The Whiteboard

Ampt

20:02

Just found a 314 line IF{}ELSEIF{}ELSEIF.....ELSIF{} Statememnt

user20683

@Ampt ick

Ampt

@FrostEngineer Such Ick. It's in C and doing string comparisons....

Jimmy Hoffa

@Ampt Welcome to standard N^3 parsing

Perhaps I should call it the NCube

So many people live and die by the NCube

Ampt

NCube? Never heard of this? Is this the O(N)?

Jimmy Hoffa

(as many as live by it die by it which speaks to how few survive it)

@Ampt Nested Nest of Nesting

Ampt

20:05

Lmfao. I love that there is a name for this

is there no better way?!?!

HAVE WE NO TREES?!

Jimmy Hoffa

if{if{if{if{}else if{}else if{}}else{if{}else{}}}else if{}else if{}else{}}else{}}}

Ampt

.....this...file.... 2439 lines...

psr

@JimmyHoffa that's clearly some variant of LISP.

Jimmy Hoffa

@Ampt The result of the NCube, it's always the same.

Ampt

@JimmyHoffa Always 2439 lines? You should write a white paper about that.. that's a serious correlation

Jimmy Hoffa

20:07

@psr Agreed, LISP may totall suffer this problem - though I would say it's considerably reduced by the fact that LISP blocks have a single statement in them rather than having all those nesting blocks with 200-400 statements each

psr

@Ampt I call it pachinko programming. Sounds like a great terminology question for the main site.

Ampt

@psr Yeah! No way this would EVER be closed!

Jimmy Hoffa

@Ampt Unfortunately causality is not correlation so I can't bring myself to do it... I have a phobia of wrongly reported causal relationships.

(TVs scare me)

Ampt

@JimmyHoffa I'm just saying if every chunk of code that uses this methodology ends up at the same file size that's amazing.

I'm not even mad.

Jimmy Hoffa

@Ampt I'm mad as a hatter.

Ampt

20:09

@JimmyHoffa It's the Mercury.

psr

@JimmyHoffa Oh, that's totally caused by your phobia.

user55340

@Ampt Mercury is in retrograde?

Ampt

@MichaelT Freddie Mercury to be exact.

Jimmy Hoffa

@psr You just like to see me whince :(

psr

@notsureatthispoint I give Freddie Mercury a retrograde of "A".

user55340

20:10

@Ampt I should remind you... I worked in a file that was 10k lines long and had several hundred (as in 6 to 9 hundred) else ifs.

Ampt

@MichaelT Yeah, cuz you're a sucker!

user55340

Sep 4 '13 at 14:16, by MichaelT

I remeber working on receipt traversal where I worked before. It was written as a painful interpreter for an awful DSL. methods with 1000x "if(token.equals("foo")) { } else if (token.equals("bar") {} else if (token.equals("qux") {} ..."

Jimmy Hoffa

@psr I would go closer to "AAA", he seems to have a significantly more stable half-life than you give him credit for.

user55340

Jan 8 '13 at 20:22, by MichaelT

The part I work with most closely is the receipt printing, which is its own interpreter (nightmarishly done). How about a class file that is 10k lines long with methods 2k lines long that are of the format if(token.equals("something")) { code } else if(token.equals("somethingelse")) { code }

Jimmy Hoffa

Also he totally loves bicycles.

@MichaelT Been there, done that, please no bumperstickers; anything that reminds me more of it may reduce my half-life..

user55340

20:13

@JimmyHoffa embalming fluid (and twikies) can prolong that...

user20683

@MichaelT Happy recentish Birthday

Jimmy Hoffa

@MichaelT embalming fluid and twinkies are turing complete?

2

user55340

@JimmyHoffa Nope... they prolong your half life... unless you work at Valve in which case they're all about TF2 and DoTA (and micro-transactions) now.

user20683

@MichaelT Also consoles

user55340

I'm looking forward to the controller if it will work on a mac.

Jimmy Hoffa

20:16

I welcome our Robot Overlords. I will allow the mouse and keyboard to be removed over my dead body however. No more perfect gaming control system has ever existed.

psr

@MichaelT Me too if it's based on embalming fluid and twinkies.

Ampt

@MichaelT Funny, this code is exactly like that.

3 characters

did you ever come up with a better way of doing it?

user55340

@JimmyHoffa keyboard and mouse work very nicely... though there's something to be said for a nice controller... the dual stick games really don't work well without a dual stick.

psr

@Ampt There are definitely better ways of doing it.

Ampt

@psr such as???

