Forgot your password?
typodupeerror
Programming IT Technology

Are Quirky Developers Brilliant Or Dangerous? 1134

Posted by CmdrTaco
from the hey-wait-a-minute dept.
jammag writes "Most developers have worked with a dude like Josh, who's so brilliant the management fawns over him even as he takes a dump in the lobby flowerpot. Eric Spiegel tells of one such Josh, who wears T-shirts with offensive slogans, insults female co-workers and, when asked about documentation, smirks, "What documentation?' Sure, he was whipsmart and could churn out code that saved the company millions, but can we please stop enabling these people?"
This discussion has been archived. No new comments can be posted.

Are Quirky Developers Brilliant Or Dangerous?

Comments Filter:
  • by p3on (1245484) on Monday March 16, 2009 @10:15AM (#27209629)
    why are the mutually exclusive?
  • Funny... (Score:5, Insightful)

    by mdm-adph (1030332) <[mdmadph] [at] [gmail.com]> on Monday March 16, 2009 @10:22AM (#27209737) Homepage

    I've never met one of these coders in real life. For that matter, I've never been with a company who's internal politics would even allow such a person to exist.

    What cyberpunk novel does this hypothetical "Josh" live in?

  • simple economics (Score:4, Insightful)

    by circletimessquare (444983) <circletimessquare&gmail,com> on Monday March 16, 2009 @10:23AM (#27209771) Homepage Journal

    i remember a book from the dot com boom days claiming that a company in san francisco hired a network engineer who stipulated in his contract that he:

    1. would only work in the middle of the night
    2. got to work completely nude

    he got away with it, because it was simple economics: his services were needed badly

    any employee who has quirky behavior that is somehow provocative to fellow employees gets away with their oddball offenses to the extent that their services are needed that badly. beginning and ending of issue. you don't have any power or influence over the guy if he is that valuable. you just don't. so accept his behavior. you can moan all you want, but if you want the guy to disappear or act more uniformly, then just hope for a sudden influx of really good programmers from some magical place. thats the only way his behavior becomes a liability

  • Dr. House Syndrome (Score:5, Insightful)

    by EastCoastSurfer (310758) on Monday March 16, 2009 @10:23AM (#27209775)

    Sounds like Dr. House for developers. People think because they are smart and/or great at their craft they can basically do anything they want. This ties back to the /. article about the younger generation being more narcissistic than ever. Shows like 'House' glorify it and apparently make people think it is okay to be an asshole as long as you get the job done.

  • by Anonymous Coward on Monday March 16, 2009 @10:26AM (#27209813)

    As an antisocial mindshare person, I resent this topic. Because perpetuation of my antisocial liberties is the precise reason I developed subject matter expertise in the first place.

  • by Hozza (1073224) on Monday March 16, 2009 @10:27AM (#27209827)

    Translation: Control is more important than productivity.

    Err..No that doesn't really match what he's trying to say. By being so belligerent "Josh" was controlling the whole process.

    So the choice is: control by a passive-aggressive mentant who refuses to talk to you, or control by management , who should (in theory) be much more approachable.

    Of course, if you management team has fewer social skills than an unwashed anti-social 16 year old, then go with the mentant every time.

  • right tool (Score:5, Insightful)

    by trb (8509) on Monday March 16, 2009 @10:28AM (#27209833)
    If you need to cut, there's no tool as good as a sharp knife. If you need to turn a screw, a sharp knife probably isn't the right tool. If you have a guy who's a sharp knife, and you're using him to turn screws, maybe the problem isn't him. Maybe the problem is you.
  • by tgatliff (311583) on Monday March 16, 2009 @10:28AM (#27209837)

    Exactly... To the average layman, the thought of a "Dr House" type principle always applies. For the people who actually do high end development or research work, however, they realize that intelligence is only useful if the person can work with other people or can effectively communicate his work. Also, documentation of that work is essential...

    In short... it is only mutually exclusive if you are in a room full of a bunch of business MBAs who apparently as a whole still think that solutions come out of some magic hat somewhere...

  • by Lumpy (12016) on Monday March 16, 2009 @10:28AM (#27209839) Homepage

    Honestly....

    I dont care how good you are. TAKE A FRICKING SHOWER AND WASH YOUR CLOTHES.

    Really is it that hard to spend 10 minutes in the morning, EVERY MORNING to bathe yourself??

    and honestly, "really good coders" are not worth it. Give me medicore coders that understand business and can do what they are asked to do over a stinky quirky great coder any day.

  • by DrLang21 (900992) on Monday March 16, 2009 @10:32AM (#27209875)

    if the author has such a problem with this guy, maybe he needs to be skilled enough to replace him

    That's part of the problem. Having irreplaceable people on your staff is bad for business long term. If someone is laughing at you for asking for non-existent documentation that they should have written, they should be fired immediately. The cost to business if this guy were to leave will only get worse with time and probably already outweighs the savings of keeping him on.

    Lesson, you are replaceable. If you are not replaceable, then you are too dangerous to have.

  • by BlueBoxSW.com (745855) on Monday March 16, 2009 @10:33AM (#27209879) Homepage

    ... but they always seem to self-destruct on their own.

    They either:

    1) Take of too much work because they never know how to balance things, and burn themselves out.

    2) Stop working on needed projects, and only focus on the fun ones, which loses their value in the company

    3) Get Hooked on drugs and/or alchohol, and mess up their own future (MODERATION, people, moderation).

    4) Piss off management by sh!tting one to many times in the lobby.

    5) Get shown-up by some newbie coder who knows less than them, but is willing to learn new things (Josh doesn't like to learn new things, because it would imply that he wasn't a master of everything in the universe).

  • Rent-seeking (Score:5, Insightful)

    by Baldrson (78598) * on Monday March 16, 2009 @10:33AM (#27209887) Homepage Journal
    "What documentation?"

    The story ends there. "Josh" is no coding genius. He's a business genius. He understands that business nowadays is all about rent-seeking. Rent-seeking is looking for a parasitic niche from which you can milk the system with impunity, until the system collapses.

    How could anyone learn any other lesson from the goings-on in Washington, D.C. and Wall St. nowadays?

  • by Maximum Prophet (716608) on Monday March 16, 2009 @10:34AM (#27209911)
    If your competitor hires this guy they might be able to outproduce you just long enough to put you out of business. Doing things right is important, but staying in business is the *most* important thing. (It's a gamble, like all of life, you roll the dice and take your chances.)
  • by Joe the Lesser (533425) on Monday March 16, 2009 @10:35AM (#27209915) Homepage Journal

    Maybe there would be more documentation if you established reasonable deadlines.

    Just sayin' sometimes there's another story.

  • by Trifthen (40989) on Monday March 16, 2009 @10:35AM (#27209919) Homepage

    I was going to moderate this "funny," but thought the same thing myself. My answer to the OP's question was "Yes." Because anyone, really, can be these things, and we need to stop with the fallacy that only IT people can be self-absorbed assholes.

    Anyone can be brilliant. Anyone can be a jerk. Sometimes these two things overlap. I'm not convinced that there's a higher penetration of this in IT than any other profession.

  • by Don853 (978535) on Monday March 16, 2009 @10:36AM (#27209947)
    That's not true. The guy is refusing to document code and skips work on a whim. He's not dependable but he tries to tie his coworkers to his capricious tendencies. He's arrogant and socially inept. Most of the most brilliant people I've worked with are very confident, but they're not all assholes. This "Josh" doesn't sound like someone I'd want on my team. The code doesn't need documenting? Seriously? Brooks thought that was outdated in 1970.
  • by vlm (69642) on Monday March 16, 2009 @10:37AM (#27209951)

    What a piece of journalism.

    Quirky = rare habits and/or rare hobbies and/or rare background/culture that bring a smile to co workers faces or make them interesting to talk to, at least compared to an average drone.

    vs

    guy in the article = a-hole that everyone hates but has the redeeming characteristic of being somewhat productive (at the cost of ruining everyone elses productivity)

  • by neapolitan (1100101) on Monday March 16, 2009 @10:38AM (#27209965)

    Agreed totally. I wish more people realized this and thought like you.

    I, too, can write obfuscated code and appear "genius-like." It is a whole lot harder to bring *everybody* along than to rocket yourself ahead, make yourself appear to be esoteric and "invaluable," and, in a sense, bully everybody else into compliance. Now, we don't have enough details on the particular story to know if his colleagues actually were bad.

    However, I spend a good deal of every day helping people that may be not as quick or sharp as me in many ways, but that is my job.

    Finally a point regarding documentation -- I'm sure that every programmer here has come back to code that he/she wrote, and thought, "Man, this guy (me) is a genius. However, it just took me 30 minutes to understand how I did this!"

    Early on in my programming life, I thought this was indicative of my awesomeness as a programmer. Now, I just think it is poor documentation, and largely a waste of time. If I can't figure out how I did something a year ago, it would take other people twice as long... They may appreciate the clever implementation, but in the large scheme of things that is not efficient, nor awesome.

  • by russotto (537200) on Monday March 16, 2009 @10:38AM (#27209967) Journal

    Sounds like Dr. House for developers. People think because they are smart and/or great at their craft they can basically do anything they want.

    Right. And that must be stopped. Because extraordinary results shouldn't result in extraordinary rewards. Genius developers who can solve problems in an hour which could take the rest of your team a month or more should get the same cubicles and be subject to the same strictures as everyone else.

    Sorry, I'm not buying it. It's hard to compensate a quirky genius developer. You can pay them well (and usually have to), but that only goes so far -- they generally aren't like CEOs for whom money is the end rather than a means. Perks like an office rather than a cubicle are perfectly reasonable incentives, and so is "slack". If your genius developer doesn't document his code, a lesser developer can document it in far less time it would take any number of lesser developers to write and document it, or at least one of them isn't worth his salt.

    Spiegel has rigged the question by choosing, embellishing, or inventing out of whole cloth a "quirky developer" who Spiegel claims caused most of the problems he solved and went beyond what any company could tolerate (open sexual harassment). But just because his probably-fictional "Josh" wasn't worth the trouble doesn't mean it's a good idea to treat your best developers like interchangable code-monkeys for whom following procedures is more important than brilliance.

  • Re:Perhaps (Score:5, Insightful)

    by ultranova (717540) on Monday March 16, 2009 @10:41AM (#27210049)

    It's not necessarily "enabling". I've known a lot of people who are just eccentric but incredibly bright (and have been told I'm one, which surprises the hell out of me), and it's probably just part of the territory.

