The Whiteboard

user55340

02:25

@GlenH7 for that NP question...

user55340

0

Q: Why we can't have FPTAS for strong NP complete problems

I understood that we can apply FPTAS to the weak NP problems like 0-1 knapsack. But why we cant apply the same principle to the strong NP problems like bin packing? I also checked wiki page about the same but understood very less.

user55340

While algorithms are on topic, I just don't believe that the asker will get the answer he wants from P.SE compared to CS.SE.

user55340

Its possible that we could answer it... though it seams to be wanting a bit more on the theory than practical side (thats the rather fuzzy line I draw).

user41796

@MichaelT My primary point was that it's not off-topic. But I agree that CS or TCS would provide a better answer. Prog is simply too pragmatic to generate an answer based heavily in theory for that one. And I was a pit peeved at Killian's comment as the OP will just cross-post instead of requesting migration.

user55340

6

Q: Why are all problems in FPTAS also in FPT?

According to the Wikipedia article on polynomial-time approximation schemes: All problems in FPTAS are fixed-parameter tractable. This result surprises me - these classes seem to be totally different from one another. FPTAS characterizes problems by how easy they are to approximate, while ...

complexity-theory reference-request approximation parametrized-complexity

user55340

02:35

And, of course, as a corollary you get "The NP optimization problems that are $W[1]$-hard under the uniform reduction have no fully polynomial-time approximation scheme unless $W[1] = FPT$." — Pål GD Aug 8 at 22:15

user55340

I suspect that his question would get duped to that one.

psr

03:20

@JimmyHoffa With the hoops you have to jump through to be allowed to do academic research, someone good enough to actually do such research could get a real industry job to pay the bills and just do the research. This would be a ton of work - which just means it's the academia experience with a bigger salary. Unless you need expensive equipment the only drawback is that academics tend to ignore even qualified non-academics - which isn't really the greatest selling point for academia.

It sort of saddens me that @jozefg wants to go the academic route.

psr

03:36

@MichaelT Just recently checked in code using the Interpreter pattern. I did it to give the MUMPS developers who maintain it some hope (vs. writing a parser generator and letting them figure that out). Let me tell you - writing one of those is even more tedious then it sounds. So much harder than using a parser generator.

Jimmy Hoffa

04:47

@psr Ah c'mon, he'll love it. He's 17 and already loves thinking about category theory, he'll spend all his time as an academic writing papers proving theorems to integrate intuitionistic type theory and linear logic or whatever other obscenely abstract thing he can get to spend his time doing which is what he does with his free time already.

Matt D

when i was 17

i rode bikes, ate dirt, and played guitar badly

Jimmy Hoffa

@MattD when I was 17 I worked 40 hours a week for a tiny website development 'venture' in the basement of a mexican imports shop

Matt D

making porn sites?

Jimmy Hoffa

@MattD Insurance policy filing site for spanish community insurance company, local advertising supported CMS site, one web-store site for some local car parts dealer to manage their inventory and allow them to sell their inventory on-line, web-store for a company that sold these doodads used for making elevated roofs, employee time-tracking website for some local company, online bail-bonds management system, stuff like that

Matt D

online bail bonds

lol

so american

Jimmy Hoffa

04:56

Somebody came to us and wanted us to put it together to sell to the local gov to win a bid for getting off of pen and paper for all the governments bailbonding stuff, I don't even remember the details of the system anymore, was one of the larger ones I did

Not like a corporate thing, it was supposed to be for gov

gnat

09:17

Carthago delenda est...

now that the feature request and implementation notes have been updated to reflect your clarifications, does it make sense to also update decline justification? (BTW your note #4, on denormalized field was spot on - well done!) I can open a separate MSO discussion for that if it feels too big to discuss in comments... — gnat 1 min ago

...Regarding analysis answer, it's probably worth noting that, although somewhat outdated, it still feels in the ballpark. For first 3 hours, the example question had up to 12 answer, making calculation pretty close to those that would be there with 10 posts limit. And even after that, number of answers has been up to 16 which is, again, not too far from 10. Denominator, denoting... — gnat 1 min ago

...age-drag effect should go away (ie should stay at 1 through all the time), but again, it's not that far from it now (between 0.8 and 1.3) and, which is most important, it doesn't really impact the core point made about "relative effect of QScore * NumAnswers / 5 versus SumAnswers". Thing is, this relative effect simply doesn't depend on the age drag: both values are scaled so that drag effect, whatever it is or isn't, simply disappears when one divides them to get relative effect — gnat 1 min ago

Yannis Rizos

10:51

> This was only a test, designed to make sure you were paying attention. You passed. Closing is appropriate for questions with serious problems.

gaussblurinc

11:46

Does anybody work with iOS?

user41796

13:38

@psr Remember that he hasn't been involved in any of the drama of academia yet. Even most undergrads aren't exposed all that much to the challenges. It's only when you hit grad school and start thinking about it as a career option that you start seeing how twisted it is. Easily half of the friends I had when I was on a PhD track went to industry instead of staying with academia. Slogging through to get a graduate degree really exposes academia's underbelly.