In C mind you

user55340

20:19

@Ampt There are two options - one is to recognize that you've got a interpreter and redo it as a proper language, the other is throw out the custom logic and hard code it.

Ampt

@MichaelT no language changes. Embedded system. You get C or you get Assembly

(I'm guessing you don't want assembly)

user55340

For the receipt, it doesn't need to have a custom programming language once you've sat down and said "I want to write code, not a broken DSL"

user55340

@Ampt lex & yacc (or flex & bison) come to mind.

user55340

Now, if those will fit nicely in your system, dunno...

psr

@Ampt MichaelT is right. And if you do a proper language (your own mini language, not C), use a parser generator (in a functional language you can use JimmyHoffa's prefered method, parser combinators). But usually you shouldn't do it at all.

Jimmy Hoffa

20:22

@MichaelT wrong. wrong. WR-o_OnONGOGGG.

psr

It's something programmers often do to entertain themselves but it generally means writing code in your own horrible buggy language instead of the better language you wrote the parser in. Make a good API instead, if that can possibly work.

Jimmy Hoffa

@psr the typelessness of C allows for a bit of simplistic parser combinator approaches. You would have to know C fairly well to do it though, and you'd be throwing around function pointers willy nilly but it oughtn't be particularly difficult, provided you have a prior basis in parser combinators

Though in reality the more normal approach for precisely his situation which is totally valid (I'm stuck with C and in embedded how do I get a language?) Parser Generator.

user55340

Modernists would tend to go for antlr though (rather than the yacc/bison).

Jimmy Hoffa

The simplicity and unrestraint facilities of C allow for some FP approachs to be rather straightforward, more so than in strictly typed modern languages often

psr

@JimmyHoffa I've tried faking some functional stuff in MUMPS and the big issue is that lack of closures makes any could that would otherwise have used closures really ugly. I think that would also happen in C.

Jimmy Hoffa

20:28

@psr I doubt as much. It would be like as in LISP, you declare your locals and thread those with parameters that are passed in, and return/evaluate/pass-as-parameter function pointers. in C a closure is as simple as function pointer + struct pointer (binding and executing environment)

github.com/wbhart/Cesium3/tree/combinators

Alternatively there's a parser combinator library in C...

So you could just use somebody elses...

Ampt

Interesting... Not sure if we're ready to go there but it's something to keep in mind

psr

@JimmyHoffa But you still have to set up the struct for the executing environment, and populate it. Not sure how painful that is in practice.

Jimmy Hoffa

@psr Looking at the code for Cesium3 there actually, it's very little code and looks pretty clean and straightforward... neat to read a bit

user55340

@Ampt The key is to recognize that it is a language of some sort, either regular (possible, though unlikely), context free (likely), or context sensitive (possible but unlikely)... and write the appropriate parser for it. If its simple enough, you could do it by hand, and better than an if-else if chain.

Jimmy Hoffa

@psr The executing environment is simply a function pointer (no-anonymous functions but big deal), the binding environment in the struct is the tricky part if you're not familiar enough with monads. A normal person would create a new struct for each little parsing step that encapsulates the particulars of that step, monadic approach would dictate a single struct that your combinators have to create instances of

user55340

20:32

If it gets complex, best use a tool to write the parser for you from a description of the language.

psr

@Ampt If you are parsing you should really consider not doing it or doing it right. The mess MichaelT had to work with happened slowly. It's always easier to add one more if/else. Also, the parser generator tools are really good. It's not that bad.

Jimmy Hoffa

@Ampt yeah @MichaelT and @psr are both right on one point: Recognize it's a language problem- then figure out how to make it not a language problem. Language problems are very poorly solved in LOB software

user55340

The thing is that when you don't recognize what you're writing is a language, you end up with awful code that just keeps growing in that direction.

psr

@JimmyHoffa How would the single struct work?

Having a single struct could potentially solve the problems, but I don't see how it would work.

Jimmy Hoffa

I like Either:
struct Either { t_input * input, bool success, t_error * failure } (Look I *really* don't know C so ignore how terribly incorrect that must be)

user55340

20:35

@Ampt glance at publib.boulder.ibm.com/infocenter/aix/v6r1/… (ug, thats an ugly url... ibm clean your stuff up, you're a good resource otherwise) for a lex + yacc calculator.

Jimmy Hoffa

@psr That's the trick of the monad, you create a structure that encapsulates a single state of your control-flow, and then you design your functions to uniformly use that structure for dictating the state of the algorithm at each step which allows the control-flow to naturally move step by step, left or right through control-decisions based on that single uniform structure

user55340

epaperpress.com/lexandyacc/download/LexAndYaccTutorial.pdf is probably easier reading though.

