Where do they FIND these people?



  • A company I used to work for, back before I moved halfway across the country, has lost 3 of their most senior techs to either illness or retirement in the last year. The one guy left doesn't have too much time after doing the work of 4 people to train new hires in the wonders of legacy systems. (This place still has three PDP-8 systems in production...)



    I'm not interested in moving again, so they can't have me. But they offered me some cold, hard cash to do one more service for them: Could I go through some resumes, do some phone calls, and screen out the hopeless causes?



    Finding a guy in his twenties who's got a few years experience writing in FORTRAN-77 is hard enough. Try finding any who can do FORTRAN-IV, which is STILL in use there on a few systems. They've finally, at least, realized the wisdom of why I quit: refusal to upgrade is a stupid idea after a while. So they'll be satisfied with someone who at least has enough of a clue about older systems to be able to learn to read FORTRAN-IV and FORTRAN-77 and COBOL and MUMPS and ... Well, I used to have a line: "This is not a heterogenous environment. This is an orgy." when referring to their total setup.



    If they can code C and C++ then they're of potential use for migration. Same if they can do Perl and Java. Now, testing for that stuff isn't hard. Testing someone's ability to get a grip on FORTRAN-IV is a little trickier, especially when you deal with people who respond to "Define MS-DOS" with "Oh, that's that command thingie in Windows 95!" because they're too young to remember ... actual DOS.



    I come up with a short questionnaire. Anyone smart enough to find this stuff in Google fast enough to "fake it" while talking to me is actually not going to do too badly here, but I did have to include a few questions that couldn't be so rapidly, easily figured out. Not many people got "What is a VT102?" wrong, but I could tell which ones were on a dialup connection...



    Here is a sampling of questions, and some of the answers I was given.



    Question: What is the difference between FORTRAN-IV and FORTRAN-77?

    Answer: FORTRAN-77 is 74 more than FORTRAN-IV.

    Answer: FORTRAN-77 is ... Could you use that in a sentence, please?

    Answer: FORTRAN-IV has one less digit in it.

    Answer: FORTRAN-IV must be older. DUUUUUH



    Question: What is the difference between FORTRAN and COBOL?

    Answer: The name.

    Answer: One has a longer name.

    Answer: COBOL comes before FORTRAN alphabetically.

    Answer: FORTRAN has two different vowels and COBOL only has one.

    (At this point I wonder if these people think these are trick questions...)



    Question: What does IBM stand for? (This was to catch people thinking it was a trick question.)

    Answer: IBM stands for something?

    Answer: What's IBM? WHAT?!?!



    Question: Who invented FORTRAN, IBM or Microsoft? (Obvious idiot weed-out here...)

    Answer: NEITHER! IT WAS BANK OF AMERICA!

    Answer: That's a trick question, FORTRAN is what the C Preprocessor used to be called, so it was invented by Kernighan and Bitchie. At this point, I couldn't actually not laugh at the guy. Especially since he said Bitchie...



    And the last question... Honestly? I expected virtually all applicants to say "I have no idea" or similar to this question: The company still has two machines in service that use One's Complement math. I.e., a byte will not range from -128 to +127, it will range from -127 to +127, and there are TWO values for zero: all zeros in binary is positive zero, all ones in binary is ... negative zero. (Google it. Much amusement to be found on that subject online.)



    I was surprised. A few people were at least aware of it though they didn't grasp a few of the reasons this drives programmers insane. But one guy's answer took the cake. Now, this guy had already basically gotten himself crossed off the list by this point: He was the "IBM stands for something?", and he was the "Could you use that in a sentence, please?"... He even got most of the deliberate softball lobbed over the plate questions wrong. "What is the difference between a microcomputer and a supercomputer?" He responds that a microcomputer is anything running Windows, and a supercomputer is anything not running Windows. (Suck-up...)



    Question: What is the difference between one's complement and two's complement?

    Answer: One's complement is when you compliment a hot chick by herself at a bar. Two's complement is when you compliment her to her husband at the restaurant.



    At least I got paid well for this...



  •  

    Answer: FORTRAN-77 is 74 more than FORTRAN-IV.

    At least you know they're good at math.

     I don't really see questions like "What is the difference between FORTRAN and COBOL?" as relevent to anything though.  You might as well have asked them what the difference between apples and oranges is.

    (disclaimer:  I  don't know any COBOL, only did the minimal FORTRAN in college.  If they're distant cousins or something, maybe I'm wrong)



  • @vt_mruhlin said:

     

    Answer: FORTRAN-77 is 74 more than FORTRAN-IV.