user20683

@GlenH7 I spent enough time as an undergrad to see that first hand. Competition isn't quite strong enough a word for what PhDs go through.

user41796

@WorldEngineer I tried not to paint that as a universal picture as it really depends upon how involved the undergrad is. I had friends that were clueless / didn't care about the politics of academia and I had friends who were considering it so they got glimpses

user41796

I went from a top notch school that kept the politics firmly in check to a ... less-than-top-notch school were the politics were rampant. It was quite disturbing and firmly convinced me that I would be better off in academia.

user20683

@GlenH7 you mean industry?

user41796

@WorldEngineer yeah... this is what I get for multi-tasking... Yes, I knew I would be much better off in industry. Politics are still present, but nowhere near as vicious as what I saw in academia. Way easier to move around too if need be.

user41796

13:51

and truth be told, I think I've played with cooler technology since I've been in industry than I would have in academia. Neural research is pretty sweet, don't get me wrong. But it's really, really, really hard to secure funding to throw together a monstrous computing environment. OTOH, when I look back at all of the storage technologies and computing platforms I have been able to work on, it's pretty clear that industry provided me more toys to play with.

user20683

@GlenH7 and the toys are cutting edge.

user41796

@WorldEngineer indeed they are. I was on a project where we were able to influence a portion of the SCSI-3 standard. Yes, standard. We were able to demonstrate the need for a change in the proposal based upon the work we had already completed. And we couldn't envision how anyone else would be able to work with things as it had been proposed.

user55340

14:12

@gaussblurinc How "work with" are you asking?

user55340

@psr I'm just thinking of the one that I saw... ~30k lines of

if (token.equalsIgnoreCase("ifSomeCondition")) { ... } else if (token.equalsIgnoreCase("ifAnotherCondition")} { ... } else if ...

user55340

Yes, thats right... every conditional in the 'language' was explicitly hard coded as a distinct conditional. And was done with multiple conditionals on a line.

user41796

@MichaelT That's ..... inspiring.

user55340

The data structure it wrote was an ArrayList of Objects (thats right, they just put everything in an object...) StringBuffers, images, special barcode objects.... then elsewhere was

if(array.get(i) instanceof StringBuffer) { printer.send(((StringBuffer)array.get(i)).toString()); } else if (array.get(i) instanceof Barcode) { ... } ...

user55340

As I said, I've only seen the interpreter pattern implemented once, and it was just wrong.

user55340

14:20

I suspect it started out as a quick and dirty (and small) thing that had about 1k lines or so for the various receipts that could be printed. But as the system grew, the number of tokens grew without refactoring it into a "its a language, parse it"

user41796

Oh, the dangers of the quick & dirty hack....

user55340

(my previous employer might have been able to keep me a bit longer if they had gone for the "let me just rewrite it" - As the application nolonger needed the interpreter rip out all that code, write a transaction interface for the transaction object (a big ugly thing) to decouple it from the tight innards of the register...

user55340

and then make it so that it returns a receipt object, that can then be sent to a rendering engine (be it the printer itself, or a pdf, or what have you).

user41796

I thought we wanted to advocate for tight coupling because, you know, it's tight

user55340

If done correctly, it would have meant that any object that implements the transaction could get back a standard receipt object that it could print - be it the register and the receipt printer, the website to pdf, or the receipt lookup kiosk, or QA testing engine.

user41796

14:25

I'm gonna have to work on my trolling for the day though. That one didn't quite land the way it should have.

user41796

And think of how much easier it would be to support the next POS system and its variants on receipt printing

user55340

@GlenH7 Yea... the receipt code looked into the transaction, and the type of transaction (void sale? return? etc...)

user55340

@GlenH7 The thing is there are currently the register and two variants out there now (the website and the receipt lookup kiosk).

user41796

blech. I know I'll get flamed for this, but RPG-IV was kind of a standout language from that point of view. It helped decouple the application's printing requirements from the actual hardware. Kind of elegant in its own way.

user55340

The website has its own receipt generation code (that has some glitches with the ones printed in the store) and the receipt lookup kiosk has a jsp that it prints.

user41796

14:27

@MichaelT I would imagine it only added to the vendor lock-in with the POS system too

user55340

@GlenH7 Kind of... though not too much. Previous employer bought a source license to it... good thing too because the innards of that code were just wrong.

user55340

Within a few days of the "take 3, all stores by this time next year" project, we had completely made it impossible to ever get a patch or support from the vendor.

user55340

And yea, looking at the code - not sure we'd want to.

user55340

(meeting time)

user41796

@MichaelT I never had a chance to find out about the innards of my second previous employer's POS system, but I was always curious. If you can say, I would be curious to know which POS system you were able to poke inside of.

user55340

14:40

@GlenH7 Yea... that one. I can spot it on a live system now too...

user41796

It's kinda hilarious when you can spot issues from previous work lives "in the field"

user55340

We took it so far from what it was designed for... it just wasn't what should have been done. It was likely good for a mom and pop shop... though I'd still be worried with the holes that we saw in it.

