The Whiteboard - 2014-04-22 (page 2 of 2)

« first day (1327 days earlier) ← previous day next day → last day (3662 days later) »

8:07 PM

Hmm, that's interesting. Turns out I've never actually used the 10K queue.

user55340

@RobertHarvey The 10k flags? or the queue history review? or the deletes?

Whatever is not there anymore that I've never seen anyway.

user55340

Used to be that programmers.stackexchange.com/tools had a 'flags' tab - thats the gone part.

user41796

I was about to ask how long you've been a mod, but that question ignores your precious repz standing on Progs....

user55340

But there's also the "lets see what everyone else is doing in reviews"

user55340

8:09 PM

(I'll point out that given some of the bugs that I found in the 10k tools, most people don't explore the pages much beyond the surface, if that)

I'm a mod on Stack Overflow, so I don't need the 10K queue. I use the Close queue and the Delete Votes tool on Programmers.

@MichaelT I took on a project like this once, when I was less... knowledgeable. Turns out its not just a simple SQL statement. :) In fact, it's a surprisingly difficult problem to solve, and some constraint sets have no solution.

It's quite the bummer when you tell a customer that you can do something in 3 days that, in reality, is six months worth of work.

user55340

@RobertHarvey those requirements are a killer...

user55340

though if you knew prolog...

I still think you can do it with some well-chosen SQL, but the final solution is probably recursive.

user55340

Constraint programming

In computer science, constraint programming is a programming paradigm wherein relations between variables are stated in the form of constraints. Constraints differ from the common primitives of imperative programming languages in that they do not specify a step or sequence of steps to execute, but rather the properties of a solution to be found. This makes constraint programming a form of declarative programming. The constraints used in constraint programming are of various kinds: those used in constraint satisfaction problems (e.g. "A or B is true"), those solved by the simplex algorith...

8:11 PM

Needless to say, it never got implemented.

@RobertHarvey ah but recursive solutions are the easiest ones, right? :)

1

Q: Why would we use Recursion if we still can accomplish the task using loop control structure?

As I start learning Recursion different questions are crossing my mind ... Recursion uses more memory for the stack and it usually slower due to maintaining the stack .... What is the benefit of using Recursion if I still can use for loop ? We describe actions to be repeated over and over until c...

@JimmyHoffa CTE's are a pain in the ass. They probably wouldn't scale well for a constraint problem.

user55340

Back in AI class in college, we did a job-shop problem assignment in Lisp.

user55340

@RobertHarvey Given the java tag I'd say "never" given how nasty the stack frames become and the lack of TCO.

user55340

Nothing like "You have an error here and here is 1024 stack frames of recursion in the stack trace"

user55340

8:15 PM

Using eclipse, I only get 1024 recursive calls. — Norswap Mar 25 '13 at 10:57

@MichaelT [ponders writing a prolog-like DSL that's actually understandable]

user55340

Drools is a business rule management system (BRMS) with a forward chaining inference based rules engine, more correctly known as a production rule system, using an enhanced implementation of the Rete algorithm. Drools supports the JSR-94 standard for its business rule engine and enterprise framework for the construction, maintenance, and enforcement of business policies in an organization, application, or service. JBoss Enterprise BRMS JBoss Enterprise BRMS is a business rule management system and reasoning engine for business policy and rules development, access, and change managem...

Ack.

Not a big fan of the BRE's.

user55340

Microsoft has a way of making you lose your appetite for such things..

user55340

8:20 PM

The 'B' part is the application. Its a RE underneath. Prolog is a RE too.

user55340

Its a matter of using it as a rules engine and forgetting about the business aspect of it.

@MichaelT It looks like it's just trying every combination.

Interesting... the final solution is a polygon.

OK... Adding some constraints. It's got some skill.

user55340

The thing is, if I was faced with a nurse schedule problem, I'd use that.

Yeah, that would work.

user55340

Its market as a business rules engine, because that's where the money is (for consulting)...

user55340

8:27 PM

Businesses don't say "we want a rules engine that can solve the TSP" but rather "we want a business rules engine for our rules" and then start searching for BREs.

user55340

Thats why they always skip prolog.

I didn't realize that. Now it's interesting to me.

user55340

Poke at optaplanner.org (the thing behind that video)

user41796

@MichaelT awesome link. That brings me a lot closer to my idea of providing a soccer field scheduler.

user55340

I'm amused they've got nurses as an example for their 'what we do' front page.

8:34 PM

Why don't we use it to schedule a soccer team of nurses?

user55340

@RobertHarvey No, you need the nurses after the soccer game.

user41796

@RobertHarvey That would be epic.

user55340

user image

user55340

user image

user55340

Btw, @Ampt go to despair.com/demotivators.html , click one, and notice who its good for. (I'm referring to Disaffected college students ) Though you might also find despair.com/consulting.html to be a good one.