    At least you know they're good at math.

     I don't really see questions like "What is the difference between FORTRAN and COBOL?" as relevent to anything though.  You might as well have asked them what the difference between apples and oranges is.

    (disclaimer:  I  don't know any COBOL, only did the minimal FORTRAN in college.  If they're distant cousins or something, maybe I'm wrong)

    The main point in a lot of these questions, actually, was to weed out the dangerously stupid: the people who didn't know anything about either language but would pretend to if it got them a paycheck. The company knew full well that finding people who actually could read FORTRAN-IV and aren't in their 50s or 60s is asking a miracle, and was quite willing to provide books, training, and time... to any candidate who was willing to be honest and admit they'd need it. A candidate who answered with "Um... Honestly, I don't know anything about either language", but who tested well on other areas, is a good potential hire when we're expecting we probably have to train them, but a candidate who tries to fake his way through will break code in horrible ways, guaranteed.



  • @vt_mruhlin said:

     

    Answer: FORTRAN-77 is 74 more than FORTRAN-IV.

    At least you know they're good at math.

     I don't really see questions like "What is the difference between FORTRAN and COBOL?" as relevent to anything though.  You might as well have asked them what the difference between apples and oranges is.

    (disclaimer:  I  don't know any COBOL, only did the minimal FORTRAN in college.  If they're distant cousins or something, maybe I'm wrong)

     

    Dude... IV means 4, the difference is 73, not 74.



  • And "Best of the Sidebar" goes to... 



  • @vt_mruhlin said:

    I don't really see questions like "What is the difference between FORTRAN and COBOL?" as relevent to anything though.  You might as well have asked them what the difference between apples and oranges is.

    Well I've never even seen (never mind written) any FORmula TRANslator or COmmon Business-Oriented Language code, but I at least know what they stand for, and the problem domains they tend(ed) to be used in.

    If I was interviewing someone, I'd certainly have more respect for a candidate with at least that kind of minimal knowledge than people who bluster and BS their way through questions like that.



  • @Wolftaur said:

    At this point, I couldn't actually not laugh at the guy. Especially since he
    said Bitchie...

    Me too, thanks. 8D

    @Wolftaur said:

    The company still has two machines in service that use One's Complement math.

    Calculating IP checksums on those must be a bliss.



  • @Gabelstaplerfahrer said:

    @vt_mruhlin said:

     

    Answer: FORTRAN-77 is 74 more than FORTRAN-IV.

    At least you know they're good at math.

     I don't really see questions like "What is the difference between FORTRAN and COBOL?" as relevent to anything though.  You might as well have asked them what the difference between apples and oranges is.

    (disclaimer:  I  don't know any COBOL, only did the minimal FORTRAN in college.  If they're distant cousins or something, maybe I'm wrong)

     

    Dude... IV means 4, the difference is 73, not 74.

    So, ever heard of the concept of “irony”?



  • @Wolftaur said:

    \
    Answer: FORTRAN-IV must be older. DUUUUUH
     

    how is that a "Duhhh"?   FORTRAN-IV does not in any way imply the year that it was made.  FORTRAN-77 does not in any way imply the version number. 

     

    Here try this one:

    Which is older?  Windows 7 or Windows 2008?

     

    Also, if you asked me those questions I would have probably given you those kind of answers too because I would have realized that I most certainly did NOT want to work there.



  •  @Gabelstaplerfahrer said:

    Dude... IV means 4, the difference is 73, not 74.

    You must be using one of those obsolete web browsers that doesn't intuitively insert <sarcasm> tags where appropriate.



  • @AssimilatedByBorg said:

    Well I've never even seen (never mind written) any FORmula TRANslator or COmmon Business-Oriented Language code, but I at least know what they stand for, and the problem domains they tend(ed) to be used in.

     

    What would you say to a candidate who thought JavaScript had something to do with Java?  Names don't mean a lot, especially when it comes to 50 year old dead languages.



  • @Wolftaur said:

    The company knew full well that finding people who actually could read FORTRAN-IV and aren't in their 50s or 60s is asking a miracle...

    So is there anything wrong with hiring people in their 50's and 60's for a job like this?  Or were they just not even getting them as applicants.

    As a Cobol programmer who is 40-something (ahem), I'm in the same boat, but my company has simply not even tried.  We support VMS, on Alpha and Integrity platforms with the occasional hard-core-budget-strapped Vax customer.  I don't think they've tried in over a decade, but I have a feeling me and my coworkers are irreplaceable.  It's not that the young punks don't know it, I have a feeling none of them WANT to learn it.  



  • @jetcitywoman said:

    It's not that the young punks don't know it, I have a feeling none of them WANT to learn it. 

    Half the people who know it don't want to admit it. Those of us who will admit it are frequently too valuable in other capacities.

    We're starting to hit an interesting time where the world isn't all on a PC anymore, and we need people who understand centralised processing. That's all the big iron used to do, so those of us with solid VMS and MVS experience are suddenly very much in demand, making excellent money - and doing work that is ultimately quite easy. It's just that nobody else can do it. I'd have no problem with working in COBOL again, but you have to admit, it's far more interesting to work on this whole cloud thing.



  • I agree with you. 

    .... so use Cobol on the cloud...

    Seriously, I've done some amazing/stupid things in Cobol.  Some years ago I wrote a Cobol program that read the database and output webpages - sort of an old-fogey's version of Ajax, I guess.  I've also done Cobol programs that sent/received data in XML format through interfaces to other software packages.

    And you're 100% correct about the pendulum swing effect in technology.  When thin client workstations came out alot of my coworkers couldn't help but point out that they were little more than dumb terminals with a pretty name.



  •  @vt_mruhlin said:

     @Gabelstaplerfahrer said:

    Dude... IV means 4, the difference is 73, not 74.

    You must be using one of those obsolete web browsers that doesn't intuitively insert <sarcasm> tags where appropriate.

    I will admit that sarcasm and irony can be hard to detect. I almost feel stupid. Almost.



  • @vt_mruhlin said:

    What would you say to a candidate who thought JavaScript had something to do with Java?  Names don't mean a lot, especially when it comes to 50 year old dead languages.

    I would tell them to watch their step :)

    I'm quite fond of joking that the only thing JavaScript and Java have in common is the first 4 letters of their names!  Just try "declaring" an object in javascript if all you know is java objects....

    Well they do share one other thing: the stamp of marketing, from people who thought that Oak and LiveScript weren't sufficiently cool names. (Though I am still somewhat joking here, at least about Oak, see what James apparently wrote about it.)

    Edit: I never said I knew much about fortran & cobol, just that I know more than zero. :)



  •  @vt_mruhlin said:

     I don't really see questions like "What is the difference between FORTRAN and COBOL?" as relevent to anything though.  You might as well have asked them what the difference between apples and oranges is.

     

    It's a basic weed-out question.  IMHO, it has no right answer, but it has numerous wrong answers which instantly identify a candidate as unfamiliar with the subject.  For example, I might say, "FORTRAN is still popular in communities like Physics researchers because it doesn't have the aliasing problems inherent in C, which makes some types of compiler optimisations possible in FORTRAN code that are impossible in some other languages.  COBOL, however is most popular in business related software development, and is generally considered most suited to things like processing business related data."  If you interviewed a fleet of guys, you might not get the exact same answer.  Some may talk about the age of the languages, or the syntax, or the availability of libraries, or any of a zillion other differences.

     But, when you get the guy that says, "COBOL is what Microsoft used to write all of Windows 2000," or "FORTRANs are only manufactured by Sun, so using that type of hardware means being locked into a single vendor," or "both are inferior to UNIX," or "I'm not familiar with COBOL because I haven't had a chance to use it yet because I haven't upgraded to Windows Vista,"  well then you know you are dealing with somebody who just isn't well enough versed in the topic to discuss it at all.

     

    Questions like that are only marginally a chance for a candidate to impress the interviewer, and much more an opportunity for them to avoid horrible failure. 



  • @jetcitywoman said:

    .... so use Cobol on the cloud...

    Ooh! Managed COBOL.NET - that would be fun. ;)




  • I know new fortran (i.e. Fortran 90), but:


    Question: What is the difference between FORTRAN-IV and FORTRAN-77?

    FORTRAN 77 has a block IF statement. This means that I could  program in FORTRAN 77 if I used CONTINUE instead of END DO and put six spaces in front of every line (and also didn't use recursion and all the other stuff Fortran 90 added), but I would not want to use FORTRAN IV to write stuff in.


    Question: What is the difference between FORTRAN and COBOL?

    FORTRAN becam Fortran 90, which is a language I might actually use. COBOL still sucks.

    Question: What does IBM stand for? (This was to catch people thinking it was a trick question.)

    International Business Machines.


    Question: Who invented FORTRAN, IBM or Microsoft? (Obvious idiot weed-out here...)

    John Backus at IBM

    Question: What is the difference between one's complement and two's complement?

    Ones compliment has a negative zero separate from positive zero. The binary representation of a negative number in two's compliment is one more than its representation on two's compliment.





     



  • @tster said:

    @Wolftaur said:

    <br>
    Answer: FORTRAN-IV must be older. DUUUUUH
     

    how is that a "Duhhh"?   FORTRAN-IV does not in any way imply the year that it was made.  FORTRAN-77 does not in any way imply the version number. 

     

    Here try this one:

    Which is older?  Windows 7 or Windows 2008?

     

    Also, if you asked me those questions I would have probably given you those kind of answers too because I would have realized that I most certainly did NOT want to work there.

    You're absolutely right, of course. My duh on the original post wasn't that he guessed the obvious: but that he guessed, making a numerical comparison without knowing what the meaning of the numbers were. FORTRAN-IV could have quite easily been FORTRAN-2004, after all. The duh is that his thought process was invalid, and he managed to, even while getting the right answer, make it obvious he was just guessing blindly...



  • @jetcitywoman said:

    So is there anything wrong with hiring people in their 50's and 60's for a job like this?  Or were they just not even getting them as applicants.

    We did get a few, but the company had been correctly convinced of two things:

    1. This isn't 1970, and a 1970 approach to the problem (and a 1970 budget and team size) is a guaranteed failure.
    2. They're great about not age discriminating but we've found that a lot of the old COBOL guys were at places that never wanted to upgrade... this often means they never had any need to learn other, newer things. Since the company is in a big push to migrate to modern hardware and software, rather than find people to maintain FORTRAN for another 15 years... a guy who's done the last 30 years in legacy code at a company who isn't interested in fixing what isn't broken isn't actually going to fit this position. They won't have to write FORTRAN or COBOL, just understand it well enough to port it.



      Oh, and if they can look at 200 pages of "ADD T TO F GIVING D" and "MULTIPLY T BY B GIVING C" without clawing their eyes out, it'll help. :)


  • @wrosecrans said:

    It's a basic weed-out question.  IMHO, it has no right answer, but it has numerous wrong answers which instantly identify a candidate as unfamiliar with the subject.  For example, I might say, "FORTRAN is still popular in communities like Physics researchers because it doesn't have the aliasing problems inherent in C, which makes some types of compiler optimisations possible in FORTRAN code that are impossible in some other languages.  COBOL, however is most popular in business related software development, and is generally considered most suited to things like processing business related data."  If you interviewed a fleet of guys, you might not get the exact same answer.  Some may talk about the age of the languages, or the syntax, or the availability of libraries, or any of a zillion other differences.

    SNIP

    Questions like that are only marginally a chance for a candidate to impress the interviewer, and much more an opportunity for them to avoid horrible failure. 

    Maybe this depends on the interviewer, but... Those questions were more than a marginal chance for someone to impress me. There is one right answer that most people won't give, but those who will give it are worth serious consideration: "Honestly, I don't know anything about FORTRAN and COBOL." The guy who will honestly admit his limits and not try to fudge his way through is someone we can try to teach, and, someone we don't have to be nearly as afraid of-- he's much less likely to stick his hands in what he doesn't understand rather than admit he needs help. And in my experience, the ones who think they can tinker and guess their way to fixing problems are always the most damaging.

    Judging from some of the articles on TDWTF in the first place about the damage done by those who claim knowledge they don't have, I'd guess I'm probably not the only one who's more impressed by "I don't know, to be honest" than someone who correctly guesses all the answers. :)



  • @Wolftaur said:

    Maybe this depends on the interviewer, but... Those questions were more than a marginal chance for someone to impress me. There is one right answer that most people won't give, but those who will give it are worth serious consideration: "Honestly, I don't know anything about FORTRAN and COBOL." The guy who will honestly admit his limits and not try to fudge his way through is someone we can try to teach, and, someone we don't have to be nearly as afraid of-- he's much less likely to stick his hands in what he doesn't understand rather than admit he needs help.

     

    I guess you wouldn't have hired me because I knew most of the answers...



  • @tster said:

    @Wolftaur said:

    Maybe this depends on the interviewer, but... Those questions were more than a marginal chance for someone to impress me. There is one right answer that most people won't give, but those who will give it are worth serious consideration: "Honestly, I don't know anything about FORTRAN and COBOL." The guy who will honestly admit his limits and not try to fudge his way through is someone we can try to teach, and, someone we don't have to be nearly as afraid of-- he's much less likely to stick his hands in what he doesn't understand rather than admit he needs help.

     

    I guess you wouldn't have hired me because I knew most of the answers...

    Knowing the answers would put you ahead of someone who didn't. Not knowing the answers but admitting it would put you ahead of someone I could tell was guessing. Me saying 'one right answer' wasn't meant to imply that was the only right answer, just that it was one of the acceptable answers. :)



  • I was just forwarded a "late entry" for screening. They could have screened this one themselves if they'd read the cover letter... it talks about his 35 years of experience in Fortran and "Cobalt." Worse still is that he put in 35 years of work history, but ... he lists his birth year as 1983. Yay for skillful resume padding thwarted by inability to pass grade school math...



  • @Wolftaur said:

    I was just forwarded a "late entry" for screening. They could have screened this one themselves if they'd read the cover letter... it talks about his 35 years of experience in Fortran and "Cobalt." Worse still is that he put in 35 years of work history, but ... he lists his birth year as 1983. Yay for skillful resume padding thwarted by inability to pass grade school math...
     

    Perhaps he was coding when he was still and egg in mother's overies.



  • @tster said:

    Perhaps he was coding when he was still and egg in mother's overies.

    Either that, or COBALT is the object-oriented version of COBOL that they end up using to control the first time machine? :)



  • @tster said:

    Perhaps he was coding when he was still and egg in mother's overies.
     

    Real programmers can output code when their total existence consists of a vague neuron impulse pattern with the meaning "well, maybe I want children some day" in either of their future parent's minds.



  • @dhromed said:

    @tster said:

    Perhaps he was coding when he was still and egg in mother's overies.
     

    Real programmers can output code when their total existence consists of a vague neuron impulse pattern with the meaning "well, maybe I want children some day" in either of their future parent's minds.

    Pfff. A real programmer can output code even if he was a blowjob.



  • Real programmers can output code when their ancestors as a collective species has yet to evolve sexual reproduction.

    I'm going to cut this short before it turns into a 1-up fest of "Real Programmers"

     

    Real programmers can output code when the matter they're composed of is nothing more than a potential Higgs-field collapse, 10^-35 seconds before Inflation.

    There.

    1-up that, yo.



  • @dhromed said:

    Real programmers can output code when their ancestors as a collective species has yet to evolve sexual reproduction.

    I'm going to cut this short before it turns into a 1-up fest of "Real Programmers"

     

    Real programmers can output code when the matter they're composed of is nothing more than a potential Higgs-field collapse, 10^-35 seconds before Inflation.

    There.

    1-up that, yo.

    Real programmers can output code before God existed to invent the rules of physics. :)



  • @dhromed said:

    Real programmers can output code when the matter they're composed of is nothing more than a potential Higgs-field collapse, 10^-35 seconds before Inflation.

    There.

    1-up that, yo.

     

    Real programmers can output code 10-35+1 seconds before Inflation!



  • @Wolftaur said:

    Real programmers can output code before God existed to invent the rules of physics. :)
     

    I can now output code before god exists, so thats about a ∞-down.



  • @dtech said:

    @Wolftaur said:

    Real programmers can output code before God existed to invent the rules of physics. :)
     

    I can now output code before god exists, so thats about a ∞-down.

    That's a relief. Winning by invoking God would have made me a Republican!



  •  Real programmers 42.



  • @dhromed said:

     Real programmers 42.

    Aww, and you were doing so well just a few minutes ago...

    Real programmers 101010(2).



  • @Wolftaur said:

    Real programmers 101010(2).

    Wait, real programmers (84+File_Not_Found)?



  • @dtech said:

    @Wolftaur said:

    Real programmers 101010(2).

    Wait, real programmers (84+File_Not_Found)?

    Old-days notation: Number(Base). Frequently seen in printed documentation, especially from typesetters which couldn't do subscript, the other common method in stuff I read back then. 101010 is the binary representation of decimal 42.

    A real programmer would have recognized 101010 without even knowing the base. :)



  •  But the base, in any base is always 10, so 101010(10) is correct, be consistent.



  • @Ilya Ehrenburg said:

     But the base, in any base is always 10, so 101010(10) is correct, be consistent.

    Number(BaseOfNumberInBase10), happy? :P


Log in to reply