user41796

One reason I stayed away from previous employers POS system is that I saw it everywhere and I knew that I wouldn't be able to resist trying to fix field issues.

user55340

It wasn't designed for 16x registers at 300 locations... where you start worrying about failover (wasn't in the core product), or integrating with a home grown back office system.

user41796

WAT? You want live inventory updates too? Sheesh you're demanding....

user55340

14:43

The register was purchased from another company that wrote Java 1.4 code for it. You can see bits of it in there. Then new team that worked on it added 3 modules... and one of those modules got interwoven into the entire system.

user55340

That module necessitated circular dependancies... which in turn cased a bunch of reflection code to get added.

user41796

Reflection makes everything performant, right?

user41796

(don't spit your coffee out)

user55340

We didin't use that module at all... and every time we saw the code in a method we were working on, we'd try to extract it.

user55340

(ie: rip it out)

user20683

14:46

Either of you have any impressions on Scala?

user55340

@WorldEngineer Its a reasonable and reasonably popular jvm language...

user55340

One of the developers undertook a "redo the layout" project that was necessary to get at the steep learning curve of the project.

user20683

@MichaelT What I've read of the main site for it seems comprehensible so far...That's probably the Haskell talking.

user55340

Trying to figure out where something was / should be was a significant chunk of trying to fix things. And then you'd find things in the wrong place that was done because it was 'expedient' back sometime, and resulted in more reflection code.

Jimmy Hoffa

@gnat Carthago canem placito

user55340

14:50

(they stuck everything in untyped arrays or collections... and then when the class wasn't visible at that level at compile time, thats when you'd reflect it)

user20683

@MichaelT yick

user55340

@WorldEngineer The real fun was reflection based versions of if (foo instanceof SomeClass) became "get the string form of the name, and then test that" because SomeClass wasn't visible either.

Jimmy Hoffa

15:08

@MichaelT Fuck all that, get a pointer to the object, inspect it's v-table for the type id, then manually parse the bytecode on disk to find out the key information you want to know. Private shmivate!

user55340

@YannisRizos question on the collider from workplace made me think of you... (not in a bad way)

user55340

26

Q: Blacklisting resumé liars

I have been conducting interviews the last couple weeks and in some cases I had to deal with resumé liars. Most cases are just people who list in their resumés skills that they don't actually have. I don't mean having them in a begginner level, I really mean not being able to demonstrate even th...

interviewing human-resources ethics

user55340

15:45

I have this weird urge to filk "Closing Time" into "Meeting Time"

user55340

o/` Every new project comes from some other project's end o/`

Jimmy Hoffa

in short, no. Ruby doesn't have a quality threading model to support concurrent in-process work and communication. If you want a trendy hipster language (or just for this general type of domain of multi-user X servers where X is anything with stateful connections like game) Node.JS would serve you much better. — Jimmy Hoffa 25 secs ago

user20683

16:16

@JimmyHoffa If you must use Ruby, you could do JRuby + Akka

Jimmy Hoffa

@WorldEngineer Yeah, but it's just the wrong tool for the job scenario. I don't genuinely have anything against ruby (though I don't like the idiomatic humane interface they sell, the language other than that seems fine), but as soon as you start needing sessionful multi-user activity (or generally anything heavily reliant on concurrency), it's just not the right tool anymore. Some ruby runtimes do better but still at it's core that concurrency is far from what ruby is for.

Node.JS is explicitly for precisely that however, so he has an alternative that's still flavor-of-the-day (and genuinely a quality implementation), nothing to complain about there

user55340

@JimmyHoffa Awhile back I read an article about a guy working on the Netflix(?) machine learning challenge and trying to do it in ruby, with threading... and having it completely get bogged down.

Jimmy Hoffa

@MichaelT Yeah see the thing is, even when you use one of the ruby alternatives that has a better runtime model, you're still in a language that was fundamentally not created for that purpose so the knobs and dials available for even doing that stuff aren't going to be any level of maturity

user55340

I remeber the guy ending up going back to Java... still trying to find the blog post.

user55340

(it was the type of thing that ruby bogged down at less than 100 threads, so he had to change his model again and again to try to make it work... he switched back to java, and went to 1000s of threads without problem and the model he had in his mind was easy to map with that)

user55340

16:33

@WorldEngineer blam! closed and deleted fast.

user20683

18:39

theonion.com/articles/…

user20683

@GlenH7 @JimmyHoffa @MichaelT

jozefg

@JimmyHoffa Haha I laughed a lot reading that

user55340

Its snowing!

jozefg

@MichaelT Here too, isn't it miserable?

user55340

@jozefg First snow of the year.... the V-ball players are trying to figure out what the difference between 'snow' and 'sand' is. I said its a edit distance of 3.

user55340

18:45

One of the people from the manager side of the building is wondering if we can open an office in Antigua.

Jimmy Hoffa

I want snow too :(

jozefg

@JimmyHoffa but don't you get mountains, I'd take mountains over snow

Jimmy Hoffa

@jozefg Package deal, let's see...

ah no snow in the front range again until early next week

@jozefg This pic was taken ~10 minutes down the street from me

School of mines is in the foothills nestled up next to the left most mountain in that image

jozefg

@JimmyHoffa Wow.. I want mountains..