    A guy who mutters to himself while working is eccentric. A guy who insults his co-workers is an asshole. And a guy who smirks while informing others that documentation doesn't exist is just plain malicious.

    Assholes should be kicked out of any team, because no matter how bright they are, they won't be able to compensate for the lowered productivity of everyone else who has to waste their time and energy to deal with their little power games. As an added bonus, it makes every other employee happy, thus making the world a bit better place. Profitable and morally right, firing assholes is a win-win situation. Even the asshole might benefit from the wake-up call.

  • by Samalie (1016193) on Monday March 16, 2009 @10:42AM (#27210069)

    Now THAT is absolute truth right there...if I had mod points, I'd mod you up.

    I'm a SQL developer (yeah, the pansy-asses of the developer world - I admit it) - and often times my documentation is sorely behind. Of course, if I didn't have 50 projects all due within 10 minutes of the conception by the end user, I'd have time to document everything too.

    That being said, I *still* do my damndest to document my code. Its not perfect, but its better than the renegade who does nothing ever.

  • by ivan256 (17499) on Monday March 16, 2009 @10:45AM (#27210111)

    For the sake of argument, let's take for granted that nobody else can do what this guy does. Otherwise they'd have replaced him by now. Also keep in mind that he's using an extreme example to make a broad point. We'll take Mr. Speigel's words at face value, since we're to assume that he's not being hyperbolic about the behavior of said employee...

    "Sure, he was whipsmart and could churn out code that saved the company millions"

    His argument is that it's worth millions of dollars to not have to deal with this guy. Who has the bigger ego in this situation?

    If I'm running a business, and a middle manager tells me that the company should spend millions so the team doesn't have to deal with an asshole, I fire the manager, not the asshole.

  • by h4rm0ny (722443) on Monday March 16, 2009 @10:46AM (#27210117) Journal

    why are the mutually exclusive?

    Or related?

  • Re:Amazing (Score:4, Insightful)

    by Bieeanda (961632) on Monday March 16, 2009 @10:46AM (#27210125)
    "Josh" is the kind of guy who thinks he can develop the next Google, and that the shit he's taking in the lobby planter smells just like the rest of the roses. He's already missed the boat if he's in the workplace and still hasn't figured how to network himself properly.
  • Re:right tool (Score:4, Insightful)

    by Lumpy (12016) on Monday March 16, 2009 @10:47AM (#27210131) Homepage

    but my knife does not stink, mumble loudly, have action figures cluttering TWO cubicle spaces and refuses to empty that festering experiment of a mini fridge under his desk.

    My knife looks good and does it's job without offending all the other tools.

  • by morgan_greywolf (835522) on Monday March 16, 2009 @10:50AM (#27210181) Homepage Journal

    Exactly. There's always two sides to a story like this. One reason documentation often gets missed is because "make it work and make it work NOW!" and "we forgot to tell you, it also needs to Z in addition to X and Y!" gets nice'd above documentation.

    If we all had all the time we needed to do everything, the documentation would get done. But this is the real world and in the real world, IT management is definitely going to put functionality well above documentation on the importance scale.

  • by Anonymous Coward on Monday March 16, 2009 @10:55AM (#27210257)

    Agreed, I've got extensive real world experience with this. I've both been the quirky developer and now manage multiple quirky developers.

    I was definitely the arrogant whiz kid when I was younger, but my infractions were usually showing up late to work, wearing torn t-shirts, etc. Still, getting away with that stuff and still being fawned over by management makes everyone resent you. The fundamental shift came when I realized that although management viewed me as a great developer, I was never going to get promoted because they didn't trust my social skills at all. After that I put a great deal of effort into developing social skills (which was really hard) and stopped acting as if the rules didn't apply to me. Once that happened my career took off.

    I now hire plenty of the same "whiz kids" as I myself was, because I value their talent. But I do try to indoctrinate them in the ways of the real world. I try to make them have the same epiphany that I had, and show them that IQ and income/career trajectory are not proportional, and having a high IQ and being arrogant about it is a huge dead end. This is not something you can convince them of in one sitting, you have to keep pushing this idea on them for months. If it sinks in, the developer becomes a rock star at work. If not, they stay stuck at the bottom, or if their behavior is egregious enough, I fire them. But their talent is undeniable, so being dismissive of them as a manager is immature and cranky. You have to give them a shot.
  • by talldean (1038514) on Monday March 16, 2009 @10:55AM (#27210263) Homepage
    If you can't replace a relatively inexpensive employee, you're one traffic accident away from being out of business entirely. Let your competitor take that risk. "It's a gamble, like all of life, you roll the dice and take your chances." The odds of your competition hiring the guy - through a noncompete clause - and him being the tipping point of sending you out of business? Miniscule. The odds of a daily accident, or family problems, or the employee just leaving for greener pastures? Enormous.
  • by Registered Coward v2 (447531) on Monday March 16, 2009 @10:55AM (#27210265)

    Translation: Control is more important than productivity.

    I think it would be a lot harder for this guy to have made his point without such an extreme example.

    While I agree the example was extreme, the point was valid.

    It's not about control but about creating a product that is sustainable over the course of its life. That requires code that can be understood and troubleshot by others; not just the author. As was pointed out in TFA; Josh's code may have been the casue of teh problem from teh start.

    The ability to write code that works quickly is not genius, it's the mark of an idiot savant. Real genius is writing tight code that works and can be understood by others.

    Despite the pain of rewriting the code once Josh left I bet the company was better off in the long run because they had fewer customer complaints and when they did they could actually fix the problem without dealing with Josh.

  • by Rene S. Hollan (1943) on Monday March 16, 2009 @10:55AM (#27210269)
    Really is it that hard to spend 10 minutes in the morning, EVERY MORNING to bathe yourself??

    Sometimes, when you've spent the past 48-72 hours working to solve some crisis that some moron left and you have to clean up, yes.

    You look like shit, smell about as bad, and have a cranky attitude to match.

    But, shipping on time and avoiding a $250,000/day contract penalty can sometimes justify the hell (Ah, contracts with separate "code complete" and "QA Pass" dates.)

    Some coders don't shower all the time because they haven't gone home.

  • by rolfwind (528248) on Monday March 16, 2009 @10:58AM (#27210299)

    When kids are recognized as being highly intelligent and gifted, parents, extended family, and teachers go out of their to coddle them. To treat them as special. To give them far greater leniency and independence than kids with normal intelligence.

    Is it any shock that these kids grow up to think the rules don't apply to them?

    One of the pure group psychology shows I really like watching is Dog Whisperer. It's left unspoken, but I think a lot of it applies to kids and even adults in power situations.

    However, I don't think it's the gifted children that are specifically the problem, I think any type of kid treated with gloves becomes that way. The one that can't perform are merely arrogant losers as adults. While the ones that can become like Josh. The brilliant ones without the anti-social problems don't use their brilliance as a excuse and often don't call attention to it in the first place and may be skipped over as merely above average (which the Josh of the world may be but play it up, afterall, when you aren't hamstringed by stupid bullshit rules, you can do things more freely and eventually do things others never thought of in the box they've been confined in).

    But as a counter, I have to bring in the brilliant George Carlin:
    http://www.youtube.com/watch?v=w7LOCg4uKAg [youtube.com]
    http://www.youtube.com/watch?v=izE4_Jd2dOw [youtube.com]
    http://www.youtube.com/watch?v=f3XeRCAAkZY [youtube.com]

  • by mark-t (151149) <markt@@@lynx...bc...ca> on Monday March 16, 2009 @10:58AM (#27210303) Journal
    There's nothing wrong with clever programming tricks, as long as they are documented so that other people can understand them.
  • by Swizec (978239) on Monday March 16, 2009 @11:00AM (#27210327) Homepage
    You stop being a quirky genius soon as you declare yourself as one. Since then you're just a wannabe poser.


    See that's why I'm NOT a quirky genius.
  • by javelinco (652113) on Monday March 16, 2009 @11:01AM (#27210337) Journal
    I have yet to encounter a situation where anyone is THIS good (or where every other employee is THAT bad). And if a business WAS in that situation? Better to put it out of its misery - it'll get killed off sooner or later, regardless.

    I'm not sure why people feel a need to defend the "quirky" walking lawsuit that these "great" programmers are all about. Very few businesses need genius programmers in order to stay in business. And most of the time, these people keep your business one step away from being sued into an early grave - and they don't provide a good product. A good product isn't something that does things in really neat ways. It's a usable product, well documented, that does the job its designed for really well - and can be updated and maintained as necessary. None of these are true of any product worked on by the described programmer.

    I have no interest in pretending that programmers need to wear ties to work and act like your average joe. However, being anti-social and incapable of writing a maintainable product? Not interested.
  • by DrLang21 (900992) on Monday March 16, 2009 @11:02AM (#27210349)
    There are many excellent geniuses out there in the tech field that do what they're supposed to do. They document their work so that others can understand it. If they die or quit tomorrow, their company won't have to spend 2 years trying to figure out what they did. Getting a cheap geek to document these people holds its own high risk. What if the geek doesn't understand what they did? If this "genius" can't be bothered to document his own work, what makes you think they can be bothered to review someone else's documentation of their work? Mitigate your risk by paying more to hire a genius who won't put your company at risk of internal collapse.
  • Managers Job (Score:2, Insightful)

    by edivad (1186799) on Monday March 16, 2009 @11:04AM (#27210383)
    It's managers job to give the right amount of freedom to talented individuals. Individuals with well over the average IQ, with passion for the job, and that are no afraid to spend extra hours on the task. If it were for me, I'd give up ten of the 9-to-5, dumb, lazy a$$es, weasels, that are unfortunately the fabric of most of US companies, for one talented engineer. You wonder and cry because he has extra privileges than you? Meritocracy, comes not only in form of extra salary, my friend.
  • by Mr. Slippery (47854) <`ten.suomafni' `ta' `smt'> on Monday March 16, 2009 @11:04AM (#27210397) Homepage

    When kids are recognized as being highly intelligent and gifted, parents, extended family, and teachers go out of their to coddle them. To treat them as special. To give them far greater leniency and independence than kids with normal intelligence.

    Kids who are highly intelligent and gifted are special, by definition. Teachers and caregivers often find that the rules designed for age-group peers should not be applied, because the assumptions behind the rules don't fit. That's not coddling, especially when you consider the additional pressures of expectation placed on them.