8:41 PM

The NASA coding standard you cited explains exactly why they don't allow recursion, as I did in my original comment. JPL writes hard real-time software that controls space systems. You can't allow recursion in hard real-time systems because you can't predict how long the method will take to execute, or how much heap space the method will consume. Stack Overflows cannot happen in real-time systems that control rockets. But none of this applies to the average C programmer, whose programs are just as serious, if not necessarily as life-critical. — Robert Harvey ♦ 2 mins ago

user55340

8:52 PM

@RobertHarvey forums.anandtech.com/showthread.php?p=36097194 i-programmer.info/news/91-hardware/…

user55340

Also from safetyresearch.net/Library/BarrSlides_FINAL_SCRUBBED.pdf "Recursion violates a MISRA-C rule (1998: #70; 2004: #16.2)"

user55340

@RobertHarvey page 26 of that pdf link (slide presentation) specifically addresses the NASA bit... and Toyota too.

Ah, the stuff you never hear about in the news media.

But I'm still essentially correct, right? Most applications don't have these kinds of constraints. If @JimmyHoffa had to give up recursion tomorrow, he'd shit a brick.

user55340

Recursion works in some places... and not in others.

user55340

There are languages where the recursive solution is much better and not problematic (any more so than the iterative solution... or less problematic than the iterative solution)... but in other places its just "don't do recursion".

8:57 PM

and in most places you'd use recursion you can use a stack and solve it with iteration

user55340

I don't see Jimmy working at NASA or trying to push Haskell onto the next Mars Rover.

@RobertHarvey nah, I wrote plenty of OO like normal people for years, no recursion. As it stands I don't particularly write much recursion code.

Except when you're writing Haskell.

isn't haskell mostly recursion?

@ratchetfreak entirely recursion. There's no assignment, and no loops, recursion is your iteration, recursion is everywhere, and really the easiest way to do lots of things

9:00 PM

@JimmyHoffa in haskell

@RobertHarvey I do write recursive stuff in C# too, it's just usually indirected. Direct recursion only really comes up with simple graph traversal code, I do more graph generation than traversal; decision points generate next nodes etc, and in that way there's usually a layer or two between the tails

@ratchetfreak yes

even at that type of recursion though, it's still uncommon in the C# I write. I focus more on recursive types for fluent APIs and such rather than recursive functions, but more than anything I write CRUD code. I know ADO.NET and WCF as well as any senior .NET dev should and rely on that stuff more than the fun recursion etc. I just talk about all the exotic stuff in here because it's way more interesting to think about than the standard day-to-day enterprise dev crap we all deal with.

@RobertHarvey have a look at the Expression tree code I posted in that Q I had on SO earlier. It doesn't look recursive, but it's used in such a way that it generates recursive calls to itself; that's the type of recursion that does come in handy in C# - constructing control flows using composition to make decision code call itself until it bottoms out it's decision tree. The code itself doesn't look recursive because I've indirected it through the type system

nothing in CS can't be solved by more indirection besides the problem of indirection

user55340

9:19 PM

> (6) It is easier to move a problem around (for example, by moving
the problem to a different part of the overall network
architecture) than it is to solve it.

(6a) (corollary). It is always possible to add another level of
indirection.

user55340

tools.ietf.org/html/rfc1925

@MichaelT moving problem around is often a key part of problem decomposition, and problem decomposition is a key part of solving a problem.

user55340

@JimmyHoffa This doesn't say it isn't a step to solving it... its saying its easier just to move it to some other area that someone else will have to fix.

that's why refactoring so often results in simpler code and simpler solutions - you just move the problems around until they can be solved more easily

@MichaelT I know, I've seen that RFC before, just saying it's still good to nudge your problems around, unless you're working in MUMPS, moving problems in MUMPS just clones them.

user55340

"These computers are slow... we need upgrades", -- "well, since your home directory is on the network, you won't speed things up because its io bound, so its a networking problem"...