Jimmy Hoffa

(that's the trick of the monad when it comes to control-flow monads like maybe and either; cont, actor, reader, writer, state monads are all decidedly not control-flow monads)

psr

@JimmyHoffa Oh, not solving lack of closures generally, just designing to only need uniform state to pass. Still a good point, you can probably often make that work if you are careful designing it that way.

user55340

(heh... that .pdf from epaperpress goes all the way to code generation of a simple calculator program...)

psr

20:40

@JimmyHoffa I guess a parser in particular doesn't need to pass much state, and it's quite uniform, so I could see that working out.

Jimmy Hoffa

@psr In general the state can be encapsulated of most large NCubes in some object which you then write functions that create generic decisions based on, parser combinators for instance have "isDigit" which makes a decision on whether or not the strings next character is a digit which pushes to the next step in the control-flow, imagine "isCredit" when the state object isn't a string but is a log of bank transactions etc, the approach can be expanded to many control-flow NCubes

user55340

20:59

@Ampt so, have those monad guys confused you yet?

Ampt

@MichaelT Huh? Oh, sorry, I was listening to music.

user55340

-2

Q: Alternative Power, connecting 3 power sources

How can I connect solar Panels to a water wheel and a wind turbine without blowing myself up?

discussion

user55340

Thats a MSO question.

user55340

Carefully. . — Lance Roberts 1 min ago

Ampt

Wish I could downvote hahaha

user55340

21:04

Its... not even a programming question

Ampt

It's perfect!

TO THE COLLIDER!

user55340

@Ampt I'm fairly sure that MSO is collider exempt... otherwise there would be many confused people.

user55340

Ok, I am going to connect my solar panel positive to the water wheel + and connect the -/-'s and see if it lights a small light bulb! — user246666 5 mins ago

user55340

I wonder if DIY has a "minimal understanding" close reason.

user55340

Nope... just "recommendation" and "arts and crafts or decorating"

user55340

21:18

@Ampt you really are going for the troll DIY... aren't you?

Ampt

No no no, you need to connect the -'s to the ground to ensure that you're properly allowing the flow of electricity from the sun and air to the earth (Electricity flows down) Now, hook up the light bulb between the positives and then string the whole thing up to a kite and fly it during a storm. This is how Electricity came into existence after all. Thanks Ben Franklin! — Ampt 45 secs ago

@MichaelT Ever seen the "How to recharge a battery" troll?

I crack up just thinking about it

He sells it so damn well

►

user55340

Right now... I'm tracking down lost hours:

user55340

> 89 days 22 hrs 60 mins

user55340

See? There's one.

user55340

@Ampt you are having too much fun on MSO.

Ampt

21:26

@MichaelT I like to think I'm having just the right amount of fun considering it's all imaginary anyway.

user55340

MSO is not the place for fun... its the place for all seriousness. And unicorns.

Ampt

I mean if it was an actual SE site, I might feel bad for misleading an honest soul but MSO? Yeah right.

user41796

21:58

@MichaelT unless it's friday

user55340

@GlenH7 true... and apparently its always friday in iceland.

user41796

@MichaelT heck yeah. I do think that's the only reason my suggestion wasn't down voted to oblivion

user55340

22:12

How do you know that person who is 1-0 isn't better than the 20-3? 1-0 could go on to be 100-0 but 20-3 could only ever be 100-3. — Ampt 43 secs ago

user55340

@Ampt Do you mean 20-80 for the second example?

Ampt

@MichaelT ripping apart commencing in 3...2...1...

@MichaelT I was trying to point out that they will never overcome those 3 losses

user55340

Also related to the question...

user55340

TrueSkill

TrueSkill is a Bayesian ranking algorithm developed by Microsoft Research and used in the Xbox matchmaking system built to address some perceived flaws in the Elo rating system. It is an extension of the Glicko rating system to multiplayer games. TrueSkill maintains a belief on the skill of each player; every time a player plays a game, the system accordingly changes the perceived skill of the player and acquires more confidence about this perception. Calculation Published formulas for Trueskill are not complete, and some needed functions are simply shown as graphs. Some of the math is s...

Ampt

@MichaelT I think that ELO is a pretty good approximation but it has it's own problems (Score protecting by not playing)

and it's already been thought out, so ++ there

user55340

22:15

Currently, I'm having some fun with that on Yucata.de because I've got some games I'm quite good at, but have no ranking in.