    For example, I remember in elementary school (this is around 1975) it was a rule that kindergarteners could not books out of the school library. After all, reading wasn't taught until first grade, so kindergarteners can't read. When they found two of us who showed up able to read, rather than remove the rule entirely or stunt our learning potential, the rule didn't apply to us.

    Now, this has nothing to do with the sort of developer discussed in the article. A smart developer develops elegant and documented code, and is so proud of their work that they love to explain it to others. Someone who's mastered some arcane bit of technical lore and secretively builds convoluted, undocumented code around it, is neither smart nor talented nor an asset to their team. If they further behave like an asshole (not just quirky, but actively rude and abrasive), the only "special treatment" warranted is a swift kick in the ass.

  • by Maxo-Texas (864189) on Monday March 16, 2009 @11:05AM (#27210405)

    The problem is that smart people get very irritated working with fools.

    Our corporation has now cut our productivity by 75% in the last 5 years due to SOX related procedural changes. It takes 45 days to put a 1 line code change into practice.

    The smarter developers got irritated, then angry, then acted out, then most of them left. The few who remained were mostly burdened with debt and couldn't afford to take the risk. So they take anger management courses and let the corporation destroy them as people.

    There are appropriate places for smart developers-- in projects where they save millions of dollars.

    There are fewer and fewer jobs for smart developers. Corporations prefer predictable pleasant programmers over brilliant good programmers. They would prefer that a project *definately* take 16 weeks instead of taking 2 to 9 weeks.

    Even tho I was smart enough to move out of programming and into management, the culture slowly driving me insane as well. As far as I can see after a few promotions is that it is is turtles all the way up and the problem is coming from somewhere above 5 levels of management above me.

  • by ultranova (717540) on Monday March 16, 2009 @11:08AM (#27210463)

    Genius developers who can solve problems in an hour which could take the rest of your team a month or more should get the same cubicles and be subject to the same strictures as everyone else.

    Genius developers like that should be employed as designers, not coders. By the time you start writing code, there should be no problems left that take even an afternoon, much less a month, to solve. Besides, to put it bluntly, I sincerely doubt you are more than 160 times more productive than an average developer. Finally, dunno about the article, but the summary didn't talk about someone getting rewarded, it talked about someone going out of his way to be an asshole; and that should never be tolerated, since it will lower everyone's productivity.

    If your genius developer doesn't document his code, a lesser developer can document it in far less time it would take any number of lesser developers to write and document it, or at least one of them isn't worth his salt.

    Except it's more difficult to understand someone else's code than to write it in the first place. This is especially true if the code is "brillant" - meaning it has hacks and abuse of language features that make strong men cry - and even more so if it's actually brilliant, since that means it uses concepts and solutions the lesser developer could never even imagine.

    The smarter the original developer is, the more important it is that he properly documents his code, since it's less likely that your average coder will understand the underlaying idea and be able to produce meaningful documentation.

  • by Anonymous Coward on Monday March 16, 2009 @11:09AM (#27210487)

    I'm with you, I love to make my code full of documentation, but when you are given unreasonable deadlines and expected to meet them sometimes good habits just get thrown out the window.

  • Ignore... (Score:3, Insightful)

    by autocracy (192714) <slashdot2007NO@SPAMstoryinmemo.com> on Monday March 16, 2009 @11:11AM (#27210513) Homepage
    Drop down menus should come with an ok button. Undoing "overrated," which should have been "funny."
  • by radtea (464814) on Monday March 16, 2009 @11:11AM (#27210519)

    His argument is that it's worth millions of dollars to not have to deal with this guy

    Not even close. This is a complete false alternative.

    First of all, "assume no one else can do what this guy does" is about as sensible as assuming unicorns are going to come charging through your door. The number of people who can do what no one else does is extremely small. And if your company depends on something that esoteric you're doomed anyway.

    More importantly, you could replace this guy with a couple of top-rank developers who might not have his "genius" but who would be able to save the company 90% of those millions, and do so in a sustainable, maintainable way, creating far more value downstream while reducing the churn you experience on the rest of the team because no one can stand to work with this jerk.

    Team skills are empirically known to be the most important predictor of developer productivity, not technical skills. Go look it up in your copy of "Rapid Development".

    It is never "either/or": a better team player with somewhat weaker technical skills is generally a better hire than a guru who can't play nicely with others, and the notion that gurus are so singularly valuable that they can't be replaced is simply false if you are running a viable business in the long-term.

  • by bigdaisy (30400) on Monday March 16, 2009 @11:17AM (#27210639)

    it also needs to Z in addition to X and Y!"

    Or my personal favourite: "You know how we said X was critical to the success of the company and how you spent several months of your life implementing it? Well, we changed our minds. Now we think Z is critical. And can you make a few changes to Y while you're at it to make it more like W, so that when we change our minds again next week and resurrect X, Y won't work with X anymore and you'll have to redo X a different way from scratch? Thanks. Oh, and if you can have it on my desk by COB on Monday, that would be good, as I have a status meeting to go before my golf game. How's the documentation coming along? Not good? Nevermind, you can get back to it later."

    OK, so I'm reading a little between the lines, but I get that most weeks.

  • by Anonymous Coward on Monday March 16, 2009 @11:18AM (#27210655)
    I think the number of brilliant genius delicate hothouse flowers of programming genius are pretty rare. So rare, in fact, that it isn't even worth discussing (sorry Taco).

    I think a FAR larger cohort of programmers fall into a category of people who have been in one place a long time, know the systems (AND business processes) really well and have done little to no documentation.

    These people have you by the balls (as evidenced by me posting AC). If they leave you have to, in a lot of ways, start at square one. One can easily acknowledge that this was a management problem but even IF you try to institute best practices NOW these people will resist it because they know they hold power as operators of the Black Box. Every line of documentation removes a little bit of their power and leverage and it is the rare (and perhaps stupid) person who does that voluntarily.

    Yes, I suppose the ubercoders are a problem too but I think it is a much smaller, and different, problem than the one I described. If a company is depending on a coding wunderkind then they are probably on the cutting edge of SOMEthing and bad documentation is something of an expectation in that arena.
  • by Skreems (598317) on Monday March 16, 2009 @11:21AM (#27210717) Homepage
    Exactly. I'm always amazed by people who think that writing impenetrable code is "advanced". Any jackass can write something convoluted and obscure that nobody else can understand (or maintain) -- what takes actual talent is condensing complicated logic into code that's simple enough a ten year old would understand it. If you can write a complex system such that a teammate can open any random code file and think "what's so hard about that?", then you deserve some of the appreciation that "Josh" made a grab for.

    People like Josh, on the other hand, should be fired on the spot. I can't tell you how many hours I've spent cleaning up the mess left by people who thought that sheer volume of output was the measure of a good programmer, and I'm betting I'm not alone in that.
  • by qbzzt (11136) on Monday March 16, 2009 @11:22AM (#27210743)

    Paraphrased from something I read.

    Walking on water is nice - but to really bring value you need to freeze it, so other people will be able to follow behind you.

  • by Chris.Nelson (943214) on Monday March 16, 2009 @11:23AM (#27210767)
    If there is no documentation, the answer to the question, "Is it ready?" is "No." It's likely that the PHB doesn't know enough about what you're doing to disagree with you and grab your raw code from the repository and use it. If you establish a precedent for being done quickly (without documentation) then you get caught in a vicious cycle of it being expected that you'll be done quickly. It's best when the system supports proper documentation, etc. but if not, sandbag your estimates to give yourself time to do the job right, or at least half right. Over time, your productivity will catch up when you can figure out last month's or last year's code more quickly for a new feature because you took time then to document what you were doing.
  • by mcvos (645701) on Monday March 16, 2009 @11:24AM (#27210775)

    Why do people feel the need to control quirky geniuses who are doing nothing wrong? Seriously, there's nothing in this example that's out of the ordinary, except for the women's t-shirts. That's what you get for having a casual work place. My thought would be that if the author has such a problem with this guy, maybe he needs to be skilled enough to replace him.

    The T-shirts in the example are not the problem (though the hygiene might be). The problem is that he claims his code works and is self-documenting, when in fact it doesn't work the way the customer wants, and other programmers (the chief engineer at least) are unable to read his code.

    That's the problem. That and his anti-social attitude of calling them names instead of helping them to fix his incomprehensible code.

  • by ThrowAwaySociety (1351793) on Monday March 16, 2009 @11:24AM (#27210779)

    If your competitor hires this guy they might be able to outproduce you just long enough to put you out of business. Doing things right is important, but staying in business is the *most* important thing.

    The last thing you want is for your business to be dependent on one single person. Even if he's not some kind of cowboy/diva/jerk with no social skills, he may get hit by a bus, leave for personal reasons, or just get a better offer.

    Unless you're so small that you absolutely cannot hire another developer, you should start weaning your business off of such a person. Now, while it still has a chance.

  • by Hordeking (1237940) on Monday March 16, 2009 @11:31AM (#27210897)

    Lack of documentation only chains you more to a developer. It makes it that much harder for someone else to maintain the code base.

    So let me get this straight:

    If I conveniently don't document my code, I hereby increase my job security, because it'll make it harder to shitcan me because the other developers won't have a clue what I did because there's no documentation. Sweet, all I have to do is make it more expensive for them to shitcan me than keep me. It's pure brilliance!

  • Re:Perhaps (Score:3, Insightful)

    by GMFTatsujin (239569) on Monday March 16, 2009 @11:31AM (#27210909) Homepage

    Truth.

    I've always considered that the indispensable genius/jerk should be given as many independent entrepreneurial opportunities as possible. As in kicked out the door. Go run your own business, genius.

    I'll happily go head to head with someone who routinely alienates the people around around him and can't get his personal shit together.

  • by Myrimos (1495513) on Monday March 16, 2009 @11:32AM (#27210917)

    TFA is confusing "quirky" with "asshole." I love working with quirky people -- people who look at the world in entirely different ways, people who solve problems in a different manner, and people whose idiosyncrasies make them genuinely fun to be around.

    I can't say that I enjoy working with assholes, though. It doesn't matter how good your code is or how quickly you put it together, I can find somebody almost as good who's a lot less of a pain to work with. The extra little bit of efficiency isn't worth the metric cockton of assholery.

  • by mcvos (645701) on Monday March 16, 2009 @11:32AM (#27210921)

    You stop being a quirky genius soon as you declare yourself as one. Since then you're just a wannabe poser.

    See that's why I'm NOT a quirky genius.

    Exactly. There's nothing cool about trying to be a quirky genius. But if you happen to be a quirky genius, it's definitely cool to try to be a team player. That's what makes your genius valuable.

