The Whiteboard

I'd imagine you could work backwards pretty easily from the result, with something like "is X / (37 +i) an integer?" as your logic, with 37+(0,15) you aren't going to get many integers

user55340

That then means you don't have to worry about sign and % 2^64 vs % 2^63 -1 and such

user55340

Next, 910897038977002 - N % 37 == 0

user55340

The value of N is the last character in the string... Then do (910897038977002 - N)/37 and you've got a new number and repeat.

user55340

680131659347 ==> leepadg
indexOf(g) = 4.
680131659347 - 4 % 37 == 0
680131659343 / 37 = 18381936739

7:21 PM

@GlenH7 me too. Just put down my work to look at this haha

I don't even want the job but I want to do it

Heh, this is clever

Part of me wants to send them an email just for the hell of it, and basically say "hey thanks for the challenge, was fun!"

Did you get it?

yes

haha ok

I'm giving it a shot now

user41796

@enderland I'm doing something wrong. :-(

user41796

7:30 PM

The solution I came up with has 10 letters

user41796

nope, strike that - I missed the initialization of h

I bet you need to change ==0 to ==7

ha

i'm going to try to brute force it, just to see how long it runs :D

user41796

Just forgot....

user41796

Did yours come back with a vegetable?

7:34 PM

yeah

user41796

or put yours out there & delete and I'll compare

user41796

Yeah, I tested mine by running it forward through the hash again

user41796

@JimmyHoffa - if brute force takes more than 15 mins then it was faster to reverse by hand. :-)

it'd be funny to email them a scanned copy of doing that

also that's not @JimmyHoffa brute forcing it, but @MattGiltaji...

user41796

Honestly, I'd go for overkill and compare the performance of automated reversing vs. brute force forward

user41796

7:37 PM

As it becomes a stupidly parallel problem on the brute force approach

another idea which would be funny would be to get a list of all words using those letters that are 9 characters and check for "correctbatteryhorsestaple" results first

user55340

@enderland %37 isn't exactly easy to do by hand.

actually now that I think about it, showing the long division for each part WOULD be the finishing touch the ridiculousness of doing it by hand that was missing in my head, especially if it was written out on paper

user55340

Technology is a great thing. kickstarter.com/projects/bunchoballoons/…

I think we lost @Ampt

user41796

8:00 PM

@MichaelT my kids would love that

@enderland I wouldn't say I feel like I lost anything...

user41796

As a chill fills the air....

user41796

can C# recursively call a parallel for each?

user41796

(back to brute forcing the Trello hash)

durron597

hello hello hello!

8:08 PM

back

user41796

@ampt - yes

ok. let me go back and read what you guys did haha

user41796

Personally, I just used excel & reversed it

Heh

oh god damnit. if I would have read @MichaelT's though process I could have saved some time

lmao I should have kept this up

user41796

8:11 PM

@Ampt That wasn't the obvious first approach to you?! Tsk!

No, I figured that reversing the hash was the fastest way since their "hash" sucked

but I had it backwards with the modulus operation

I did 37 + N

as my modulo

Hey I posted before him about using the mod operator

I just had to do a straight divide by 37 at the beginning

user41796

Thou shalt obey the laws of operator precedence

because really, the first characters position is only multiplied the second time through

you start with 7, which gets mulitplied by 37, then you add the place of the character

ah well. It was fun to do on my own :)

the character's value only starts getting multiplied through after the first "round"

Thomas Owens

8:15 PM

I haven't read it yet, but there's an ACM article co-written the current chair of the program I graduated from and the two founders. University News Article ACM Queue Article

Oh hey. My company even gets a mention. Not bad, considering at one time, we had something like a dozen graduates from the program in my office alone. We're down to like 4 or 5 here and 3 in remote sites.

I'm sort of glad the only person on the team who knows JavaScript is leaving, it means I can write stuff like this and nobody can complain because they don't know the language at all

    controllerDeps.controllerDeps = controllerDeps
    (() ->
      directiveFunc = (dependencies...) ->
        restrict: 'E'
        transclude: true
        templateUrl: view.templateUri
        controller: ($scope) ->
          view.controller.call controllerDeps $scope
    ).call controllerDeps

granted that's coffeescript which they know even less because it doesn't have curly braces

user41796

@JimmyHoffa I CAN HAZ CURL BRAZES?

@GlenH7 can....

(I actually rather like coffeescript, it's quite simple)

from __future__ import braces

@Ampt from __past__ import braces

FTFY

8:21 PM

@JimmyHoffa whoosh

curly brace languages will be relegated to the likes of cobol in 15 years

File "<stdin>", line 1
SyntaxError: not a chance

@MattGiltaji you get used to it

if it's not haskell, and apparently coffeescript, Jimmy don't care

@Ampt but, but ... xkcd.com/353

I like python just fine

but hey @MattGiltaji, have you accepted Haskell into your home and into your heart yet? If not, I've an article for you to read.

user41796

@JimmyHoffa Pretty sure that if I were a mod I'd kick you from the room for that one.... :-P

8:37 PM

cant room owners kick??