9:26 PM

@MichaelT "the part slowing you down is all that electricity cruising around the silicon, so it's a physics problem"

user55340

@JimmyHoffa You know those '1's stand up tall and get stuck, you need to push more '0's through the network instead.

user55340

pages.cs.wisc.edu/~ballard/bofh/bofhserver.pl

user55340

user image

@MichaelT if you turn the serif off your system's default font the 1s will fit around the network better

user55340

True story: In Bodie, CA (first long distance power transmission) they put the power lines straight because they were afraid the power would not make the turn. bodiehistory.com/power.pdf

user55340

9:30 PM

Ok, that's local folklore...

user55340

>
Local folklore asserts that the transmission line had to be “absolutely straight, no angles, no curves, which might cause the power to jump off into space.” (Cain 1956, 49) This delightful anecdote is not upheld by technical literature of the period, nor does it account for vertical curves (hills and valleys) that the line traversed. “The line crosses extremely rough country,” wrote Leggett in his paper to the American Institute of Mining Engineers, “not 500 yards of which is level beyond the town-limits. Most of the ground is very rocky, over 500 pounds of dynamite being used in blasting…

user55340

(better link: bodiehistory.com/myths.pdf for that)

In my experience, 90% of the "did you know that the yokels originally believed in..." chain letters are wrong. Actually, 99% of all chain letters passing themselves off as fact are wrong.

user55340

Dealing with 1890 here... sometimes they had some strange ideas.

I knew a guy in school who prided himself on being "factual," but was always being duped by the conservative conspiracy folks. I finally trained him to "snope" the emails he received before passing them on. I got far fewer of them from him after that.

9:38 PM

...we did use to try and autogenerate schedules for nurses in this product

but no one ever uses that feature anymore

@Zeroth of course not, everyone wants their favorite nurse.

it never saved any time, because we could never get it to account for all the necessary human factors that human schedulers can adapt for

so, yeah, BRE, very interesting, but I can tell you, the market does not want it, not for scheduling nurses, at least in Canada

user55340

Too many human factors - and it never survives contact with reality.

Nope.

user55340

(My brother is a nurse assistant and he is always getting called in for additional shifts)

9:41 PM

(It might also have to do with the fact that it was a giant piece of %&^%&^)

One interesting thing we noticed nurses doing with Availability was setting themselves as not available on every day except their scheduled day off

so they'd get automatic double time for those shifts

user55340

heh...

one we just heard about from a client, was that the labor laws in california mandated a 30 minute break for every 5+ hour shift.

employees were clocking back in on the old system at 29 minutes

so they'd get an extra 2.5 hours of pay, automatically

user55340

I know people at my former employer who would head out at 5pm for something to eat and get back 1h 30 min later and do another 3h of work so that they'd get the 2nd shift pay differential (if they just kept working, they'd still be on first shift pay... even if they work until midnight).

the new system rounds to the 30 minutes, saving the company almost a million dollars in overpayments

user55340

We've got break activity tracking for California in our app.

9:44 PM

who do you work for Michael?

user55340

@Zeroth Applied Data Consultants. I primarily work on the Elite Extra application (see adc4gis.com )

ahh, I work for Kronos(in a department that was acquired then the acquiring company was acquired by Kronos)

I work mostly on workforce ESP - which is pretty much only used by hospitals in Canada

user55340

Healtcare IT is one of those "don't want to touch" -- hipaa == lots of regs.

We don't deal with patient data

user55340

PCI was enough of a headache (or would be if done to the letter). At least those, a programmer can completely comprehend.

9:50 PM

just scheduling nurses, doctors, janitorial staff, etc

user55340

Somewhere, probably deleted, there was a comment or answer I gave of "a programmer doesn't try to learn every aspect of hipaa - you write it as best as you can with the guidelines you are given and then hire an auditor to go over your application."

"Bug can cause deadly failures when anesthesia device is connected to cell phones
No, it's not clear why anyone would ever connect a phone to a medical device." Ah Ars.

Hah @MichaelT

@Zeroth I read "amnesia device" and was unsure if such a thing would qualify as a "medical" device...

Well, erasing traumatic memories is actually a field of research.

I interviewed a researcher involved in that back in the college newspaper days

user55340

the spotless mind...

9:55 PM

^^

user55340

Eternal Sunshine of the Spotless Mind

Eternal Sunshine of the Spotless Mind is a 2004 American romantic science fiction dramedy film about an estranged couple who have each other erased from their memories, scripted by Charlie Kaufman and directed by Michel Gondry. The film uses elements of science fiction, psychological thriller, and a nonlinear narrative to explore the nature of memory and romantic love. It opened in North America on March 19, 2004, and grossed over $70 million worldwide. Kaufman and Gondry wrote the story with Pierre Bismuth. The ensemble cast includes Jim Carrey, Kate Winslet, Kirsten Dunst, Mark Ruf...

And also, playing tetris is still one of the better treatments

get a panic attack, play tetris. It ends up disassociating the trauma from your stress reactions.

@MichaelT dead true. We had a regulatory group who would usually have somebody as many of our design meetings just like a fly on the wall only giving us direction when we discussed doing dangerous things. He would be quiet for the most part as people had gotten a feel for most the hipaa stuff just from having him around for so long.