    A genius asshole is just another asshole.

  • by nomadic (141991) <nomadicworld.gmail@com> on Monday March 16, 2009 @11:33AM (#27210947) Homepage
    As a quirky genius, I have to say, if you don't like the way we do it... go fucking do it yourself. Should be good for a laugh...

    Please let us know what absolute brilliant pieces of programming have come from you so as to show that you are entitled to this attitude. At least let us know what companies you've worked for so we can at least know what world class programs you've helped produce.
  • by tedrlord (95173) on Monday March 16, 2009 @11:34AM (#27210965)

    They may appreciate the clever implementation, but in the large scheme of things that is not efficient, nor awesome.

    Whenever I hear the word "clever" relating to code, I cringe. I generally use it as an insult. In any professional project, clever code generally means "unmaintainable."

  • by Dan Ost (415913) on Monday March 16, 2009 @11:35AM (#27210981)

    If you can't be replaced, you can't be promoted. Hope you like your current job.

  • by Anonymous Coward on Monday March 16, 2009 @11:36AM (#27210997)

    What's the net cost?

    Let's accept for a moment the code may save millions of dollars as a line item.

    Does this guy's attitude cause more frequent turnover in his department incurring continual overhead in dealing with recruiters or new employee hire/training costs?

    Does his likely shoot-from-the-hip coding style work for his one use case and break things in ways he didn't bother to test or account for?

    Does he drop code in the repository that's of a huge, "everything's changed" variety which causes downtime as developers have to spend time trying to understand the quirks of his new million-dollar-savings?

    What was the nature of the change? Was it purely an infrastructure or refactoring type change that allowed maximization of existing resources, or did other departments have to message changes to clients? Was there any cost in client or customer service overhead as a result of this change?

    Were there additional refactorings necessary to user interface, help materials, etc as a result of this change that hadn't been budgeted?

    Great developers are awe-inspiring. Truly, I work with a handful of people who blow my mind on a daily basis, unlike anywhere I've ever been. One thing I've seen from all of these guys though is that they communicate their skills because they're smart enough to know that if they *all* know it (to a greater or lesser extent), or at least grok the important points well, they can improve upon it further. Compared with a lot of "superstar" guys I've worked with who couldn't be bothered to deal with people who "couldn't understand their greatness". Rarely are these guys the unmatched geniuses they think they are.

    Toxic employees are a liability... "superstar" is a label that is thrown around way too liberally.

  • Re:Perhaps (Score:2, Insightful)

    by Em Emalb (452530) <ememalb@[ ]il.com ['gma' in gap]> on Monday March 16, 2009 @11:37AM (#27211005) Homepage Journal

    Oh bullshit. Being an adult means not being an asshole to your co-workers.

    As you said, you've all got to get along, so why allow one jerkoff to ruin everyone else's day?

    I have no problem firing people that suck at life. I've never suffered for it.

    Don't let one ass-clown's childish behavior cause issues in the work place. You'll have a more productive work place because of it.

  • by DrLang21 (900992) on Monday March 16, 2009 @11:37AM (#27211013)
    I used to believe this. Then I found myself in a dead-end job because I was irreplaceable. I couldn't move up because the company couldn't afford to let me. So I found a new job. Unless you are content to be in one position for the rest of your life, being irreplaceable is bad for everyone involved. Now I contend that a wise employee starts training their replacement from day one.
  • by TheCarp (96830) * <sjc&carpanet,net> on Monday March 16, 2009 @11:38AM (#27211027) Homepage

    Meh I agree and disagree. I don't think you can put a blanket over all behaviour like that.

    The fact is, most of what we see as normal, is people doing prescribed things because its what everyone else does and whats expected of them. Anyone who has spent a few minutes questioning can see how superficial many of our daily activities are.

    I have seen entirely effective and competent entire teams that don't wear suits or even particularly nice clothes. Tshirts and jeanes. That same group, dark offices, loud music.

    Frankly i think most of what people lable as "eccentric" is just mundane difference. You don't really need people to conform in those ways. You don't need everyone to be a 9 to fiver with a collared shirt...

    what you need is the work to get done, and right, and the docs to be written. You need to make the docs part of the task and not fall for the old "the code is the docs" BS.

    What probably helps the most is integrating such people with others more. More positive social interactions, more work interactions where they can see that other people are important and they contribute.

    Assholes are usually assholes because they don't understand the people they are being assholes too. It comes, often, from judging others by the standards of your own myopic world. "I am the sysadmin, you are an idiot by sysadmin standards, your request makes no sense for a sysadmin".... but you arn't a sysadmin, you are with the helpdesk.

    I think to the studies of what happens in situations of authority (like the prison gaurd experiment), and think this is a lot of the issue... people judging and making decisions for others based on their own requirements. If my requirements and world works best when you are docile and do afraid of me and do as I say.... I treat you accordingly. Hence the "Instant asshole, just add badge" syndrome thats seems so common (at least among younger cops as far as I see)

    -Steve

  • by 0xdeadbeef (28836) on Monday March 16, 2009 @11:39AM (#27211037) Homepage Journal

    Yeah, and what if the owner of the company declares you one, and it happens in more than one company, and you regularly live outside the traditional chain of command of the company, answerable only to the owners?

    Do you have an equal share of the company as the owners? No? Then I hope the pats on the head are worth it.

  • simple, really (Score:4, Insightful)

    by Tom (822) on Monday March 16, 2009 @11:40AM (#27211061) Homepage Journal

    There's absolutely nothing specific to developers here. You have the same kind of people in every other job.

    The one question you need to find an answer to is this: Teamwork or solo heroes?

    If, for whatever reason, your project needs a team to work, say you want to support it for years to come and can not 100% guarantee that the one developer is still on board by then, or it is simply so large that you need more than one person to do it, then you absolutely can not use asocial people. Any and all attempts to somehow fit them into the team, or build the project around this inherent conflict will fail. You can't go faster than the speed of light, it really is that simple.

    However, there are projects where you need a lone hero. A crash project that needs to be done with next week, and can be shut down the week later, but it absolutely must be there during that time, and there's absolutely no way you can get it done while following procedure. Or - the more common case - you inherited a project that only this one dude even understands, and you don't have the manpower to replace it or reverse-engineer it. And sometimes, you have a project you want to fail spectacularily, and absolutely no team will give you the same show for your money that a fanatical lone hero can bring.

    So if you need a hero, then enable him, empower them, and suck it up. If you need a team, kick out the hero and make sure your team knows who to thank for it. Just don't try to have both. You can't. Been there, survived it, and I did, in fact, get a T-Shirt.

  • by fugue (4373) on Monday March 16, 2009 @11:43AM (#27211121) Homepage

    People like Josh, on the other hand, should be fired on the spot.

    I don't think so. They can just be recognised for what they are, and treated accordingly. Think of him as a fire extinguisher--a pain in the ass to clean up after, but from time to time invaluable. Sometimes you need a solution NOW, and you will have time to clean it up (or re-implement it more carefully) later. Perhaps your expectations for him were too high. Understand your resources and learn to use them appropriately.

  • I read the article (Score:3, Insightful)

    by br00tus (528477) on Monday March 16, 2009 @11:45AM (#27211149)
    First of all, the writer talks about how the coder was not "businesslike", but then goes on to describe him and others as "quirky...weirdest...weird...quirky...prima donna...he was truly crazy...his appearance was like Pig Pen...smells...'offensive' t-shirts...inappropriate...dark, cave office...quirky, crazy, irrational". Now how "businesslike" is it to describe someone like this? He is just slinging mud trying to bias me against this person instead of giving examples of what he has done. So what if his office is darker than the rest of the floor, it's his office, I've had to sit in cubicles where we are practically blinded by the light. And it's his private office, this guy seems to think he's a big boss and can tell people what the lighting level of their own private office is. If he is offended by someone not having a global warming causing office, I wonder what he finds offensive on a T-shirt, "Vote Obama"?

    Amidst all of this mud-slinging, we hear some actual examples of what Josh's supposed failings are. The first one is a developer on his team, who is responsible for implementing and patching version 1.0 of the code, decides to not do his work, and goes to Josh, who is writing version 2.0 of the code, and sounds like the head developer on that, and have Josh do his work for him. Josh tells him to fuck off as he is busy, on a deadline to write 2.0. Then Spiegel walks in. Spiegel is there to reprimand Josh for not pulling off his tight developed schedule, and deal immediately (without scheduling it) with a problem that his own incompetent developer can't deal with. Spiegel is shocked Josh isn't obsequious in the face of this demand. Josh's paycheck is dependent on him getting version 2.0 on time, why should he spend more than the 50, 60, 70 hours a week than he's currently working to dump everything immediately and go deal with a problem due to an incompetent developer who can't handle the work?

    So the story is Spiegel has an incompetent developer on his team who can't figure out code and how to do his job, so the bad guy is the coder who everyone including his manager says is the best, most brilliant coder, who won't drop everything immediately and go work on Spiegel's problem. After which Josh will either miss his deadline or have to work even more hours than he has to, and Spiegel looks like a star for fixing his problem. As far as curtness, I wonder if Josh worked 40 hour weeks, had things scheduled far ahead with reasonable deadlines and a full and competent support staff in place? Why do I have a feeling that was not the case? Spiegel had a developer on his team and mentions the 2.0 team Josh is on. So why didn't his own developer or someone else on the 2.0 team look at it? Because Spiegel wants the star of the 2.0 team to drop everything and fix his problem.

  • by nine-times (778537) <nine.times@gmail.com> on Monday March 16, 2009 @11:47AM (#27211197) Homepage

    The problem is that smart people get very irritated working with fools

    Of course, really extremely smart people can outsmart fools into getting them to do what they want. Really smart people get more irritated working with other smart people who have opposing agendas.

    There are fewer and fewer jobs for smart developers. Corporations prefer predictable pleasant programmers over brilliant good programmers. They would prefer that a project *definately* take 16 weeks instead of taking 2 to 9 weeks.

    I don't do software development, but as a manager, yeah, I'd generally rather work with pleasant people who do their jobs "slow and steady" rather than the "brilliant" but unreliable guy. The real issue is often not the uncertainty about exactly how long a project will take, but uncertainty about whether you can trust what you're being told how long a project will take.

    What I mean is, yes, I'd rather have someone working for me who says, "I can get this project done in 2-9 weeks" and gets it done in <9 weeks then someone who says, "I can get it done in 16 weeks" and gets it done in 16 weeks exactly. 9 weeks is shorter, that's an easy call.