What is this IRC and 1999

user41796

@Ampt nope, only mods

@GlenH7 you mean I've been biting my tongue this entire time for nothing?!

user41796

and it's been fairly rare for this room to need to kick someone and a mod not be around

user41796

@Ampt nah, think of how often there's a mod or SE employee in this room

8:38 PM

theres 3 sitting in now haha

watching you three

Hey @GlenH7

user55340

@JimmyHoffa Non curly brace languages will be relegated to the likes of pascal in 15 years.

user41796

@Ampt Apparently not since we reversed the trello hash much more quickly than you did. :-)

@GlenH7 I'll admit I wasn't watching then at all

On the other hand, I didn't need @MichaelT and @enderland to hold my hand ;P

Jul 10 '13 at 17:11, by GlenH7

and GET OFF MY LAWN ALREADY!!!

user41796

@enderland Pretty sure he already knows. Hence the petty comments in retaliation. :-)

9:07 PM

@GlenH7 already knows? what are you two raving on about?

user41796

Feigning ignorance won't save you on this one... :-)

Err am I on your lawn?

or is enderland implying you're old?

user41796

yes?

user41796

ultimately, he's the whipper-snapper in this case. But he knows that too, so what's the point in feeding his ego by pointing it out?

Apparently I need more coffee because this is all out of my grasp at the moment haha

user41796

9:12 PM

food coma hasn't worn off yet?

yesterday, by GlenH7

I'd agree with enderland's take on the matter

@GlenH7 ah crap, I missed lunch

too busy with that hash thing

user41796

@MattGiltaji not done yet? :-)

@MattGiltaji Hash thing?

you know Hashes are just isomorphisms, right?

user41796

2 hours ago, by enderland

https://trello.com/jobs/developer

9:14 PM

The other thing that mine does that I'm not sure yours does is account for multiple possibilities. theres a chance that you could get a character that causes a false positive

A friend of mine did it in R, pretty beautiful solution with using FP

that is HASH - N % 37 == 0 for two different N

user41796

@Ampt erm, no

user41796

37 is a prime number

@GlenH7 I was working on brute forcing, but got lazy and just went the recursive inverting route

user41796

9:15 PM

And the available set is much smaller than the number itself

@MattGiltaji I totally read this as "reversive incursive route" and was wondering wtf you meant

@enderland that sort of stuff is always nice in FP, FP is awesome at data manipulation

@enderland well I leveraged the synergy of the web 2.0 parallelism to paradigm shift the answer out

@JimmyHoffa I think his solution was about 6 lines of code, maybe 7

user41796

@JimmyHoffa Did you notice I dropped a recommendation for FP (Scala) in my answer last night?

9:19 PM

@GlenH7 hrmmm

user41796

Walk the hash algorithm forward for a few steps; note that you're always dealing with an integer value

ok, then it's a feature for larger character sets lmao

user41796

As a consequence of always dealing with an integer value, we know that the "correct" answer after dividing by 37 must have a remainder of 0

Good afternoon @all

I'm saying for possibilities where the number of values to try for N is larger than 37

user41796

9:20 PM

since the allowed character set is << 37, there's only one "right" answer at each stage

It's a feature!

IT HAS TREES!

user41796

@Ampt That wasn't one of the original constraints, but if it were then the problem would be harder to reverse, yes

user41796

The brute-force forward approach will get you more trees

well then I've already solved the non-existent harder problem on accident

I hang mostly over @ Code Review SE but I was wondering if I could ask a few quick questions here

user41796

9:21 PM

hence my wondering if C#'s parallel-for-each allows recursive calls

@Phrancis go for it!

user41796

@Phrancis fire away

no guarantee we can help

unless it's about beer. Then we probably can

user41796

@Ampt or that we won't start talking about booze

Just general questions. :)

9:22 PM

(But only if the question is whether we'll help you drink some)

@Phrancis Have you accepted Haskell into your home and your heart?

"What's the best algorithm to determine the optimal drink order?"

@enderland "all of the above"

@Phrancis If you answer his questions in the nude, he tends to go away. Always works for us

user41796

@JimmyHoffa oh quit scaring new folk already

9:23 PM

@GlenH7 I entertain me.

@JimmyHoffa I thought that was why we kept @MichaelT around

lol

@Phrancis we're on the edge of our seats. fire away

I should hang out here a little more. Then again we got our share of weirdness over at The 2nd Monitor

As long as your shenanigan tolerance is high

;)

user41796

9:25 PM

3 hours ago, by GlenH7

See, we start chatting about science and scotch, and the crowds start rolling in. The Whiteboard has the coolest conversation on SE.

So: I'm designing a business tool for my media production company. I am building a database using PL/pgSQL

user41796

Sorry, we've already claimed the title

And the user will interact via an application I will write in Java

My original thought was to design the whole database and stored functions in SQL first, then build the Java afterwards, but I'm having second thoughts

have you tried jQuery?

^^ negative

9:28 PM

good, it doesn't have anything to do with the problem you are trying to solve

So you want to know how much time to spend on fleshing out the DB first

What I'm thinking now is to design functions and the application side-by-side, one feature at a time

@Ampt yeah basically.

how well are the requirements defined?