Ampt

as an aside: MAN are these head hunters BAD at their job

In the past two weeks alone, I got addressed as Hey XXXXXXXXXXXXXX, Wanted to reach out blah blah blah

this one got my current job wrong even though it's spelled out on my linkedIn (where they messaged me)

I've never even worked for the company they said I was working for

I mean holy crap, at least try

user41796

It's a good skill to learn - how to differentiate the good hunters from the bad

Ampt

Drew | Confidential
Hi xxxxxxxxxxxxxxx,

‘Glad we were able to connect via LinkedIn.

I’m hoping we can talk further. My contact information is below in my email signature. I’m an Executive Recruiter – what most folks call a Headhunter.

About Me:
I work with the Top 20% of people; those that stand-out amongst their peers, are committed to their work-life, and making significant contributions at their current firm.
I operate pretty much exclusively in the Manufacturing sector, and I specialize in all disciplines of Engineering, and Management – focused locally in Wisconsin.

user55340

Btw, there's a team lead and manager position open at my former employer...

Ampt

top 20%. Yeah. I can tell

user41796

22:21

Note their name and the firm they're with. Try to avoid in the future… :-)

Ampt

I wanted to respond and tell him that if that's what it takes to get "the top 20%" he may need to evaluate his scale a little bit

instead I sent this:
Hi Drew,

I'm not interested. Please remove me from whatever mailing lists you have me on.

Thanks,

Was I supposed to see that @Psr? totally missed it

@MichaelT you saying I should go for it?

user55340

@Ampt Consider the difficulties / expectations when a manager and team lead both get demoted at the same time and the challenge of charting the proper course for the team/department going forward.

user55340

I suspect its going to be one of those "not gonna touch that with a 10 foot pole"

user41796

@MichaelT demoted and not fired? FTS. Wouldn't even consider touching that one. Talk about a political nightmare to walk into

user55340

And the person they get to fill the role will be someone who has minimal technical knowledge from outside of the department (but within the company, looking to move up the org chart) who is unaware of all the politics involved.

Ampt

22:28

Politics: The foundation of every great bankruptcy

user41796

@MichaelT May God watch over them… Yeesh. What a horrible situation to be in.

user55340

@GlenH7 They're a bit short on trying to fill existing positions... firing the ~~incompetents~~ not cut out for management would leave far too many vacancies.

Ampt

@MichaelT Crash, crash, burn

user55340

The former team lead wasn't bad, just... wasn't really able to take charge and be a team lead... partly because she didn't have support from her manager to actually change things that needed to be changed.

user41796

@MichaelT Run, don't walk from that type of a situation. My previous employer could fit into that situation too

user55340

22:30

@GlenH7 this weekend I finished getting 150 points on Careers 2.0 so I could get invites to send to former cow-orkers.

psr

Hi Recruiter/PersonalAgent/Concierge,

Thanks for contacting me! That sounds like a fantastic job opportunit/entertainment opportunity/bit of dry-cleaning! I'd love to continue this process!

I am not aware of your current situation or identity, but wanted to have a lovely conversation about how you can further my life, which I can tell you are uniquely qualified to do. All interaction between us will be entirely imaginary, and confidential to you alone.

You can reach me via the following:

3

user41796

@MichaelT I didn't realize there were points to it. Then again, I haven't touched my profile there in a little bit. Perhaps I ought to...

user41796

Oh, and I know I should VTC that hackathon question but I'm curious to know the answer.

Ampt

Careers 2.0 vs linkedin
Is one betteR?

user55340

The former manager (M) got booted down to team lead in a different department. The team lead (T) of that team was looking to move up, but because of family ties (her (T) brother is a developer on the other team), couldn't become manager there... so they accelerated creating a new team and moved the team lead (T) into that team that had higher visibility and thus better chance at future proportions...

user41796

22:32

@Ampt LinkedIn has much greater visibility.

user41796

I'd do both, but LinkedIn would be my first choice if I could only do one.

Ampt

@GlenH7 That's what I figured. I'll continue to trudge then. Eventually I'll get tired of correcting the over-eager headhunters but until then it gives me some laughs

user55340

However, in doing that, the developers in T's former team aren't moving over, so in effect they're setting T up in a new, visible team without the people she's been working with for a long time... and putting M in charge of them when it was shown that M isn't really a good manager.... or really team lead either.

user55340

The reason M worked well as a team lead was that she had a manager before who she worked very well with...

user41796

@Ampt You only need "the one" to make it all worthwhile. So what if you blow off 99 others in the process.

user55340

22:34

@Ampt He's already on Linked in... just giving him another resource to use.

user55340

That said, the offers that have come through Careers 2.0 for me are ones that I would certainly consider if looking... while Linked in is very hit or miss... and recently, very miss.