    The problem is dealing with someone who says, "I can get it done in under 9 weeks," and then sometimes it takes 2 weeks, sometimes it take 9 weeks, sometimes it takes 23 weeks, and sometimes it never gets done. I'd generally rather take the steady 16 weeks over that sort of thing. A steady-16-weeks allows me to make other plans, make promises to other people, and set other deadlines. With the theoretically-9-weeks-but-who-knows answer, everyone would actually be better off being told, "I have no clue how long it will take," because at least then there would be no false expectations.

    All of this is just to say, I'd rather have people that I can rely on than theoretically brilliant people who are just going to do whatever the hell they feel like.

  • by fdrebin (846000) on Monday March 16, 2009 @11:48AM (#27211223)
    I have a simple policy, for myself and for people who work for me.
    If it isn't fully documented, it isn't done. There are no excuses, period.
    /F
  • by somersault (912633) on Monday March 16, 2009 @11:49AM (#27211239) Homepage Journal

    perhaps brilliant people are angry because life sucks.

    You poor fools who die boring, polite and pointless never realized the tragedy of it all.

    Funny how what you think of as the boring and polite people don't actually think life sucks, eh? I get frustrated at the attitude of some sheeple too, but there is something to be said for being happy.

    Sometimes these people are not actually boring either, just respectful. Speak to them in an informal situation and they might not be as boring as you had previously thought. You don't have to be a jerk to have fun.

  • by oliderid (710055) on Monday March 16, 2009 @11:50AM (#27211251) Journal

    Yeah, and what if the owner of the company declares you one, and it happens in more than one company, and you regularly live outside the traditional chain of command of the company, answerable only to the owners?

    It happens especially in little company. I remember a client of mine. They had an "in house" programmer. It was probably the shittiest codes I have ever seen. No logic, redundant functions, etc. I remember a meeting with him, he played the arrogant know it all in front of me. He finally noticed that I was a programmer just like him. The tone changed.

    He thought he was irreplaceable but the management has changed and the reason of my presence was to well outsource his work or to make him less "irreplaceable" due to the difficulties to get things right without crawling in front of him.

    There are a lot of people like him, as soon as they know "a little" more than their fellow co-workers, they feel like genius...Until they meet another professional who didn't past the last ten years sleeping on its knowledge without ever documenting himself about the last techniques/tools.

  • Re:Amazing (Score:3, Insightful)

    by mcvos (645701) on Monday March 16, 2009 @11:52AM (#27211279)

    "Josh" is the kind of guy that develops Googles, Yahoo, etc.

    No he isn't. Well, I don't know about Yahoo, but Google invests in smart, maintainable code. Josh writes convoluted code that nobody else can maintain, and he's unable to work with others. You can't build a company out of that.

    And there are far better coders out there who write self-documenting code that the other coders, the "average" ones, are able to maintain and fix.

  • by taustin (171655) on Monday March 16, 2009 @11:54AM (#27211307) Homepage Journal

    The first job responsibility in our employee handbook is that we have to get along with other employees. In other words, yes, it is my job (and everyone else's job) to not piss people off. Unless Josh can literally do every job in the company, he's not worth losing other employees over. No matter how productive he is, he's not the entire company.

    Plus, creating a hostile work environment is illegal.

    Any company that tolerates assholes like this will have no other competent employees.

  • by Dionysus (12737) on Monday March 16, 2009 @11:55AM (#27211345) Homepage

    Sometimes you need a solution NOW, and you will have time to clean it up (or re-implement it more carefully) later.

    Except, cleanup (or re-implementation) never happens. What will happen is layer upon layer to work around bugs and problems. Because you can (almost) never justify to upper management that you need to reimplement something that works and the finish product is basically the same you started out with (with cleaner code, maybe).

  • by jellomizer (103300) on Monday March 16, 2009 @11:58AM (#27211373)

    Um no.
    Because there are a lot of good people who can do the work and better and be a company player too. You are assuming that Josh's skills are irreplaceable. And that a good employee cannot do what he does. I am sorry, he is replaceable, and you can get a more professional guy to so the same job just as well, if not better because he is not so high on himself. I too have cleaned up messes after people like him. And let me tell you I have never seen any work by these guys that make me go wow this guy is my superior, in programming. Usually after a couple week I figure out the flow and I am just as productive as the guy was before, except people are willing to talk to me. Ask questions and raise problems that the other guy made them to afraid to mention.

  • Unmaintainable (Score:5, Insightful)

    by WED Fan (911325) <akahige.trashmail@net> on Monday March 16, 2009 @11:59AM (#27211387) Homepage Journal

    My local "Josh" is a genius, has gone from Athiest to American Indian to Christian to converted Jew (the last because he doesn't believe the miracles that Christians talk about), has a habit of telling the most inappropriate jokes, shows up when he wants, leaves when he wants, cannot/will not explain his code, will not code with others, insists the DB be designed to his standards, and produces code that does the job very well, but is utterly unmaintainable.

    He also collects the bonuses and gets the trips and training money. (The last, training trips and seminars that he usually ends up walking out of because they don't go along his lines of thinking.)

  • by adrenaline_junky (243428) on Monday March 16, 2009 @11:59AM (#27211389)

    Convoluted or obfuscated code is no benefit to anyone and deserves scorn.

    However, there is nothing about being a brilliant programmer that requires one to write convoluted or obfuscated code.

    The best code needs very little documentation because it is immediately obvious to any other programmer what the code is doing, and long comments are only required in sections of the code where the purpose of the code is not immediately obvious.

    It does require a dose of humility to intentionally add minor inefficiencies to a program to make it more clear, though, and it sounds like the Josh in this story was sorely lacking such humility.

    Using five lines of code to show the intermediate assignments in a long calculation instead of doing it all in one line of code can make the logic flow easier to follow by orders of magnitude, but some programmers are unwilling to deface the elegance of their design with such simplicities.

    Ultimately the true genius is the person who can write brilliant code that is also easy to read and modify.

  • Exaggeration? (Score:4, Insightful)

    by MBGMorden (803437) on Monday March 16, 2009 @12:00PM (#27211411)

    I wonder to what degree "Josh" here is an exaggeration. I'll say that I'm probably the "Josh" of our office - to some degree. I don't wear offensive t-shirts, but I don't normally do suit/tie (normally a polo and jeans - though often the polo is untucked). I am regarded as one of the few "people persons" in IT though (contrary to the "Josh" example).

    But, I am the odd ball out at work. I typically embrace new versions of software much quicker compared to many of our older workers who must be dragged kicking and screaming into an upgrade. I'll code stuff in whatever language I feel suits my mood and the requirements of a particular problem. I keep up to date with the latest technology trends.

    The "documentation" thing is one that I hear a lot. I try to document what I do. But some people can be unreasonable. For instance, I setup an amavisd-new email filter. Now, amavisd-new is a well known open source tool. It's ALREADY documented. Is that enough? Nope. I'm expected to write NEW documentation for the tool so that the rest of the IT department (who doesn't understand much Unix) can use it. I don't mean a diagram as in "here's how I setup the system", as in I'm expected to produce documentation on "How to release a quarantined message" when it already exists.

    Not to mention that when you look at how some of these people code and/or setup systems/databases, it's obvious just WHY they need so much documentation. The darned things don't make any sense. Without some ancient codex you can't make heads or tails of the system. So rather than do a clean and logical implementation, they'll do something that makes no sense and then go about it as if everything is OK so long as there's a written explanation of the kludge on file somewhere.

    I think far too often "Josh" might simply be getting mud slung his way due to the shortcomings of his peers.

  • by Anonymous Coward on Monday March 16, 2009 @12:01PM (#27211447)

    The key word is "definite". And the parent poster is correct, it's all about hitting dates. Any date will do, but hit it. Management has fallen in love with Gantt charts and Outlook calendars. Unfortunately, they have also fallen in love with outsourcing. This "love triangle" between management, calendar, and outsourcing results in satisfaction for no one.

  • by Anonymous Coward on Monday March 16, 2009 @12:02PM (#27211461)

    Productivity without control is like an accelerator without a brake. On a less extreme note, I worked with a guy who could crank out what seemed like klocs per day. Yet, when you looked at his code it mostly worked but had some dangerous problems, like servlets with member variables that had race condition issues.

    I inherited his code and told my boss I would need at least 4 weeks to clean it up, otherwise every little change would take a couple of weeks. The previous developer had a good mental map of the horror show that was his code so he could fix it in a day or two.

    My boss swallowed hard and I did the cleanup. I probably ripped half the code out and I reworked the rest into something approaching quality code. With the code base half as large, and appropriate javadoc comments, combined with a static class diagram, any idiot following after me should have been able to maintain it.

    And that's what being an engineer is. That's what makes engineering different from a guy tinkering in his garage. The guy tinkering in his garage may be able to bolt stuff together and make a gizmo, but it's a one of a kind only he understand how to make. An engineer may take weeks to do the same thing but it's a repeatable process, with standard instructions, that can be replicated and repaired as necessary.

  • by DaveV1.0 (203135) on Monday March 16, 2009 @12:06PM (#27211527) Journal

    Josh is paid to do his job. Part of the job of a programmer is to work with others. Part of most people's job description includes things about hygiene, appropriate work behavior, teamwork, etc.

    If Josh's coding results in ten thousand lines of spaghetti code with no documentation and riddled with single and double character variable names that result in other people not being able to do their jobs, then Josh is not doing his job.
    If Josh and his office smell like a pile of garbage, he is not doing his job.
    If Josh is saying offensive things to people, he is not doing his job.

    In other words, Josh is not doing his job.

  • by buddyglass (925859) on Monday March 16, 2009 @12:09PM (#27211577)
    I'd fire Josh just because he's jackass and I don't like working with jackasses.
  • by Weaselmancer (533834) on Monday March 16, 2009 @12:13PM (#27211639)

    Yeah, I was all angsty in my 20's too. I had long frizzled hair and wore an army jacket with patches all over it, and hated the world and all the stupid morons in it.

    I'm now in my 40's. I have a haircut, I'm sitting in an office cube wearing a polo shirt.

    And I've got some news for you. It's *all* pointless. The end is the same for everybody. We're all worm food. Doesn't matter if you rage against the machine or oil its gears. In a hundred years, I promise you it won't matter one whit.

    What does matter is what you do with the time your have. And I'll say this - I'm happier now at 40 with a nice job, nice house, nice car and a family I love dearly - however boring and polite it may be - than I *ever* was at 20 running around rebelling against everything mocking the stupid sheeple.