Jimmy Hoffa

19:00

@jozefg highway 93 is what runs from boulder (in that pic) about 20-30 miles south to Golden (where mines is) and... it's one of the most dangerous roads in the country

user41796

@WorldEngineer - the comment software bundle contained on 35 floppy disks brought back memories of some of my first redhat linux distributions

Jimmy Hoffa

It's not the snow that's the problem, it has a huge valley from mountain up to it on the one side. It's a gorgeous drive, but that valley gets absolutely demolishing winds at times. I had to dodge a christmas tree blowing off someone's roof one time

I once saw a roof topper blow off someone's car and explode on the side of the road (shovels and rakes were in it apparently, not a light thing for the wind to carry away)

user55340

Gah, this makes me nostalgic for SF. sobadsogood.com/2013/07/09/…

Jimmy Hoffa

couldn't be a nicer drive though, this is a pic taken from the highway, this valley between mountains and 93 basically follows it the entire 20-30 mile stretch

user55340

(I was looking for this thing when I found that one - slate.com/blogs/bad_astronomy/2013/08/10/… )

user41796

19:08

Daily close vote limit reached <sigh> Have I ever mentioned that I need more close votes?

3

Jimmy Hoffa

They should really onebox the onion in here..

theonion.com/video/…

user55340

That fog one, I especially like the ~3:30 where the fog 'breaks' over a hill.

Jimmy Hoffa

@MichaelT Doesn't render for me for some reason

It's just a slideshow, a frame comes up every 45 seconds

which is a type of time lapse I guess...

user55340

@JimmyHoffa It would be close to realtime then.

Jimmy Hoffa

@MichaelT yes, real time lapse.

jozefg

19:15

@JimmyHoffa I've almost dualized lenses

user55340

@jozefg Is that some haskell thing you're making up words for again?

Jimmy Hoffa

@jozefg What does that get you? (I still haven't gotten to lenses any further than: they're setters and accessors)

very very general setters and getters

jozefg

@MichaelT Dualizing stuff is an idea from category, basically you flip all the arrows of a constructions and you get free (dual) theorems

Jimmy Hoffa

so the dual would be, constructors and destructors?

jozefg

@JimmyHoffa I think they're lenses on sum types vs product types

So I do believe I can prove that lenses are the categorical dual of prisms

Jimmy Hoffa

19:17

@jozefg lenses work on both though

granted product types are definable as sum types

jozefg

@JimmyHoffa Not necessarily, you can't satisfy the get put law on a maybe lens.

What if it's nothing?

Jimmy Hoffa

@jozefg You'll have to hop on #haskell on Freenode and talk to EKmett about that

jozefg

@JimmyHoffa I think I shall once my proof is less hand wavy

Jimmy Hoffa

@jozefg He's plenty friendly, as well as McCann and plenty of others in there anyway. He'd likely help you out with it.

Have you not stopped in there before?

If they're willing to talk to me about stuff helping me understand shit like that, I'm sure they'd be more than happy to talk to someone who actually knows what they're talking about

@MichaelT Ok that's cool. I had to re-open the tab but yeah, reminds me of the cloud forest in Peru

user55340

@JimmyHoffa I remember just sitting on top of a parking ramp and watching the fog roll in one evening.

jozefg

19:26

@JimmyHoffa yeah I'll have to ask him about this.. It's weird, lenses appear to morphisms in this category

Jimmy Hoffa

@MichaelT About the only equivalent we have here is going up to the foothills before dawn and watching the sun come up and light the entire plateau, denver is a big flat plateau running up to it

(in the picture there you can see I think the south table in Golden, you can hike up there anytime)

can never remember the south table from the north table

But yeah, watching a cloud roll in at ground level is something else altogether

user55340

For the green tech, the federal building in SF is rather interesting (fog plays a role in this)...

user55340

The mesh catches the wind off the ocean at night, and forces it down along the outside of the building. At night, the windows open a little bit. The inside is concrete and the cold outside air cools down the concrete to ambient night temp.

user55340

This way AC during the day can be limited because the concrete is able to absorb the heat.

Jimmy Hoffa

19:41

and it looks like a work of art too

user55340

The mesh is also all south facing, this cuts down the amount of solar heating the interior gets during the day

Jimmy Hoffa

@MichaelT See if they were going to go with something like that though, I have to wonder why they didn't go a little thicker with the bars, make them hollow pipes, and just have some water pumps pushing/pulling brown water through on it's way to the treatment plant. That would cool the building significantly more, and in the climate they don't have to worry about pipes freezing (besides, constantly running it wouldn't freeze)

I remember reading at one point the two most effective house cooling techniques people found were A) house shutters (which was extremely surprising) and B) having water misters around their house (which is a horrible waste of water)

from some study. Unless I remember for shit

user55340

@JimmyHoffa But thats an active system - this uses all passive technologies.

user55340

Another fun bit from it - elevators only step every 3rd floor. Most people have to walk up or down a flight of stairs to get to the cube (encourage interaction between people on the stairways and promote some healthy walking).

Jimmy Hoffa

19:59

@MichaelT Untrue though, there's still an active HVAC system throughout the building, what I refer to would act as an arm of the HVAC significantly reducing the demands on the rest of the HVAC

user55340

@JimmyHoffa Actually... I think they scrapped that.

Jimmy Hoffa

@MichaelT Scrapped what? They have no HVAC?

user55340

"The architect's decision to eliminate the usual HVAC system saved $11 million in construction costs."

Jimmy Hoffa

holy crap

that's awesome

user55340

ever seen/heard about the block of ice cooling system?

user55340

20:03

The idea being you put a block of ice in the basement, run the AC through that. It melts during the day, but at night, you refreeze it when power costs are cheap.

user55340

Jimmy Hoffa

20:15

0

Q: What should be the architecture of an urban game system?

I'm going to develop an urban game using a telco API for phone geolocation and sending/receiving messages. A player would pick up one of the scenarios, move around the city and when he hits a given location, he gets a message and possibly has to answer it. I'm wondering, what approach would be t...

architecture web-applications ruby ruby-on-rails

Turns out he was asking about architectural design not just technology stack platform I guess

pmichna

@JimmyHoffa So, what should be the architecture with node.js?

Jimmy Hoffa

@pmichna Don't ask me! :D

Your question wasn't super clear and a bit broad, likely you won't get an answer on it as it is, but we can definitely discuss it here. If you edit it a bit though you may get a hand full of more comprehensive thought out answers in the question itself

What exactly are you trying to do?

pmichna

@JimmyHoffa A bachelor thesis project ;)

As I said I have access to telco API

Jimmy Hoffa

It's not really clear, I just hear game server and immediately think long-standing sessionful connections with shared memory for concurrent state updates which is a problem you really don't want to use Ruby for

pmichna

Probably you are right

my game server should be able to host concurrent games

and should be stable so that noone gets an error during playing, like you said

So my idea is to create a web interface for users to registers, edit game scenarios, start/stop games, view rankings etc.

Jimmy Hoffa

20:24

Aye, Node.JS works great for that type of thing, Java will work fine too, or Python plausibly though I'm not super familiar on Python's runtime mechanisms

@pmichna and what is a "game" going to be?

pmichna

pinging players, say, each 5 sec to get their location

and send them SMS messages ro do something else depending where they are and they have done before

user20683

@pmichna Python concurrency is...eh

user55340

From the mobil side of the world, location fixes every 5 second is... battery consuming.

pmichna

@WorldEngineer Ok, im rather not interested in python

@MichaelT It doesn't use phone's battery

it's GSM location

not very accurate, but pretty good in cities

user55340

So this isn't a smartphone running an app...

pmichna

20:28

not necessarly :)

could old nokia 3310

could be*

the communication would be over SMS messages

and as I said, there would web interface and after reading a little bit about node.js I think it may be a could candidate for the game server. However, my question is: should the game server and web application (user interface) be separate or one node.js application?

user55340

So you're polling the phone company for the information on the placement of a given phone rapidly... you might just want to double check the API to make sure you don't get ratelimited there.

Steven A. Lowe

Hi Pmichna! Hi MichaelT!

pmichna

You're correct. The rates can be changed for my needs :)

Steven A. Lowe

Mind if I hop into this conversation for a moment?

pmichna

but I may lower the pinging polling frequency

user55340

20:30

@StevenA.Lowe Go ahead...

pmichna

@StevenA.Lowe no problem :)

Steven A. Lowe

Thanks - I think you're asking two different questions at once

what's a good architecture

what's a good language

pmichna

good point

Steven A. Lowe

until you understand/define the architecture, 'what language' is premature

you know the limits of the geoloc API and how it operates, so if you must use that library then that's a hard constraint on the architecture

which may dictate the process/thread structure

Jimmy Hoffa

@pmichna Separate apps make sense for this to me

Steven A. Lowe

20:33

but the protocols and traffic levels between the game app and game server (however that functionality is physically distributed) may also influence the architecture

in terms of queueing, state-management, and so on

i think you've got a lot more architecting to do ;)

pmichna

I get the same feeling :). As of traffic - it won't be big, because it won't be a real life project (at least for now) - just a project for bachelor thesis

Steven A. Lowe

hmmm... thesis requirements may be very different from real-world requirements!

pmichna

The very basic requirement is that it has to work :)

I feel I'm not a big fan of Java and I wanted to learn something new, so I thought about Ruby and RoR

Steven A. Lowe

if it just has to work, then pick a tool you know, use TDD, and just start building it ;)

while learning a new language might be fun, it's not the point of the thesis, is it?

and it will make the job more difficult unnecessarily

pmichna

yeah, you are right

Steven A. Lowe

20:40

on the other hand it may make it more fun, entirely up to you to decide an acceptable level of risk

pmichna

but I'm not a master in any other language/framework

that makes the situation more open for other options

still

damn

wish there was a simple answer :P

Steven A. Lowe

can Ruby do it? probably. are there better choices for a 'real' app? probably. Since this is a thesis project, does it matter? I doubt it.

