• Neo (unregistered)

    Hmm, deja-vu ... Must be a glitch in the Matrix...

  • Anonymous (unregistered)

    It boggles the mind.

  • Fregas (unregistered)

    I love it!  What a great customer WTF!

  • Martin (unregistered) in reply to Neo

    I also have strong feeling I saw it somewhere...

  • Nooto (unregistered)

    Hmmz strange..
    i swear i've read this one before...

     

  • (cs)

    ROFLMAO! It never ceases to amaze me how so pany people can fail to comprehend the order in which instructions are executed. I also love how even though all data points to one thing (working), they still let their gut rule their mind! Stupid stupid users lol...

  • Neo (unregistered)
  • JoeyLemur (unregistered)

    I'm reminded of the #BLOAT directive in PowerBasic...

  • Brian (unregistered) in reply to Neo

    The real WTF... and it was barely over a month ago. I think we need an apology from the webmaster ala Slashdot when a duplicate story is posted!

  • (cs)

    Holy hell, what are the chances this would happen to Jeremy twice???

  • (cs)
  • (cs) in reply to Manni

    <font face="Verdana"><font size="2">'Twas a "side bar" post that made its way to the front page today.



    </font></font>

  • (cs)
    Alex Papadimoulis:
    The customer complained that the analysis was "way too fast" and "must be broken".  They were especially concerned because the analysis appeared to take the same amount of time no matter how much data was collected.  I tried explaining the changes I had made but they were unshakeable in their belief that the analysis "couldn't possibly be that fast."  They couldn't seem to understand that the analysis was taking place WHILE the scan was going on.


    It can be a problem being competent.

    Six months later, they complained that the analysis was taking too long and could we speed it up.  I cut the fake delay in half and they were overjoyed.  It was not my decision to charge them $5,000 for "optimization and refinement of algorithms".


    Maybe, it should have been.

    Sincerely,

    Gene Wirchenko

  • (cs) in reply to pbounaix
    pbounaix:

    Hey Farva what's the name of that restaurant you like with all the goofy shit on the walls and the mozzarella sticks?

  • Jeremy H (unregistered)

    Explanation for why this appeared twice:  I submitted it for the daily WTF a couple months ago and when it didn't make it in a few days, I assumed it had been rejected and I posted it to the sidebar.  In the interim, I guess the WTFmaster decided it was worth the front page, or else maybe there's just a big backlog?

  • (cs)

    It's called "Premature Optimization".  Optimizing your algorithm before the user understands the shortness of the real time needed.

  • (cs) in reply to Neo

    Anonymous:

    Nice, a promoted sidebar!! So, if you post to the sidebar and you get enough replies do you get shoved to the frontpage like digg??!

  • Trent (unregistered)

    Reminds me of an old rule of programming: always put a sleep(5000); statement in your code when you show the first demo, and then take it out before the next demo to show what an awesome job you did optimizing it.

  • chaos_engineer (unregistered)

    That's not a very elegant solution. If you use a hard-coded delay, then the delay will stay constant even if you upgrade to a faster processor.

    Assuming that we don't know the number of scan lines in advance, i'd have a "realtime_percent" configuration parameter. At 50% we'd process one line of data for every two lines read. At the end of the scan half the data would be processed, effectively trimming 50% off the original processing time.

    The default would be 100% for maximum speed, so 0% would emulate the program's original behavior.
    <font size="3">  
    </font>

  • Hrishikesh (unregistered)

    This isn't limited to software...

    In fact, I had read somewhere that those making vacuum cleaners can tone down the motor drone quite a bit, but won't do so because their preliminary customer reviews showed that the users felt that the motor was underpowered because of the lower sound!

    So there, that's market dictating the law. And no harm so long as you can profit from it!

  • (cs) in reply to mrsticks1982

    I notice nobody's actually talking about the WTF in question...

  • (cs) in reply to Martin
    Anonymous:
    I also have strong feeling I saw it somewhere...

    And I'll miss you most of all, Scarecrow!
  • (cs) in reply to Brian
    Anonymous:
    I think we need an apology from the webmaster ala Slashdot when a duplicate story is posted!


    Excuse me, but in what alternate reality do you read Slashdot?

  • Unklegwar (unregistered)

    The best thing about these customers is that you can charge them $5,000 and they think it's a bargain.

    Call it a Stupidity Tax.

    The invoice may have read "for optimization and refinement" but it's really a fee for just annoying us with such idiocy.