    My advice would be to take whatever brilliance you may have and apply it to your own life, if you're able. Solve your own problems. Find whatever happiness you can. Because sitting around picking at your own wounds to keep them fresh doesn't do a single bit of good.

    I have friends who never "sold out". They're miserable. Most are too poor to fix their missing teeth. If you sit around and tend a harvest of misery your whole life, then that will be your reward.

    To sum up, life only sucks if you work at making it suck. Let it go.

  • by Greg_D (138979) on Monday March 16, 2009 @12:13PM (#27211651)

    They should be recognized as douchebags and fired on the spot.

    Proper management and planning means you don't need a Josh on your team. The guy should have been fired before he was ever allowed to become so integral to their solutions that getting rid of him would mean pain for the group.

    There are very few irreplaceable workers in this world, and none of them work on code.

  • by Anonymous Coward on Monday March 16, 2009 @12:16PM (#27211685)
    mh, true and false. a pseudo code for a fast Fourier transform would be undocumentable - extracting radix: wtf is a radix? but a clean call to a method called fft() with a one liner pointing at the algorithm (eg Cooley-Tukey implementation of a fast Fourier transform) would change the mess in a functional block easily testable, replaceable and maintainable
  • by Maxo-Texas (864189) on Monday March 16, 2009 @12:19PM (#27211721)

    Close.. but not what I'm saying.

    Given a choice of:

    * Definitely 2 to 9 weeks. Over 9 weeks 5% of the time- file a project change with reason at week 7.
    * Definitely 16 weeks. Over 16 weeks 5% of the time- file a project change with reason at week 14.

    Managers choose 16 weeks.

    And if they do have a 2 to 9 week person, they'll tend to let that person sit around doing nothing for 7 weeks (actually vetoing projects because they don't want to "spend money" for a person that will instead sit at their desk doing nothing, worrying about their job. i.e. the money is already frakking spent).

    You have to let go or it will drive you crazy. You just deliver on the deadlines- with all of your paperwork correct (7 documents these days) and the appropriate status reports and weekly meetings attended.

    Any truly brilliant developer will be driven insane by this. The only option in my opinion is for them to go to consultant firms or small companies.
    When I worked at a small company, I was free to change code, put it into production-- on the SAME day. When I worked at a small company, I was free to tell my manager, "Hey I'm not busy right now so I'm going to improve the code". We didn't have to schedule a series of meetings. When I was a consultant, at least I could say, "I'm done- bye" instead of being tied to a desk with no work. And as a consultant, the paperwork was SOOOOO much lower.

  • by Miamicanes (730264) on Monday March 16, 2009 @12:20PM (#27211749)

    > The problem is dealing with someone who says, "I can get it done in under 9 weeks," and then
    > sometimes it takes 2 weeks, sometimes it take 9 weeks, sometimes it takes 23 weeks, and sometimes
    > it never gets done.

    In most cases like that, here's what really happens...

    Mgt: "How long do you think this will take?"

    Programmer: "Er, I guess 3 months or so, assuming nothing major goes wrong along the way."

    Mgt: "That's too long! We need it in 8 weeks. Can it be done?"

    Programmer: "I doubt it. Maybe if god parts the skies and makes a miracle happen."

    Mgt: "It's really, really important. In fact, it really needs to be done in 6 weeks."

    Programmer: "You're insane. There's no way in hell it's going to happen."

    Mgt: "OK, I'll allocate 8 weeks."

    Programmer: (sigh) "Whatever."

    8 weeks later ...

    Mgt: "Is the program done?"

    Programmer: "No. We'll probably be done in another month, maybe two at worst."

    I think you see where this is going. The programmer had a decent idea of how long it would take, and could have probably given a more realistic estimate within a few days had he been encouraged to identify the riskiest parts of the project (specifically, third-party libraries and things constrained by real-world hardware/network performance) and try to tackle them *first*. However, if management twists his arm backwards, or keeps pressuring him for a "better" (ie, shorter) estimate, he'll eventually get disgusted and throw them the number management wants... rationalizing that it's not *quite* a lie since miracles occasionally happen, and absolving himself of any moral responsibility for actually agreeing to a deadline he views as ridiculous since he was coerced into it.

    That, IMHO, is the root of more miscommunication between management and developers. Far too many managers don't quite understand that programmers *hate* interpersonal conflict, and will casually agree to just about *anything* if they think it will get the person to quit bothering them. The constructive way to deal with it is to begin by asking the programmer for a range (best case vs likely worst case), then ask him to identify the riskiest factors influencing the range, then nudge him to tackle those factors first so a better estimate can be refined quickly. Just don't make him feel like you're twisting his arm or browbeating him, because estimates are like information from interrogation -- torture will get you the answer you want quickly, but the answer itself will likely prove to be worthless.

  • by Eunuchswear (210685) on Monday March 16, 2009 @12:22PM (#27211769) Journal

    Only sheeple use the word sheeple.

    Baaa.

  • by Rorschach1 (174480) on Monday March 16, 2009 @12:24PM (#27211805) Homepage

    Generally true. Sometimes clever is necessary, though. I do most of my programming these days on embedded systems, where size and speed are absolutely critical. I'll occasionally do something horribly non-standard and convoluted (usually to avoid writing even more annoying inline assembly code), but I've learned to allow about a 3-to-1 comment to code ratio in those cases. Even something not that complicated but just unusual (casting a char array to a function pointer and calling it because that particular buffer is the only one available to hold the flash programming code that has to be copied down to RAM, for example) warrants a clear, concise description of what the hell is going on.

    No matter how sure I am that I'll remember how something works and why I did it, I still try to always comment it. I'm sure everyone here (who's been programming for more than 3 years anyway) has gone back to code they wrote 3 years ago and thought "what the hell is this, and what was I thinking?". In my experience that's usually followed by a quick correction, and then after a few hours of chasing down some obscure bug that subsequently appeared, remembering why you did that in the first place and putting it back the way it was.

  • by squoozer (730327) on Monday March 16, 2009 @12:27PM (#27211851)

    There is a flip side to this though. One place I worked had a habit of trying to make sure any code they wrote was always as flexible and general as possible so that future requirements could be satisfied easily. The problem was that everything took twice as long to develop and was unnecessarily complex. I noticed that the developers were generally fairly poor at predicting what the new requirements would be or when they would be wanted. I'm not advocating that every piece of code should be written as the shortest path from A to B but I think the best solution is somewhere between general and rigid. Your in class example isn't great because you knew up front that there would be new requirements in the near future therefore hinting to you that a flexible base is a good thing. In the real world you rarely get such hints.

  • by aurispector (530273) on Monday March 16, 2009 @12:38PM (#27212053)

    The key difference is the willingness to work well with others. In any organization you need cooperation or long term it just won't work. Coding strikes me as a task that's particularly vulnerable to long term maintenance issues if it isn't properly documented/commented.

    You can blame management for putting up with difficult people, though perhaps they simply don't understand the depth and breadth of the problem. Sometimes all it takes is for someone to confront the person in question with a set of expectations - and be willing to pull the trigger if the standards aren't met.

  • by Venik (915777) on Monday March 16, 2009 @12:42PM (#27212141)
    It is equally amazing how programmers of average ability insist on branding brilliant code they have trouble understanding as convoluted and obscure. The only thing that matters here is the bottom line. If Josh produces code that makes the company millions, then this is all that matters. It is entirely irrelevant if some of Josh's obtuse co-workers with a pronounced inferiority complex think that his code is convoluted. Most managers I know would rather fire every idiot complaining about Josh's shenanigans, than to fire their obnoxious but talented cash cow. I had the privilege of working with a couple guys like Josh. Understanding their work and their methods may be challenging, but it is also rewarding. Most can't stand brilliant co-workers, but not because of their alleged eccentricity. Bell curve riders feel inadequate and threatened working with talent. They demand endless meetings, ceaseless telecons, and detailed documentation, as if reading documentation would actually make them understand genius. People like Josh rarely get fired: they just get tired working every day with the same morons and go for a better-paying job elsewhere.
  • by Kell Bengal (711123) on Monday March 16, 2009 @12:48PM (#27212257)
    I've been in exactly this situation: we were an custom GPS electronics company where one very talented electrical engineer built the hardware from the ground up (and he and a whole team of software guys did the code). I signed on as his lackey to do additional electronics development on the side because his time was 'so valuable' and they needed more stuff done besides

    .

    The -very- first thing they had me do when I arrived was produce page after page of documentation on how the hardware actually worked so that the software guys could understand it. It wasn't ground-breaking design, it wasn't super complicated, but it was subtle and you couldn't get the whole idea of what was going on without being able to speak Engineer (specifically the EE dialect). A lot of people in the company were terrified that he'd walk out one day and get hit by a bus and the company would have to spend a fortune it didn't have for a team of engineers to come in and tell everyone else how their own system worked.

    When I asked him why there was no documentation (or very poor documentation when there was) the answer was a combination of "You shouldn't need documentation" and "I'm not paid to document things."

    Well, actually... you are.

    A few early experiences counseled me very strongly to enforce good documentation practices in my code and hardware design. Any design more complicated than a blinking LED (the hardware equivalent of 'Hello World') requires it - if you aren't documenting, you're not doing what you're paid to do. As TFA says, End of Story.

  • by Maxo-Texas (864189) on Monday March 16, 2009 @12:53PM (#27212343)

    I've worked with many brilliant programmers over the course of 25 years.

    Good coding *is* difficult.

    Managers want to think that programmers are a generic goo. Our 2nd to last reorg 3 years ago took everyone away from their areas of competence because some management company said it would be a more efficient allocation of our resources. After suffering with it 3 years (I think because they didn't want to embarrass the executive that backed the plan), we have a brilliant new organizational plan-- we put people in their areas of competence.

    A good programmer can do quality work and quality documentation in 1/10th the time, with less architectural/design issues than a pleasant competent programmer. A brilliant programmer MUST write quality code- sub-standard code and design pisses them off and drives them crazy. You don't want *two* brilliant programmers in the same area. You need a brilliant programmer and 3 to 4 solid pleasant programmers and 1 to 2 eager novices.

    You don't need an asshat like Josh, but you may find your brilliant programmer does his ( I've met some brilliant female programmers recently but the vast majority are male ) best work from 1pm to 1am so you have to focus on project deadlines, not arrival times.