@Zeroth very interesting

Basically, the tetris treatment is about retraining your brain's response to the trauma(either a single inciting incident, or long exposure to stress and trauma) from an overreactive stress response, to a less emotional response.

Its pretty cool - we're beginning to figure out how to hack our brains.

@Zeroth wouldn't work for me; I fucking hate tetris. Not less emotional at all.

9:59 PM

Haha. :)

@Zeroth pretty sure the vikings figured this one out long ago, o different hack.. ;P

Well, not just the vikings :P

this is fun, I'm learning more and more about .NET Expression trees. Building these is basically like building S-Expressions but all through a bunch of ungainly Builder style classes and functions... at least I know how I want the S-Expressions to look from playing with LISP; the only way I can create extra layers is by nesting lambdas further and further like LISP.

user55340

@JimmyHoffa Viking fans however... they've gotta hack their brains to deal with such a disappointing team.

user55340

I wonder if @enderland is a viking fan?

10:10 PM

@MichaelT he works in VBA, he's not a viking fan; he's a viking.

@RobertHarvey Haskell has recursion in the source code, but the compiler generates machine code that is a unicorn with DNA made of unicorns.

user55340

@JimmyHoffa trying to figure out how to make that 'Viking B A'..

@MichaelT VIKING BASH ACCESS!

user55340

@psr at least it doesn't make machine code out of alligator eggs... then it would be confusing.

@MichaelT seriously. Alligator eggs are totally confusing. In other news I bet I could very easily encode the Alligator eggs into Haskell

actually, I guarantee you somebody else has... it wouldn't be but 10 lines of code if that to define all the alligator types and their interactions

user55340

10:15 PM

Btw, did you poke at haskellforall.com/2014/04/… ?

actually speaking of alligator eggs and haskell, I was looking for an implementation for the heck of it and stumbled across this: apparently alligator eggs is a result of one of the Haskell creators asking a colleague for help teaching calculus to his kid

user55340

calc? or lambda calc?

@MichaelT sorry, lambda calc

user55340

That's borderline child abuse then.

user55340

Which makes sense with the alligator eggs.

10:20 PM

lol

user55340

user image

user55340

Note the advertisement for a book on the side. I wonder if thats coincidence.

@JimmyHoffa Technically the language doesn't clone problems for you automatically, but most MUMPS programers have programmed themselves to never write any code without first copying at least 20 lines of old code to use as a starting point. Renaming the global variable XGPOL to XGPOM in the new code is optional, based on the programmer's mood at the time.

Just over an hour left at work sigh

user55340

0

Q: Can't log in from Windows Phone 7

Yes, I still have a Windows Phone 7. Laugh if you want. Whatever, I like it. I can't log in to stackoverflow.com. Tapping "Log in using Google" does nothing. I can't even select the text.

bug login mobile-web windows-phone

10:32 PM

@MichaelT probably not, coincedence the guy is one of the authors of that book (and the original Java Generics extension). It's logical he may have brought up alligators when talking to the guy about the lambda calc as he apparently has an association with them already.

user55340

@JimmyHoffa I think the aligator eggs predates the ora book... just an amusing bit that was there that I went 'huh' at.

I'd laugh, but I loved it too. +1. Well sort of; you can't see if I voted +1 or not anyway so you can just assume I did. — Jimmy Hoffa 26 secs ago

@MichaelT Eh? No way.

He worked on the Java Generics extension ~2000 or so (to my recollection - help me out here @gnat ?)

user55340

2006 was when the book was published.

@MichaelT well there ya go, and that blog post is dated '07

user55340

Yep... just saw that. Ok, similar timeframes then.

10:35 PM

the alligator was fresh in his mind from working on that book I guess. Strange the book came that long after the extension.

I'd laugh, but I loved it too. +1. Well sort of; you can't see if I voted +1 or not anyway so you can just assume I did. — Jimmy Hoffa 7 mins ago

Ugh, brain burn out.

@JimmyHoffa Nice comment. +1 on all your answers.

10:53 PM

ah stupid me pasted it twice.

It's getting late and today there's been too many fires. I'm so out of here.

11:08 PM

You two guys' icons look too alike.

11:21 PM

heading home

OK, Jimmy.

Jimmy + Jimmy's Sock puppet? haha

Francisco Noriega

11:51 PM

Anybody know if there's a studiostyl.es/schemes style / name for this color scheme? blogs.adobe.com/typblography/files/2012/09/SourceCodeSplash.png it's a really nice dark theme

« first day (1327 days earlier) ← previous day next day → last day (3662 days later) »

Transcript for

Apr '1422

The Whiteboard

General Discussion for softwareengineering.stackexchange.com P...

coffee-day horror-stories monads my-code-is-compiling scotch the-other-monitor

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile