland Ugh. Tell me that's a stand-up meeting limited to 15 minutes.
<- NOT a fan of weekly status meetings. We have a monthly 30 minute briefing and everything else is called as needed
@voretaq7 they vary depending on how much we have to talk about. It's generally 20-45 minutes.
luckily the ad server team is its own meeting now. We used to be part of the Careers team meeting, and the meetings were starting to become over an hour regularly.
yeah, any meeting that lasts more than an hour needs to stop happening
Frankly IMHO any meeting that takes more than 30 minutes is pushing the boundary if "Wasting more time than it's worth"
We don't normally have anything longer than an hour, and we try not to have many in general
At $job[-1] I had 3 standing meetings on Monday, Tuesday, and Friday. I was essentially completely unproductive because all my time was spent preparing for hour-plus meetings
The last two weeks we had a lot to talk about, including hardware we're about to purchase for ad server since we're getting very close to being live.
01:25
@BretCopeland (formal) meetings I call have a clock (literally - I set a countdown in my phone for 1 hour - if the alarm goes off the meeting ends)
I also try to make sure I have something I can't move scheduled immediately after meetings (like a call with an outside vendor)
@voretaq7 sometimes that's why our meetings end too.
(yes, I strategize about avoiding being stuck in interminable bikeshed circle-jerk meetings. NEVER. AGAIN.)
Usually when David is in the meeting, he has something afterward (typically another meeting or an interview)
@BretCopeland It's seriously the best strategy! Nobody argues when a tier 1 vendor is expecting your call :-)
I'm kind of excited about getting hardware.
01:29
@BretCopeland Don't be. It'll fail. Hardware always fails.
(especially when it's all plugged into a power squid sharing an outlet with the toaster oven....)
@voretaq7 that's okay. We're probably getting three boxes for the NY DC. Just one should be able to take the full production load, so as long as they don't all three go down, we should be alright. If the DC powers down, then it probably doesn't matter that we can't serve ads at that moment.
What is nice about it is that they'll be dedicated just to my team. That's pretty rare in our infrastructure.
Because it's a node.js project, running it on our .NET web tier doesn't really make sense.
I am kinda surprised they didn't just carve you out VMs
@voretaq7 we've been running on VM's for testing in production, but with the number of cores we need, and the amount of network traffic that's going to go through it, it just makes sense to get physical machines.
01:36
@BretCopeland "Get off my hypervisor you greedy bastards!"
Consider that every single page load on our big sites is going to send a request to the ad server.
possibly several I assume - you guys serving the image data too or is that coming off Cloudflare?
@voretaq7 we're definitely not going to serve images. That would be a ridiculous waste of resources.
what's a little bandwidth between... hey how come the router is smoking?
 
9 hours later…
10:18
0
Q: overlapping tags (e.g. ballooning, lighter-than-air)

RedGrittyBrickThere seem to be several sets of overlapping tags - are these useful? For example lighter than air ballooning but no airship or zepppelin tags unpowered glider hang-glider so we get questions tagged unpowered ballooning lighter-than-air. is this desirable?

10:46
posted on May 22, 2014 by Bryan Swopes

22 May 1937: After her Lockheed Electra 10E, NR16020, was repaired at Tucson, Arizona after its left engine, a Pratt and Whitney R-1430-S3H1 air-cooled 9-cylinder radial, caught fire while restarting after a fuel stop, Amelia Earhart and her Navigator, Fred Noonan, arrive at New Orleans, Louisiana. Although she is on the third leg of her […] The post 22 May 1937 appeared first on This Da

 
1 hour later…
11:49
posted on May 22, 2014 by Bryan Swopes

22 May 1948: Jackie Cochran flew her green North American Aviation P-51B-15-NA Mustang, USAAF serial number 43-24760, civil registration NX23888, over a 2,000 kilometer (1,242.743 miles) closed circuit at Palm Springs, California. Her Mustang averaged 720.13 kilometers per hour (447.468 miles per hour) setting a new Fédération Aéronautique Internationale (FAI) World Speed Record for its class,