Ampt

@GlenH7 Yep. Only need one job. And I've already got at least the next 18 months lined up so no rush from me to make recruiters happy haha

Ideally I want to be in the position where something like linkedin isn't necessary but time will tell on that front

Jimmy Hoffa

Last time I checked, if you want to be a good programmer in general, it's required that you never sleep. Don't worry, a programmers salary tends to afford one the necessary amount of mountain jolt, enjoy your last few night's sleep if you plan to start programming now. You'll not have more until submission at 'Ye Home of Old Folks' — Jimmy Hoffa 6 secs ago

user55340

> Men wanted for hazardous journey. Low wages, bitter cold, long hours of complete darkness. Safe return doubtful. Honour and recognition in event of success.

user55340

user55340

22:43

(that said, that precise wording doesn't appear to ever have existed - web.archive.org/web/20120505143103/http://… -- antarctic-circle.org/advert.htm )

Ampt

you know, I wonder if you can calculate your worth by the quality & number of head hunters after you....

user55340

@Ampt Sounds like a good senior research paper in machine learning.

Jimmy Hoffa

Trying to come up with a valuation method for selling shares or something?

user55340

@JimmyHoffa amazon.com/Unincorporated-Man-Dani-Kollin/dp/0765327244

Ampt

@MichaelT Research Paper? I have to make an actual working product for a company to sell.... I'm pretty sure I should be getting paid.

user55340

22:46

@Ampt Good point, if you did it as part of school you'd likely have that owned by the school...

Jimmy Hoffa

3 shares in Ampt for 30k each, man will they be mad though when you make another offering...

Ampt

I just have to bite this bullet for another 6 months. This is all so unrepresentative of the real world and it's painfully obvious to me now.

j0h

hi guys
i have a fortran program
581 Format(f7.3,6(f11.7))
772 FORMAT(f7.3,x,16(f8.4,x))
the 581 is in the first colum
and gives me this error
At line 784 of file rang2.f
Fortran runtime error: Expected REAL for item 2 in formatted transfer, got INTEGER
(f7.3,6(f11.7))

Ampt

EXCEPTION: FORTRAN NOT FOUND ON LINE 3

Jimmy Hoffa

Welcome to what makes going BACK to school a right bitch, @Ampt

user55340

22:49

@JimmyHoffa Undervalued... wired.com/business/2013/03/ipo-man/all

user55340

Consider that if I was to buy one, I want dividends for life based on your actual income (pre-expenses) and if I have two, I have a majority vote in what you do... guess who's cleaning my house.

Ampt

@MichaelT Clearly I would always hold 51% of the shares (For that exact reason)

user55340

@Ampt Read the wired article?

Ampt

@MichaelT reading it now.

> McCormick says she and Borenstein reached an understanding that he’d vote in line with her wishes. With Borenstein’s votes behind her, she managed to get the vasectomy rejected by a margin of 456 to 387.

I don't even... what.

That's the point where I ask myself if I've got this shit in writing anywhere...

psr

@MichaelT Uh, I guess if you decide that random strangers can make better life decisions for you than you can for yourself then it's probably true?

Jimmy Hoffa

23:00

The problem is they aren't making them for you, they're making them for them

psr

True, but given your track record of letting them do that...

You clearly have no capacity to make real life decisions.

user55340

> They decided against continuing the relationship but gave a 97 percent approval rating to a guy referred to as Jordon California. Feeling the weight of investor expectations, Merrill spent a drunken night “fooling around” with him. Like companies that stray from their core business to pursue an ill-advised acquisition (AOL/Time Warner, eBay/Skype), Merrill was slightly hung over the next morning

Ampt

I wish there was more

Very interesting social experiment.

psr

@Ampt Buy some shares!

Ampt

@psr Not for me I'm afraid. I'm old fashioned like that.

user55340

23:05

@Ampt read that book I linked a bit ago.

Ampt

@MichaelT I will, but the real thing is always better. I'm impressed at his resolve with this whole thing. The vasectomy would have sent up warning flags for sure but the dating thing probably would have broke it for me

user55340

@Ampt The book lets you go off into other aspects... though its the other way around. One normal person, everyone else living like that.

user55340

Though consider this... instead of paying to go to the university, the university gets some shares in you - they now have a vested interest in making you the most productive wage earner they can.

Matt D

23:22

morning

Transcript for