user20683

Scala

user20683

if you want massive concurrency

Steven A. Lowe

So pick something you'll enjoy learning

user20683

20:41

but that's probably overkill

user55340

Ruby isn't known for good concurrency models.

pmichna

ok, Scala is to hardcore for this project

Ruby is already gone - everyone discourages me from it

user55340

igvita.com/2008/11/13/concurrency-is-a-myth-in-ruby

user55340

@pmichna (there's a good reason for that)

pmichna

node.js seems great because of its ease of use of concurrency

Steven A. Lowe

20:43

I don't think the language choice for a toy project is critical. I'm assuming that the thesis is about either the architecture, or just the experience

what is the point of the thesis? and/or which class is it for?

pmichna

I wrote some code (very little) in Erlang and it seems to have similar idea

Steven A. Lowe

the 'requirements' in this case are very much a function of how the thesis will be graded ;)

pmichna

@StevenA.Lowe The thesis is my bachelar thesis, for the end of undergraduate studies

user55340

Oct 8 at 23:16, by psr

Hello World is tough with Erlang. Due to concurrency you have to say hello to the whole world at once.

pmichna

the thesis subject is "urban game using telco api and crowdsourcing" or something like this

@MichaelT Yeah, I wasn't so much attracted to Erlang, that's why I don't want to use it in my project. :)

Jimmy Hoffa

20:45

@pmichna Erlang is a fantastic language for these things, concurrency is the point of the whole language

pmichna

it's cool, but not know :P

Jimmy Hoffa

but it's a much steeper learning curve than say node.js

pmichna

so maybe coming back to the beginning and putting the detailed architectural talk aside

I know some Java, done some small projects (mainly some lab tasks on tutorials during my studies)

and there is node.js, which I have noe experience at

user41796

@pmichna What Steven has been getting at is you need to scope your project in one way or the other.

pmichna

but it seems fun

user41796

20:47

Either pick a language and architect around that. Or architect the project and pick the language that best fits it.

Jimmy Hoffa

@pmichna A web front end in Ruby would be fine, it's just the part that's doing all the concurrent communication and updating the current serverside game state of each user that you need to not do in Ruby

user41796

Trying to do both will leave you in an infinite loop

Jimmy Hoffa

@GlenH7 sound advice.

(well, to an extent... fundamental requirements dictate some portion of the architecture in an overriding fashion that you need to pick the language based on)

pmichna

@GlenH7 I'm trying to get out of it ;)

user41796

It's just a thesis. I'd recommend picking a language you know moderately well and bang the thing out. Picking a new language adds a lot of risk which translates into stress for you.

user41796

20:49

full disclosure: My masters thesis was essentially rewriting the code from someone else's thesis. And yes, his code was ... poor.

Jimmy Hoffa

If it's just a thesis, who cares. Nobody will ever see or use the code. It could be (probably usually are) rife with concurrency bugs and nobody will know the difference

pmichna

I think I'll just need to think about it

thanks for all the advice

user20683

@pmichna Just use LaTeX and a sufficient number of sigmas and no one will notice.

pmichna

:)

user55340

79

Q: Can we make a love heart with LaTeX?

I understand that this post may cause some down votes because it seems a bit off-topic, but having seen the discussion on Christmas Tree with LaTeX I am still willing to try. My Girlfriend's birthday is coming (OK, actually less than 12 hrs!). I would like to write her a little love note. Both...

tikz-pgf diagrams fun

user55340

20:52

(a question that has been sticking to the collider for awhile)

user41796

@pmichna On a more practical note, I don't believe your question is answerable currently on the main site. Would you consider closing or deleting it please? You're more than welcome to re-open / edit once you've plugged in some of the details

pmichna

@WorldEngineer It would be a nice feeling if everything worked fine, though

user55340

@StevenA.Lowe And thank you for stopping by.

pmichna

@GlenH7 sure

@StevenA.Lowe exactly, thanks for your help

user55340

Ug - old question, has upvotes on answers that are non-trivial... requires four votes to delete.

gnat

21:00

@GlenH7 I got a question regarding your hotness formula analysis. In Raw basis of question table, collider column, score values - look like ones observed from collider, correct? If yes, how do you take into account that these may be slightly out of sync with values you calculate?...

... I mean, value for collider was calculated by the system in a different moment and post scores could have change since. 2-3 votes for question or an extra answer posed can make a noticeable difference between what you calculate and what is shown at collider, right?

user55340

Actually, seeing a couple of those in the 'most delete votes'

user55340

71

Q: Crunch-time Sleep Deprivation: Survival Strategies?