First approach would make the DB fully functional quicker, but I would be the only user able to interact with it, which is not good

user41796

Separate out the two problems

user41796

9:30 PM

Design your DB & stored procs. Model the DB according to whatever problem you're solving

user41796

Better yet, build a unit test suite to validate that the DB works the way you expect it to

user41796

Then write some services to expose CRUD access to your DB

OK.

user41796

The objects the services use will become the business objects that the Java app or whatever other app will consume

user41796

Then when your DB or App or business needs change, you can change each piece in isolation of the others

user41796

9:32 PM

It's more work up front, but buys you a lot more flexibility and ease of maintenance later

user41796

Directly calling into the DB from your app is a recipe for a brittle disaster

That's good advice right there, thanks for your input

Have you tried storing everything in text files? That always takes care of my DB Problems

@Ampt at least use a csv file, jeesh

@GlenH7 yeah I've learned that very fast reviewing [php-mysqli] scripts lol.

9:33 PM

@enderland use HTML, then parse it with regex

user41796

@MattGiltaji So he'll now have 2, 3, or 4 more problems?!

@MattGiltaji that was pretty much what regex was invented for

A: RegEx match open tags except XHTML self-contained tags

nah, was fishing for this

4428

You can't parse [X]HTML with regex. Because HTML can't be parsed by regex. Regex is not a tool that can be used to correctly parse HTML. As I have answered in HTML-and-regex questions here so many times before, the use of regex will not allow you to consume HTML. Regular expressions are a tool th...

Q: What is meant by "Now you have two problems"?

I always say that if your regex can't parse your HTML, your html is wrong

user41796

106

There is a popular quote by Jamie Zawinski: Some people, when confronted with a problem, think "I know, I'll use regular expressions." Now they have two problems. How is this quote supposed to be understood?

9:43 PM

Thanks Santa!

@Phrancis yes.

one feature at a time

but don't make any of them perfect before moving on, get them working, then get them right

changing something after it's been polished is way more a pain in the ass than while it's still got the construction support framing in place

Getting conflicting information @JimmyHoffa @GlenH7 - Pros vs cons?

@GlenH7's approach is more ammenable to folks who don't need to come in here and ask for help on their approach - those more likely to get it right in less tries are better off driving towards completion from the start, those who are struggling a bit are generally better served aiming for small milestones and not trying to fit a whole design concept in their head at the beginning

user55340

In person interview on the 29th.

@MichaelT which person will you be in?

9:48 PM

@MichaelT grats

@JimmyHoffa 0.o

You know... I wonder where all Jon Skeet's extra rep goes. I mean he repcaps every day be noon

user55340

@Ampt he has a meta post about that somewhere...

I bet Joel has a giant pool scroodge mcduck style filled with all that extra rep

2

@JimmyHoffa I may go that route. I'm pretty solid with SQL but know nothing about Java so this will be a learning experience for me

let @MichaelT = CaptainAmerica in suit <*> @MichaelT >>= interview

9:53 PM

What's everyone's top-pick beer?

Personally I dig most IPAs, especially the very hoppy ones

user55340

Send email back double checking the reimbursement on milage... if I did my math right, they'd be reimbursing me about $200, when it would only cost me about $25 for gas (do people really have that sucky of mpg?)

@enderland I had a MUMPS REPL open so I just did it there in one line. It did print it backwards. Do I get the job?

DELETE FROM Work WHERE User = 'Phrancis'; GO

See you all next time, thanks for the advice!

@MichaelT how far did you drive? $200 in gas would drive my car most of the way across the US

user55340

@MattGiltaji I would drive about 180 miles down and another 180 back. At $0.55/mile this is $99/leg which is... well, high.

10:02 PM

@MichaelT thats a pretty intense daily commute

user55340

My car gets upwards of 45mpg... and I can do that trip on 8 gallons. Assuming that gas is $4/g thats $32

user55340

@MattGiltaji thats for an interview - not commute.

@MichaelT ah, key difference

@MichaelT No. But your car wears out faster. The number is based on some portion of the overhead being amortized into the mileage as well.

user55340

@psr I know... still, $0.55/m just seems a bit high.

10:04 PM

@MichaelT I think the IRS comes up with that number and everyone just uses that.

user55340

I can check (I keep tabs on car costs for all parts), but I doubt that my car is > $0.10/mile.

@MichaelT did you take resale depreciation into account as well?

user55340

For the last 6 months (gas, tuneups, etc..) I have $0.038/mile

@MichaelT or at least the sticker price/years you think it will last.

user55340

Over the entire lifetime of the car, its been $0.043/mile

10:06 PM

@MichaelT counting purchase price?

user55340

Including the purchase price of the car, its $3.02/mile

@MichaelT So depending on how long the car lasts .55/m could be reasonable

@MichaelT the fact that you keep track of this data is really cool

user55340

Possibly. Wish I still had my old car data... my '96 geo metro ($8k new) is still in good shape (my brother has it).

user55340

10:16 PM

You can see quite clearly the seasonal variation there.

@MichaelT what is the name of that app?!

user55340

@MattGiltaji itunes.apple.com/us/app/road-trip-mpg-mileage-fuel/id298398207