  • by Beardo the Bearded (321478) on Monday March 16, 2009 @12:55PM (#27212397)

    It's not 1985. Comment space is cheap.

    First, most functionality should be in functions (or methods, depending on your language). You wouldn't put stuff inline because that's a nightmare. However, at some point, you're going to have to write code that actually performs some kind of operation upon the data.

    If you called the function fast_fourier_transform() or fastFourierTransform() (depending on your coding style) it would make it a lot easier on the maintainers and cost you nothing.

    No matter what you called it, you'd still have to document the transforming function so that if you'd made a mistake in it, the next person looking at the code would be able to say, "oh, hey, this is supposed to be multiplied by -1 here"

  • by RightSaidFred99 (874576) on Monday March 16, 2009 @12:59PM (#27212465)

    Well, it's a myth that these great coders are valuable, as well. High level software development requires more than the ability to manage complexity. You won't find any Josh's developing high quality, vast enterprise applications. You won't find them developing a modern RDBMS, or anything that's _truly_ complex in terms of architecture, scope, and interaction with other systems.

    You'll find Josh's hacking away on some custom application developed by a small or medium sized company and it'll be their one trick pony.

    The reason is simple - to develop a large system it takes many people, you can't have a one-man show on large software products.

  • by Hoi Polloi (522990) on Monday March 16, 2009 @01:01PM (#27212515) Journal

    That "millions" number was probably pulled out of thin air but anyway...

    How much would it cost to replace the good coders who might leave for a better work environment? How much would it cost to debug or port that guy's code after he left with no documentation? How much would it cost to defend the company in, say, a hostile workplace lawsuit?

    Even the president of the US can be replaced as the last election showed. No one is irreplaceable, especially a hostile person.

  • by jellomizer (103300) on Monday March 16, 2009 @01:04PM (#27212561)

    The reasons that the "Josh" usually doesn't get fired is because the company has a small enough Dev Team that makes him the big fish in the little pond. And most of the people don't know that he is actually quite bad. His code may make millions but the truth is someone else can write code that will make the millions too, and make change over that much smoother. So they keep him as they know the mess it would be when he leaves and hopes they can change jobs before it falls on their head.

  • by fugue (4373) on Monday March 16, 2009 @01:09PM (#27212647) Homepage

    Because there are a lot of good people who can do the work and better and be a company player too.

    Um no.

    The article was about someone who can do an incredible amount of coding in a very short time. Indeed, more coding in less time than most anyone else.

    It isn't that they weren't smart. In every case, these "great" developers were the most talented in the group. Their intellectual abilities and problem solving capabilities were unparalleled.

    Obviously, if there are a lot of people who are equally fast and one can't work with teammates, then fire the asshole. But that wasn't the question. The question was how to deal with one asshole who can churn out more code faster than everyone else. You can tell me you're as good as anyone else until you're blue in the face--in which case your employer is lucky to have you--but that's not what the article was about. It was about someone who was much better than you (for a certain very circumscribed--but potentially useful--definition of "better").

    Sure, if you spent sufficient time and money, you could find someone better than Josh. And ideally you do. There's always someone better on the market, out there, somewhere--especially in this economy. But if you have a limited budget for finding and hiring the best of the best, then an acceptable compromise is learning how to use the people you have. Understand what he is, and isn't, good for, and offer him what he's worth to the company. And learn to use him wisely.

  • by NormalVisual (565491) on Monday March 16, 2009 @01:13PM (#27212703)
    It is entirely irrelevant if some of Josh's obtuse co-workers with a pronounced inferiority complex think that his code is convoluted.

    Until he quits, is hit by a bus, or otherwise becomes unavailable to maintain the undocumented, uncommented spaghetti mess that usually comes from these kinds of people. When the total cost of writing and maintaining "brilliant" code exceeds that of "average" code that provides adequate performance, then there's a problem.

    On the other hand, I've also had the pleasure of working with developers that are far and beyond my skills, but also are decent people to work with and document their work so that it's understandable without having to spend a week tracing through it. Such people are worth their weight in gold.
  • by mdielmann (514750) on Monday March 16, 2009 @01:14PM (#27212733) Homepage Journal

    And another Einstein quote I try to live by in my programming career: "Make everything as simple as possible, but not simpler."
    IMO, MS Office ribbons fail on the second part.

  • by jellomizer (103300) on Monday March 16, 2009 @01:24PM (#27212899)

    Having Asperger's isn't a good excuse to do a poor job or to be anti-social, or unprofessional. Yes you may have hard time following the right non-verbal queues. But things such as dressing appropriately for work, using the bathroom in the right spots, and a lot of the quarks that happen are due to bad behavior that people even with serious Asperger's can work one and minimize and be at a professional level. I don't take the idea, that I have a disability so you need to deal with my Crap mentality, it is basically reinforcing that they can behave badly, without having them work on improving themselves.

  • by shutdown -p now (807394) on Monday March 16, 2009 @01:25PM (#27212919) Journal

    Except, cleanup (or re-implementation) never happens. What will happen is layer upon layer to work around bugs and problems. Because you can (almost) never justify to upper management that you need to reimplement something that works and the finish product is basically the same you started out with (with cleaner code, maybe).

    Agreed; which is why this statement from TFS: "... could churn out code that saved the company millions" - is nonsense. It may look that way on the surface, but when accounting for all the code maintenance pains that inevitably follow, I've yet to see a single such "genius" that wasn't a net loss. What's worse, the expenses are quietly swept under the rug, or, even worse, shouldered by the rest of the team who gets flak when they can't keep up with the "genius" (because they're cleaning up after him).

  • by MartinSchou (1360093) on Monday March 16, 2009 @01:44PM (#27213317)

    I'd place myself as slightly behind the bell curve (out of practice) at the moment (slightly ahead when up to speed). I'm quite capable of recognizing brilliant code. I'm also able to recognize code that makes absolutely no sense to me. At times those two are the same thing (meaning that the code is doing things way above my comprehension, not that the code doesn't work).

    The thing is - if that brilliant code turns out to have a subtle flaw or needs to be redesigned, how can you be certain that the originator is still with the company or the project? Sure, that brilliant code may have saved your company millions, but when the flaw allows people to siphon money directly from your bank account, how would you rather go about fixing it? Stepping through convoluted code or reading the documentation? Sure, "my code just works" is a nice reponse. I'm also certain that Einstein was a lot smarter than most of the Josh' out there, and if he'd just said "E=MC^2 - trust me" people would have told him to fuck off and come back when he'd shown the math that proved it.

    I've worked with people quite a few rungs above me. All of them are capable of writing documentation that I can understand. Documentation that cuts the time spent on my comprehension of how their code works by 90+%.

    Their job isn't just whipping out code. It's also showing that it works. How it works. The upside is I don't ask them nearly as many "stupid" questions, because while their code still in Klingon - but it comes with subtitles. It also means that once I've looked through this Klingon enough times, it'll start making sense. I might actually learn how to write some stuff in Klingon by reading what they've written (with subtitles). But in the software industry that's just a waste of time - who needs people actually learning stuff at work?

    Imagine the following scenario:
    Ed is a brilliant engineer and architect. He comes up with a way for us to build a trans atlantic maglev train route that runs under water in essentially vacuum tubes. He's even figured out how to make it cheap enough that trip from London to New York city will cost you 200$ and will take about two hours. His brilliance even allows the project to scale, so that if we swing the tubes upwards and really punch it, we can send stuff into orbit for a price of 1000$/ton.

    Now, instead of writing up the designs, specs for the materials, how to build the materials and so on, Ed's just going to tell the people involved how to do it by phone. Because of Ed's absurd brilliance and genius, this actually works for a full week, and his super human skills in JIT means we're now 12 miles into this tunnel.

    The 8th day however, Ed's rather unfortunate. Seems he decided to drop by the post office the same day that Dan the mail man went postal and killed everyone in the office. Including Ed.

    Since noone else knows how any of this stuff is supposed to work, we now have to give up on the tunnel project while we siphon through the few things Ed actually left behind.

    Now, in the real world Ed's demise would be somewhat of a setback, as we've now lost the lead engineer on the project. However, since Ed was a good engineer and architect, he knew he was supposed to put all these things down on paper before we started the project and put billions or trillions of dollars on the line.

    Now, in the software industry, we're very fond of calling ourselves engineers and architects. Unfortunately most of us (even in companies) really don't reach that level of excellence - we don't document what we do, either because we're too lazy or because the companies don't want to spend money doing that. That's fine - just don't consider yourself or what you're doing software engineering.

    I've actually had the pleasure of working for an engineering company as the only software developer/programmer. Imagine how flabbergasted I was when my boss asked me for documentation as well as actual schematics for the software I was making. Schematics as in fl

  • by 91degrees (207121) on Monday March 16, 2009 @01:53PM (#27213527) Journal
    The article was about someone who can do an incredible amount of coding in a very short time. Indeed, more coding in less time than most anyone else.

    Because all he was doing was writing code. He took an hour to solve a problem that took the team 2 days. "The team" must have been at least 3 people. So that's occupying 6 programmer days. 40-45 hours. It would have taken him less than an hour to document or explain what the solution was. Is he really worth 40-45 times as much as the other programmers?

    If the guy produces a lot of unmaintainable code then he's costing almost as much as he's making for the company. His personality problems will increase staff turnover, and he will eventually leave. Nobody lasts forever. When he leaves everything that he wrote will have to be documented or replaced at considerable cost.

    Most programmers will be able to do most tasks. There are some highly specialised tasks that will require an expert in that area, but you can always find the appropriate expert. Anything else can be learned. You'll lose a developer for a few days while he learns but you'll gain a developer with extra knowledge, and the half decent ones will be happy to stick with a company that allows them to develop.
  • by computational super (740265) on Monday March 16, 2009 @01:55PM (#27213557)

    How do you know you're not a jackass?

  • by squoozer (730327) on Monday March 16, 2009 @02:02PM (#27213701)

    It's true enough that requirements often change during development but if requirements are changing drastically and / or frequently during development someone has screwed up.

    Like a lot of young developers I used to think that the "best" way to deliver a software project was just to jump straight in and start coding. As I've gained more experience and worked on progressively larger projects I now see that software development is a lot more like civil engineering.

    If you went up to the engineering team that was responsible for the Viaduc de Millau bridge and said I want it in purple half way through the project he would say fine give me X euros to re-paint it. If you instead said I want it to span the straights of Gibraltar you've screwed up with the requirements.

  • by noidentity (188756) on Monday March 16, 2009 @02:03PM (#27213739)
    Being required to produce documentation along with one's design often makes the design better, because it encourages a design that's easy to document clearly. The lazyness principle makes you want to avoid tricky things that take pages of documentation. So even if you think the documentation as some crap that goes along with your nice, clean design, think of the task of writing it as no different than a unit test or other QA; it exposes problems you might not otherwise notice.
  • by JoeMerchant (803320) on Monday March 16, 2009 @02:08PM (#27213831)

    Sometimes you need a solution NOW, and you will have time to clean it up (or re-implement it more carefully) later.

    Except, cleanup (or re-implementation) never happens.

    I'd say that's clearly not Josh's fault. If you hire a team of paratroopers to build you a bridge, then you try to use it as critical infrastructure for the next 30 years, it won't be the paratroopers' fault when a bunch of trucks fall in the river.

  • by CrazedSanity (872448) on Monday March 16, 2009 @02:16PM (#27213993) Homepage Journal

    I absolutely agree!

    At one point at a previous job, I was tasked with putting all of our projects into our project management software and prioritize. I built a tree structure with parent projects and sub-projects, where the furthest-out projects needed to be completed before the parent project could be done (so the root projects were easy to understand for the PHB, and we could deal with the smaller bits). Each level was prioritized based on the level, so I could tell what piece should be completed first (I worked that part out with the other developers so we all understood what it all meant, along with figuring out some of the lower-level priorities and building best-guess timelines).

    After a week of prioritizing, arranging, and setting timelines, I brought it to the PHB. I explained the logic of the thing and how much I'd worked with the other developers in order to get it organized as such. He gave me a blank stare, asking why there were so many sub-projects and why he couldn't find the project he was looking for, etc. I explained the organizational logic, and he just gave me that blank, glazed-eye stares and then excused me.

    The next morning I was called into his office, where he showed me (with a huge smile on his face) how he'd rearranged everything. There were no trees (projects with sub-projects) that explained dependencies. The timelines were changed to what he wanted them to be, causing 10-12 projects to overlap on very tight timeframes. EVERYTHING was a project (the sub-projects that were dependencies of parents were suddenly their own projects with incredibly low priority). Only the projects he was interested were prioritized, and there were dozens of projects set with the highest priority.

    The funniest thing? Some time later we had a meeting where he told us adamantly that we should only EVER have ONE priority 1 (highest priority) item and that we shouldn't work on anything else until that priority 1 project was done. *sigh*

  • by Jherico (39763) * <bdavis@sai[ ]ndreas.org ['nta' in gap]> on Monday March 16, 2009 @02:31PM (#27214235) Homepage

    Agreed; which is why this statement from TFS: "... could churn out code that saved the company millions" - is nonsense. It may look that way on the surface, but when accounting for all the code maintenance pains that inevitably follow, I've yet to see a single such "genius" that wasn't a net loss

    First off, sometimes there is a time basis that cannot be avoided and a solution, however dirty, is required right away in order to complete a contract or open a web storefront or the like. In these cases the original statement is literally true, millions could be made or lost depending on whether you can flip the on switch tommorow or next week. At that point, you're making money in the long term, regardless of whether you have to reimplement.

    Of course companies are usually, imo, too focused on the here and now results anyway and this is a double edged sword. It can get you to market quicker, but I have time and again seen companies shopping for development libraries or other similar tools go against the selection of one vendor by EVERYONE who was going to use the product and go for another cheaper vendor that no one liked because it saves the company 100k right now, even though the cost of developing locally all the missing functionality from the cheaper solution will easily end up costing more that the saved amount in the long run.

  • by Venik (915777) on Monday March 16, 2009 @02:37PM (#27214313)

    Interesting analogy but probably not applicable here. You see, Ed didn't just tell you how to build the train - he actually built it for you. True, the bastard hasn't produced a shred of documentation. And now, after Ed's unfortunate demise, you and your team of average engineers are scratching your heads, trying to figure out how to extend Ed's design to reach Tokyo. And the reason for your predicament is obvious: as an employer you lucked out to have a truly talented engineer working for you, but, being an idiot, you made no attempt to understand his work. You should have been searching high and low for engineers capable of understanding Ed's designs and working with him, however difficult that might have been. Instead, you hired some random guys off the street and let Ed work alone. Your problem is entirely your own fault.

    Talented people get bored easily. Something that you or I may find intellectually stimulating, they find obvious and prosaic. This is why, as the fortunate employer of a genius, you hire capable people who complement his talents. You don't make soup out of your goose just because he won't document the process of laying golden eggs.

  • by cromar (1103585) on Monday March 16, 2009 @02:38PM (#27214327)
    Overall, I agree, but saying that

    It's *all* pointless. The end is the same for everybody. We're all worm food.

    is a great plan for mediocrity. There are struggles that are worthwhile and there are struggles that are pointless, but to say that no struggle matters is speaking from both ignorance and arrogance. I mean no offense, we're all ignorant and arrogant to some extent.

    Doesn't matter if you rage against the machine or oil its gears. In a hundred years, I promise you it won't matter one whit.

    I'll just give a few examples of why this isn't true: Martin Luther King, the Buddha, Jesus, Krishna (to whatever extent those last three are flesh and blood historical figures), Ghandi, US founding fathers, those who participated in the Tiannamen square [youtube.com] incident, etc., etc., etc. All these people have had and will continue to affect life for humanity.

    So, while it's true that blindly "raging against the machine" is pointless, just because you have

    a nice job, nice house, nice car and a family I love dearly

    doesn't make you any better or make your life more worthwhile or valueable than someone who can't afford to fix their teeth. Their pain and alienation may be far more meaningful than your "boring life" (your words).

    All I'm saying is you don't have to settle for assimilation, blind hate, futility, alienation, mediocrity or ambivalence or comfort. Do something with your life and make the world a better place, but don't "sell out" or become so bitter that you are divorced from the world. It's not worth it for you or anybody else.

  • by Kell Bengal (711123) on Monday March 16, 2009 @02:51PM (#27214569)
    Well, he was employee #1 in the company (both value-wise and chornologically). Since he didn't document at the start, it was much cheaper to hire a post-grad than spare a more expensive employee from actual development work.

    I wonder if perhaps there's an argument for pairing senior employees who do the critical design work with fresh hires to document the what and why of it. That way, the higher-up engineers don't have to write anything down and the junior engineers get to absorb some of their insight by osmosis.

  • by Thaelon (250687) on Monday March 16, 2009 @03:27PM (#27215133)

    If my code takes more than a few seconds to figure out what it's doing, I keep rewriting it until it doesn't.

    The golden rule of being a truly good programmer is that you write the code that you wish other people would write.

    Write code that is self documenting by way of having very descriptive method and variable names. Because while it may take longer to write, you'll invariably end up reading it ten or more times more than you'll be editing it. So time making it readable & grokable is time exceedingly well spent.

    here's a quick sample of what I mean:

    Instead of this

            if (thingy.getLastTime() - thingy.getCurrentTime() > 3600000) {
                    for (Calendar calendar : thingy.getCalendars()) {
                            calendar.update();
                    }
            }

    write this:

            boolean isChangingTimeZones = hasChangedTimezones(thingy);
            if (hasChangedTimeZones) {
                    updateAllCalendarsWithTimeZoneChange(thingy);
            }

    It's debugger friendly (you can see the value of isChangingTimeZones before you step into the if-then-block, and the logic is described via the variable name which has a corresponding method that can be easily unit tested) and all the names are descriptive as hell. Not to mention the action is decoupled from the logic, and could easily be more so. Either could be modified without ever changing that particular code. And yes, the boolean method should use a constant for that magic number. You can probably infer what that constant is, but why waste people's time making them infer when you can very easily tell them explicitly?

  • by sfcat (872532) on Monday March 16, 2009 @03:38PM (#27215277)
    And you are 100% wrong. Programming is 100% technical. Working in an IT shop might be 40% technical and 60% people but that just means that you only spend 40% of your time programming.

    But here's the catch, someone must write the code in the end. And someone must maintain it. And the code that is written is of varying quality. If someone is simply a better, faster programming, then their code will be cheaper to maintain because it will break less often and scale better (or whatever your metrics are for code quality). I find that a "nice programmer" might be easier to work with, but those Saturday night production outages make me hate that person all the same. And I'm much more likely to fire him at that point because I think that person is unlikely to have to skills to keep the code working.

    Finally, I think this entire argument is a bit of a crutch. I've seen people who match Josh's description, but usually the best programmers are just crabby because so much of the work falls to them. Then they get painted with "Josh's brush" and labeled as having bad people skills. When really, they are just tired and overworked. If the people with "people skills" had to deal with even 1/10th the work, they would go on a killing spree within a week.

  • by stephanruby (542433) on Tuesday March 17, 2009 @12:01AM (#27221295)

    If you really want to do the armpit-to-armpit teamwork go back to Yourdon's original structured programming team. You had a senior guy, a junior guy, and a librarian. Today that would be senior guy, junior guy, and documentor. It works in threes, but not in twos for some reason. I think it has something to do with allowing intelligent people to lead design, rather than have to check around to see if what they're doing is ok. In pair programming you have no leader. With no leader you have no direction and thus no progress.

    That metaphor can be extended to a surgical team, where you have one chief surgeon and everyone else around the table has a specific role and is there to assist. Or it could be extended to the 911 phone operators, where there is usually one operator on the line, and two assistant operators listening on the call and following the directions of the first (although, that part is almost never shown on 911 reenactments).

    Personally, I have no problem letting another programmer take the lead when pair programming, my only two requirements are that we set some time aside for debriefing each other beforehand (so that I know where we're going) and that we set some time aside for debriefing each other afterward. I don't usually interrupt (unless I have to), and besides I take copious notes when sitting shotgun -- this is a trick I use to keep on paying attention -- while keeping the things I say to a bare minimum until later.

    I find it also helps to let the person typing make their own mistakes, a typo, or what have you. Usually the guy typing will correct himself without interruption needed on my part. So if I see an error, I take a quick note of it in the margin of my notebook, and it's only after 10 or 15 seconds or just before the compile cycle, that I'll point out the errors as tactfully as I can.

    That being said, that Amadeus reference you cited scares me. Most programmers are not at the Amadeus-level, and yet most programmers think that they are. And I can't tell you the number of times I've had to stop a fellow programmer from coding because he had no clue where he was going, and no clue on how to get there, he just wanted to make himself feel better by coding something -- anything -- right away.

I am the wandering glitch -- catch me if you can.

Working...