 
4 hours later…
15:55
Anybody here willing to make a "Wing Design and Aerodynamics for Dummies" recommendation for me? I'm looking for a book to learn the subject. Please keep in mind, I majored in religious history in college and the farthest I got in math was Trig.... I'm not bad at math, I just have no background in it.
I have some text books on fluid dynamics. Can't say I recommend them though. They're pretty much as dry as they come :)
I have a number of books on fluid dynamics
BTW, how much of aerodynamics you should know to fly a plane?
"Fluid Mechanics" Kundu, Cohen. "An Introduction to Fluid Dynamics" Batchelor, "Geophysical Fluid Dynamics" Pedlosky, "Atmospheric and Ocean Fluid Dynamics" Villis, "Turbulent Flows" Pope, "Turbulence in the Atmosphere" Wyngaard , "Dynamic Meteorology" Holton, "Numerical Methods for Fluid Dynamics" Duran. And those are just the ones I can see from here
oh, how could I forget "Vectors, Tensors and the Basic Equations of Fluid Mechanics", Aris.
@Farhan I think that depends on what kind of flying you'll be doing
a test pilot should probably be very knowledgeable on aerodynamics, wheras the guy putting around in a 172 probably just needs to know some basics.
@casey At present, just getting towards PPL.
16:05
Just the basics. flow over the wings == lift, lift varies with AoA, at some AoA lift starts diminishing (stall), etc
@casey and what flow separation is
@casey Mine are in a box in my parents attic, couldn't tell you what they're called much less who wrote them :)
@JayCarr "Fundamentals of Aerodynamics" Anderson, John D. This is what I used in school.
17:11
@voretaq7 I popped over to the Comms Room to see if anyone was excited about the bump in traffic SF is getting after the new Panda release. Nope, just bitterness.
@BretCopeland Panda? <confused>
We got a new panda for the national zoo?
@Farhan Pull stick = go up. Pull stick more = Go down. Pull stick even more = Air France 447.
@voretaq7 To be fair, he didn't pull even more. He just held it there for a very long time.
@falstro There's also not much of an "even more" on the airbus since it's basically a video game with great graphics and motion :)
(I wonder if the HUD says "INSERT COIN" when waiting for an engine start?)
17:27
@voretaq7 HAH! That'd be an awesome easter egg if you could make it do that somehow.
17:40
@falstro Probably he forgot some aerodynamics rule.
@BretCopeland I noticed that whenever you make an entry to this room, you only start talking to @voretaq7. Are you two ... ummm ... related or ... something like that?
@voretaq7 Panda as in Google new search ranking algorithm.
@Farhan well, he's the only one in this room I know in-person.
@BretCopeland Ah, makes sense. A personal relation. ;)
also, I didn't have anything to add to the discussion that was already going on (and had seem to have come to a close)
18:15
@casey Are any of those books going to make sense to someone who only understands Trig? Or do I need to find a calculus primer?
@JayCarr probably not. Most of those books are considered "grad" texts and you'll find them full of tensors and partial differential equations. You could probably follow some of the concepts and some of the derivations but the basic stuff isn't always explicit. Holton for example is notorious for doing things like "F=ma (1). From equation (1) it can be shown that du/dt = -1/\rho \nabla p -2\Omega \cross u +\nu\nabla^2 u + g (2)"
Note to self, avoid Holton.
I guess I just need to accept that unless I'm willing to go to grad school, my understanding is just going to be limited.
if you care to see what that equation actually looks like rendered properly, see my A here: earthscience.stackexchange.com/questions/541/…
10
A: Where does wind come from?

caseyWind is caused by pressure differences. Think of a balloon full of air; poke a hole in it and the air comes out. Why? Because the pressure in the balloon is higher than outside, and so to regain equal pressure, mass moves and that is the wind. There is a bit more to this in the atmosphere as ...

Well that does look cool, I'll give you that.
@JayCarr you don't need to go to grad school to learn the math (and grad school isn't going to hold your hand there either)
you just need to read, practice and probably find a mentor / tutor for concept you have trouble with.
If you get through vector calculus and differential equations, you have most of what you need.
and for those you can just go grab a used textbook and work through it
You can find editions one or two revisions behind whatever is current for dirt cheap usually
18:28
@casey Well, any suggestions for learning that then?
I got my diffeq book for $5 one edition behind current which sold for $100 new
wow
How about MITs online classes?
Are those a good idea?
The free ones.
It took some trips to the library to check problem sets to make sure they hadn't changed, but it was 99% the same
anything free is nice
Anything free and accurate is nice.
But maybe I come from too much of a history background.
I'm always weary of things that aren't cross verified 8 ways from Sunday...
@JayCarr one of the grad students in my meteo program had a BA in History before enrolling for a MS in meteorology
18:31
How did he make the transition? I'd be curious to talk to that guy...
I don't recall his story offhand
I have a similar example, but not quite the same
I took about a 10 year break during my BS (in which I programmed, instructed and flew for an airline)
I had integral calculus in 1999 in High school
and the next real math course I took was vector calculus in 2009
followed by diffeq the next semester
before I took the vector cal class I grabbed my old calculus book and worked through it cover to cover, filled up 2 notebooks with notes and working through some of the problems
did it over 2 months on my days off
@BretCopeland ...why would they call it Panda?!
@voretaq7 because naming things is one of the two hardest problems in computer science.
(and I'm reserving judgment until I see if any decent questions result from it - we get so many "how do I windowze?" questions :-( )
@BretCopeland that's why everything gets a number :P
@BretCopeland thats why my fortran library for ingesting model data from CM1 is called "ingest_cm1" :)
18:38
"Search Algorithm 2.0" :)
@casey All of my Fortran code is called why_do_you_hate_me
@voretaq7 :)
F2008 is where its at, though I don't think any compiler fully implements it yet :(
gcc 4.9.0 just finished (i think, it might not have) implementing F2003
@casey Fortran 2008 is like COBOL 2002 to me. "Hey - we updated your boat anchor! Now with 20% more chain!!"
For the record through, I hate F77 FORTRAN
For the record, I hate Cobol97 - though it was quite profitable for me back in the day.
@voretaq7 I'll agree with that, mostly because polymorphic variables and certain allocatable / pointer variables when used in certain combinations expose bugs in many compilers. but its getting better!
e.g. my ingest_cm1 library does very weird and wrong things in gcc 4.7 and probably wont compile at all before 4.7, but it is right in 4.8
s/gcc/gfortran
18:42
"It's not the compiler's fault you suck at pointers!" - Me (to an undergraduate class that couldn't quite grasp the concept of pointers)
@casey It is however your fault that you're using a GNU compiler :)
pretty sure it doesn't matter what language you write in, if there is a spec, then nothing fully implements it.
fortran pointers are nice because you almost never need to use them unless you are doing C interop. allocatable is nice :)
one possible exception may be C#
@voretaq7 I have ifort 14.0 as well, and I think the same code needs at least 13.1 or something to work right
@BretCopeland Disproven by counter-example: Scheme implements its spec :)
18:44
on the department machines I have pgi and some other compiler as well, but I don't use those
@voretaq7 yeah, but the spec just reads something like (((((((((((((((((((((((((((((put code here)))))))))))))))))))))))))))))
@BretCopeland it's lisp?
I hope those parens are balanced
@BretCopeland no, it's p(u(t((c(o(d(e((h(e(r(e(x))))))))))))).x :-)
@ratchetfreak very related
18:46
@casey that's what the % key is for :)
:)
@casey they are, actually... because I copied the opening parens into an editor, then used replace mode to type over them... yep.
@BretCopeland :effort:
@BretCopeland you don't have an auto-closing editor?
or a highlighting editor
M-x electric-pair-mode
(i went there)
18:49
@ratchetfreak I don't like auto-closing actually, it could highlight paren pairs, but it was just as easy to do it the way I did it. It's not like that's actual code.
And besides, you'd have to pay me at 2-3x what I currently make to write scheme for a living
so you'll do it
@BretCopeland you have to learn COBOL to demand those kind of rates!
I would say 3x for scheme or COBOL, 2.5x for Perl, 2x for ruby
1.5 for VB
@BretCopeland I'd take a university job teaching Scheme and Lambda Calculus for 1.5x what I'm making now.
18:54
I seem to have stumbled into the CS room again... ;)
@casey M-x viper-mode
(is betters now!)
Anyone here have an opinion on the importance of learning functional programming? We've been debating this a bit at work
For us that would mean clojure, btw.
Unicorn shirts just showed up at our office for no reason.
@voretaq7 I'll have to check it out. I'm much faster at editing with vim, but I like a lot of the features of emacs...
18:58
@BretCopeland Stack Exchange: Because UNICORNS.
@JayCarr "Is there a specific problem it solves better than your existing tools?"
@JayCarr the pinnacle of academic functional programming is Haskell
@voretaq7 Probably not. I guess that means we could learn Haskell if we wanted, I think we just preferred the idea of clojure, just incase we found some use for functional programming.
This is something we are discussing doing in our free time.
You know, to be mind expanding.
@casey Haskell is useful in certain corner-cases
@BretCopeland If you guys don't want those, I'll take one ;)
So, is it useful to learn some functional programming techniques? Will it "expand my mind" in a useful way?
JavaScript gives me all the functional programming I want, without limiting me to it.
19:03
@voretaq7 I figured it would be. I've never written anything in it though.
@BretCopeland Javascript gives me nothing I want, unless someone else is writing it in which case I don't care about the language :)
I stay far away from js
and php
and use the absolute bare minimum needed to get the basic interactivity out of the page I'm coding
@casey well, if you're currently doing work in Fortran, I wouldn't imagine JS would be a good substitute.
2
However, it's a really good general-purpose language.
@BretCopeland I had someone try to convince me it would be once. They didn't know a lot about fortran except that it was old so must be inferior.
I use python for my general-purpose stuff but mostly because of the ease of visualization I get with it
since my general-purpose is mostly "turn this data into a movie or a really cool (publication quality) plot"
@casey . . . is this person one of those people who tries to take my HP calculator away because the TI is "so much better"?
19:08
@voretaq7 probably. no one wants to borrow my HP a second time.
Calculator Requirement: Give the correct answer to 2^128. For our purposes "E" or "OVERFLOW" are not "correct".
<3 RPN
@casey My graphing calculator is usually in "equation mode" unless I've been using it to do math
("do math" as opposed to "Fuck, simplify this mess for me will ya?")
@voretaq7 I'm usually doing math with mine though lately I've been using an ipython notebook for that
5 years of calculus, 2 years of trig, and I-don't-know-how-many years of algebra. What do I do when I need to simplify a formula? Feed it to the HP CAS and let it condense everything for me. That's Progress :-)
19:11
@BretCopeland here's a guy trying to use js where fortran would be appropriate
0
Q: Node.js and MPI (Message Passing Interface)

ChewOnThis_TridentAre there any javascript libraries for MPI? I am new to MPI and am learning about intensive parallel processing. We have a supercomputer at school that we are using that only uses MPI and I was wondering if I could code my programs using Node.js instead of c++. Thanks.

posted on May 22, 2014 by Bryan Swopes

22 May 1968: Los Angeles Airways Flight 841, a Sikorsky S-61L, N303Y, was enroute from Disneyland, Anaheim, California, to Los Angeles international Airport (LAX). Captain John E. Dupies and First Officer Terry R. Herrington were in the cockpit, while Flight Attendant Donald P. Bergman was in the passenger cabin with twenty passengers. The flight was […] The post 22 May 1968 appeared fir

posted on May 22, 2014 by Bryan Swopes

22 May 1969, 21:30:43 UTC: Just over 100 hours after launch from Kennedy Space Center, Snoopy, the Lunar Module for the Apollo 10 mission came within 47,400 feet (14,447.5 meters) of the Lunar surface during a full dress rehearsal for the upcoming Apollo 11 landing. Mission Commander Thomas P. Stafford and Lunar Module Pilot Eugene A. Cernan […] The post 22 May 1969, 21:30:43 UTC, T + 10

posted on May 22, 2014 by Bryan Swopes

22 May 1991: After nearly 30 years in service with West Germany, the F-104 Starfighter made its last flight before retirement. The Luftwaffe was the largest single operator of the Lockheed F-104 with nearly 35% of the total worldwide production in West German service. 915 F-104F two-place trainers and F-104G fighter-bombers were built, with most […] The post 22 May 1991 appeared first on

Ew. MPI. That's still around?
Is it still terrible?
@voretaq7 yes and yes
I use a lot of MPI code, but I only write OpenMP which can use lots of cores but only the local node
and is much easier to deal with
since I often just needs loops parallelized
I worked with MPI a little bit in college. I remember disliking it intensely and wondering why it offered no "normal" parallel programming primitives.
I also recall saying "I could just put work units into AMQ. It would be cheaper and easier!" and getting a dirty look from the professor.
F2008 has coarrays which natively handle multi-node stuff, but it is implemented as MPI behind the scenes by the compilers that support it (ifort)
19:15
chat's moving to Oregon...
@BretCopeland if you ever talk to Joel, tell him I appreciated his comments about my earthsicience.SE post in the last podcast :)
just listened to that
I don't think I've even listened to it.
but if I remember, I'll say something to him.
@BretCopeland Nobody wants to live in Oregon :(
I had it on as background noise but it caught my attention when he started talking about ES
@voretaq7 StackExchange: Portlandia
where you can go to your datacenter and ask if all of the components are "local"
19:38
@casey ...is that where these artisanal hand-crafted backup scripts people keep asking about on Server Fault come from? :)
19:52
@casey I actually have an HP 49g+
Everyone should use RPN :)
Also, PostScript.

« first day (155 days earlier)      last day (3926 days later) »