Fore-warning: I'm typing this question on 3 hours' sleep in the last 48, after a month straight of long hours and sacrificed weekends. Launch deadline D-day was 12 hours ago, now we're in red time. What do you do to make it through delivery deadline crunches? Survival strategies? Tips? Caffeine...

sleep deadlines

user55340

Two votes cast, 8 more needed.

user41796

@gnat Noted the assumption and moved on.... Granted, that's hand-waving, but it works in this case. Especially since I wasn't privy to the "actual" algorithm at that point.

gnat

@MichaelT mine is there already. Although I would rather prefer a his lock for it (writing flag messages for h/l is rather tricky, that's why I didn't did it yet on that Q)...

yesterday, by gnat

@MichaelT you know, when I look at old, currently inappropriate questions with solid, good quality answers, I ask myself, why do we need to blindly follow SO stringent requirements for historical lock? "The post is stellar... large number of views, upvotes and inbound links" oy wey. These may be fair for a site that changed smoothly, slowly and gradually, but Programmers are not like that!

user55340

21:04

While its upvoted, it just doesn't have the views to indicate the long tail of a useful h/l.

gnat

@GlenH7 all right that makes sense. As long as you aware of the possible diff, I see no reason to worry :)

user55340

Only 1 linked question, and one related question - so its lacking on the incoming too.

user41796

@gnat The time decay portion wasn't working either (duh, wrong formula) so I had to wave my hands over that one too. Sometimes you just have to acknowledge that some data is crap and do what you can with the rest.

gnat

@MichaelT well, to me h/l at Programmers is not so much about usefulness (I'd leave that to SO where was no scope change like ours) but for civil peace and smoother transition between the past and present (and the future for that matter, I won't bet that current scope is cast in stone, think about workplacey topics)

user55340

There are, admiditaly good answers in there... If I was a mod (no, @WorldEngineer don't go sticking a diamond here) I'd delete most of the answers and then h/l it.

user55340

21:07

As it is, it likely won't get deleted, its closed... and isn't terribly embarrassing. Just more an observation on questions with lots of upvotes and delete votes.

user55340

7

A: Crunch-time Sleep Deprivation: Survival Strategies?

Haylem's Survival Guide to Code-Rushes Master your Sleep Be Active and Healthy Be a Team Sleep! Master your Sleep The most important thing is to master the art of sleeping well. Listen to your biological clock and be aware of your Circadian rythm. Learn to understand your sleep cycles and sl...

user55340

Thats the good answer in there.

gnat

@GlenH7 "revelation" about time decay didn't bother me much, as soon as I figured it only makes the issues with score more prominent. And, of course it was obvious from the very beginning that the core point about relative effect is invariant to time decay...

...age-drag effect should go away (ie should stay at 1 through all the time), but again, it's not that far from it now (between 0.8 and 1.3) and, which is most important, it doesn't really impact the core point made about "relative effect of QScore * NumAnswers / 5 versus SumAnswers". Thing is, this relative effect simply doesn't depend on the age drag: both values are scaled so that drag effect, whatever it is or isn't, simply disappears when one divides them to get relative effect — gnat 12 hours ago

@MichaelT if memory serves I once seen quite compelling explanation (Show? Yannis? maybe ChrisF) why cleanup doesn't make much sense for h/l. Idea was like if you cleanup, it would better stay unlocked. If you lock, heaps of garbage serve as sort of proof why it's not appropriate and on the other hand (which is maybe most important), lock protects things from becoming broken windows - it's intended and designed for that

"Show" = Shog

that one won't go to h/l, for sure. I'll CV it tomorrow, will flag if closing won't come through the regular way, will wait until it's eligible for deletion, and will del-vote as soon as I can...

3

Q: where do you track team Decisions

I have been on many development teams and as the team matures decisions about direction are made. These decisions often come back up over and over. Like why don't we fill in this field why didn't we use memcache over a custom solutions. These decisions add up over time and become a significant...

development-process configuration-management

barbarians at the gate

guys who learned that recommendations don't fly at SO as good as before, now try to leak them through to Programmers

user55340

21:23

Interesting bit on the apple.SE site with SO people realizing that different sites have different accepted questions.

user55340

37

Q: What un(der)-documented features have you stumbled upon in Mavericks?

This question serves to share and collect the enhancements which are not documented by Apple, or documented poorly. Please post one feature per answer. Please also check to see if your answer has already been posted - duplicate answers will be deleted. To search answers for this question use inq...

osx hidden-features mavericks

user55340

Coming from StackOverflow, this would be closed as subjective immediately. Does this community tolerate "favorite feature" type "questions"? It hardly seems to be something answerable. — p.campbell 24 hours ago

user55340

@p.campbell: you can find some discussion of this here and here - the short answer is that hidden feature wikis are tolerated here, under limited conditions - a straight up "favorites" poll isn't something that you should find on any SE site however. I've edited this question to bring it closer inline with what is normally found here. — Shog9 19 hours ago

Steven A. Lowe

@pmichna @MichaelT sorry for the dropout, a client called with a problem. Looks like y'all got it sorted out. Good luck!

user55340

@StevenA.Lowe Yep. And yea, thank you for stopping by. Its always nice seeing people in chat working out things that would be far too awkward for comments.

user55340

21:25

@p.campbell - It's one of the fascinating things about SE. Each board has its own personality, and slightly differing rules. At Money.SE this gets shutdown quickly. — JoeTaxpayer 3 hours ago

gnat

@MichaelT yes it's the distinct feature of their site. When I studied sticky questions, a similar question of their was the only one where lots of answers felt all right. They somehow found a way to manage things like that, in their context...

79

Q: What un(der)-documented features have you stumbled upon in Mountain Lion?

This question serves to share and collect the enhancements which are not documented by Apple, or documented poorly. Please justify your answer; if it is something that is well documented by Apple and elsewhere on the web, it does not belong here. Your answer should put the feature into context, ...

osx hidden-features mountain-lion

user55340

319

Q: Please share your hidden OS X features or tips and tricks

Do you know any hidden or little-known nice feature of Mac OS X? It doesn't matter what it is—maybe just a short terminal command or a keyboard shortcut. Share your experiences on hidden Mac OS X features with us.. Please post one tip per answer. Please also check to see if your answer has alrea...

osx hidden-features

user55340

(615 favs, 127 answers)

gnat

"Ask Different" - name fits :)

their questions differ

user55340

Yep... its up to the community really to decide how open/closed they are to certain types of questions... within the SE framework.

gnat

21:31

on the other hand, they've got objective criteria to judge answers and drop off meh / crap, meaning more leeway for questions

it's not a coincidence I think that Programmers and TWP suffer most from collider, while other sites more or less clean the crap answers - per their objective criteria

user55340

P.SE is partly at the point where we are because of our close association with SO, its history - what we're closing now with questions showing up is in part a reaction to that history.

gnat

@MichaelT maybe. Part of the history is SO oldtimers making comments under crappy questions, "belongs to programmers"

I flagged here one cross-post like that today, along with flagging a comment at SO "origin" that tripped OP into cross-posting

one post two helpful flags at two sites

user55340

21:47

github.com/MrMEEE/bumblebee-Old-and-abbandoned/commit/…

user55340

I need more close votes.

tylerl

22:06

@MichaelT how did I never see this before?

user20683

@MichaelT There's an easy way to fix that...

user20683

:)

Robert Harvey

I am reporting you to jeff atwood — user104971 1 min ago

Good luck with that. — Robert Harvey 1 min ago

Jimmy Hoffa

@RobertHarvey If you didn't catch the sarcasm in my tone about Node.JS being a language according to hipsters; it was just a matter of being out of the edit window when I noticed what I said

Robert Harvey

22:22

@JimmyHoffa I kinda half-thought it was mild sarcasm, until you pointed out that some hipsters really do think of it as a language. I guess you get street cred if your thing is a language, or at least a domain-specific one. Then again, some folks still think Node is single-threaded.

I actually checked. You jest, but some bloggers really do talk about Node like it's a language.

Jimmy Hoffa

@RobertHarvey and even more people think Node is magical eventomatic elf-driven shit and that threads are the evil santa claus that only gives people coal and sends every child to the krampus

@RobertHarvey I didn't have to check, I don't doubt it for a moment. People call JQuery a language for fucks sake (I had a contractor once argue with me that it was when he was asking me to fix something he was doing for our company)

it's one thing to be misinformed, it's another entirely to fail at making something work, ask somebody who is said to know what they're doing to help fix it, and then tell them they're wrong regardless of the fact that they solved your problem when you couldn't

Robert Harvey

Unfortunately, I too have been guilty of talking out of my ass once in awhile. A couple of days ago, someone asked this question, and I answered with a perfectly sensible usage of type classes in Scala. I was getting upvotes on my answer, too, until someone correctly pointed out that you could accomplish the same thing with ordinary generics and interfaces.

So it turned into "stump the chump."

tylerl

> Hello kind sirs. I'm looking at building a ______ program. How do I do this? Please send the codes.

- Half the questions I read

Jimmy Hoffa

@RobertHarvey With Jeff and Joel the Stackert Brothers? We're all guilty of it sometimes. Something makes sense so you throw it out there, other people think it makes sense too, then someone corrects you. Can't know what you don't know is all.

> Due haste, my program does not work, plz2fixthx!

- The other half

Robert Harvey

@JimmyHoffa Yes, that.

Most questions are simply underspecified.

I suddenly realized that SO made a tactical error a couple of years ago with "no question is too simple." The fact is, you need basic knowledge of your craft (and some decent communication skills) to formulate a good question.

Jimmy Hoffa

22:31

@RobertHarvey It's really hard to specify programming problems when you don't know what pro-gra-mee-nguh is

Robert Harvey

Exactly.

tylerl

@RobertHarvey puts it generously

I want to...

user20683

22:47

In other news, I cannot run the JDK for Java 7

user20683

I think it's time to upgrade

Jimmy Hoffa

@WorldEngineer In other news Java has only pretended to run for years.

user20683

@JimmyHoffa fair enough but it's still annoying.

Jimmy Hoffa

@tylerl My favorite is Why did... because we're all intrinsically aware of what caused every major company to make every technical decision it's ever made

gnat

@RobertHarvey did they forgot to put the comma back then? And make it into close reason: "no, question is too simple"

Transcript for