  • Tei (unregistered)

    User response table levels:

    L1) Warning: short operation takes long
    L2) Danger: server smoke
    L3) Panic: typo error on a form

    Theres a reason because BOFH whas invented.

    --Tei



  • ParkinT (unregistered)

     It was not my decision to charge them $5,000 for "optimization and refinement of algorithms".

    I am sure the customer feels like they got a bargain.

    It can be expensive to remove excess electrons from a working system!

  • dazed (unregistered)

    Ah well, that makes a nice change - instead of broken code, a broken customer. I've had a few of those, but none quite in this category.

  • (cs)
    Alex Papadimoulis:
    The customer complained that the analysis was "way too fast" and "must be broken".


    1. The customer is always right.
    2. The customer says the code must be broken.
    => 3. The code is broken.
    QED
  • (cs) in reply to Brian
    Anonymous:
    The real WTF... and it was barely over a month ago. I think we need an apology from the webmaster ala Slashdot when a duplicate story is posted!


    Ask for a refund instead.  Alex will deduct a reasonable charge for administrative handling and send you the rest.  Right, Alex?

    Sincerely,

    Gene Wirchenko

  • (cs) in reply to kipthegreat
    kipthegreat:
    Alex Papadimoulis:
    The customer complained that the analysis was "way too fast" and "must be broken".


    1. The customer is always right.
    2. The customer says the code must be broken.
    => 3. The code is broken.
    QED


    Jeremy did break the code per the customer's request.

    Sincerely,

    Gene Wirchenko

  • (cs) in reply to Hrishikesh
    Anonymous:
    This isn't limited to software...

    In fact, I had read somewhere that those making vacuum cleaners can tone down the motor drone quite a bit, but won't do so because their preliminary customer reviews showed that the users felt that the motor was underpowered because of the lower sound!

    So there, that's market dictating the law. And no harm so long as you can profit from it!



    About 4 years ago, I had an internship at a cell-phone company, and they were trying to strike a deal with Verizon to support their phones.  These phones had an internal antenna, but Verizon made us put an antenna on there because "people feel like they are getting a better signal if they raise an antenna."  The antenna was not connected to any internal electronics--just a placebo!

  • (cs)

    <FONT face=Tahoma>well, they got what they wanted...

    i'm sure they'll ask for continous improvement upto the point where the delay part is removed...

    customer is always right!


    </FONT>

  • david (unregistered)

    Heh ... I had a VERY similar experience many years ago.

    I was making a minor change to a report that a customer ran every day ... at the time it took a few hours to run.

    I noticed that the original programmer had gotten the blocking factors wrong such that the report read one record into the buffer for each sequential read it did ... and a thousand records into the buffer for every random read it made.

    Since this report was run in batch, that means it was very inefficent.

    I reversed the blocking settings and re-ran the report, with my minor modifications to the content, and the report ran in 45 minutes.

    Of course the customer was dubious ... and had me spend 2 days verifying that the report was still accurate.  No amount of explaining could convince them that the content change I made was trivial and could no way effect the amount of time it took to run.

    Oh well.

  • david (unregistered) in reply to kipthegreat
    kipthegreat:
    1. The customer is always right.


    Contrary to popular belief ... this is absolutely NOT the case.

    The customer is usually WRONG, otherwise they wouldn't need to hire consultants.

    The customer is, however, always the customer.
  • (cs)

    My only suggestion was that they didn't charge the customer enough. If they're that stupid of a customer, the charge for changes should be punitive. Also helps make the customer think twice about asking for changes... though thinking doesn't seem to be their strong suit in the first place.

  • JSteiner (unregistered) in reply to OneMHz
    OneMHz:
    My only suggestion was that they didn't charge the customer enough. If they're that stupid of a customer, the charge for changes should be punitive. Also helps make the customer think twice about asking for changes... though thinking doesn't seem to be their strong suit in the first place.


    Thinking like Labor, and not like Management! Shouldn't you want to price changes as high as possible, but low enough that a good number of customers almost never think twice about asking for changes, so you can get as much money from them as possible on an on-going basis?

  • Tilendor (unregistered) in reply to david

    I claim that the customer is not always right.  But the customer is always worth humoring.

    ~Til

  • (cs) in reply to JSteiner
    Anonymous:
    OneMHz:
    My only suggestion was that they didn't charge the customer enough. If they're that stupid of a customer, the charge for changes should be punitive. Also helps make the customer think twice about asking for changes... though thinking doesn't seem to be their strong suit in the first place.


    Thinking like Labor, and not like Management! Shouldn't you want to price changes as high as possible, but low enough that a good number of customers almost never think twice about asking for changes, so you can get as much money from them as possible on an on-going basis?

    Well, I am labor, not management. Since I don't get more money for more hours worked on a customer's request, I'd rather see the customer get monetarily abused into silence when they explicitly ask for the system to be crippled, or slightly fix a cripple they asked for. Save me the frustration and the cash on maalox.

  • Approaching Two Gallons (unregistered)

    Sounds like the Q & A system used by the Red Cross for blood donors.  Regardless of the length of the question there is a fixed delay befor the answer buttons are enabled.

    "Have you used mass transit which may have carried a midget exposed to someone vaccinated for smallpox within the last 73 months who has travelled outside the United States for more than 1200 hours in the last 3.511 years?"  Let's see.  Does the 73 months refer to the midget's exposure?

    "Do you feel healthy today?"  Wait for it... .  Wait for it ... .  Now you may answer the question!

  • (cs) in reply to david
    Anonymous:
    Heh ... I had a VERY similar experience many years ago.

    I was making a minor change to a report that a customer ran every day ... at the time it took a few hours to run.

    I noticed that the original programmer had gotten the blocking factors wrong such that the report read one record into the buffer for each sequential read it did ... and a thousand records into the buffer for every random read it made.

    Since this report was run in batch, that means it was very inefficent.

    I reversed the blocking settings and re-ran the report, with my minor modifications to the content, and the report ran in 45 minutes.

    Of course the customer was dubious ... and had me spend 2 days verifying that the report was still accurate.  No amount of explaining could convince them that the content change I made was trivial and could no way effect the amount of time it took to run.

    Oh well.

    Surely you mean "in no way affect the results"? Obviously it did affect the amount of time it took to run, or they wouldn't have noticed...

  • 3 Hurried 2 login (unregistered) in reply to Albatross
    Albatross:
    I notice nobody's actually talking about the WTF in question...


    Your point being?
  • (cs) in reply to david

    The real WTF is that they didn't charge more for the "upgrade".

  • qbolec (unregistered) in reply to Manni

    Manni:
    Holy hell, what are the chances this would happen to Jeremy twice???

    ROFTL :)))))))))))))

    hahahahahaha

     

  • (cs)

    Intentional slowness, eh? I do that all the time, only for me I have a reason: I'm learning as I go.

    I'm a systems management guy, I can create a beautiful solution that integrates Nagios with Cacti with MOM 2005 with OpenView with Sitescope - but the code behind it isn't pretty.

    So usually I start out with horrible horrible code, and each revision it gets smaller and faster.
    Right now my beast is PHP. You do NOT want to see what the code behind this SNMP Poller is like. But they love it, it will see a Compaq Windows server vs a Sun Solaris server vs a Cisco router vs an HP Procurve switch and display only the info about that item. And it makes me happy hehe.


    Oh, here's 2 of the biggest WTFs in my company:

    One of the net engineers, the Principal one, had written scripts that would build MRTG config files (For windows MRTG). His initial scripts were written in AWK.
    People complained about the lack of awk in the company and told him to rewrite it - so he did.
    Now it's an excel spreadsheet that's dumped to a word doc, then edited by hand, that queries 3 SQL servers, then runs a mailmerge in Word that spits out the config - which he then has to go through and remove all the WIncentric CRs.

    The reports on how we are doing, our metrics, is created by several Excell spreadsheets, which talk to an access DB, which talks to several production SQL DBs, which is then cut and pasted to a new excel, which is then... well, it's nasty - let's just say that.

  • (cs) in reply to kipthegreat
    kipthegreat:
    Anonymous:
    This isn't limited to software...

    In fact, I had read somewhere that those making vacuum cleaners can tone down the motor drone quite a bit, but won't do so because their preliminary customer reviews showed that the users felt that the motor was underpowered because of the lower sound!

    So there, that's market dictating the law. And no harm so long as you can profit from it!



    About 4 years ago, I had an internship at a cell-phone company, and they were trying to strike a deal with Verizon to support their phones.  These phones had an internal antenna, but Verizon made us put an antenna on there because "people feel like they are getting a better signal if they raise an antenna."  The antenna was not connected to any internal electronics--just a placebo!


    That's funny because about 2 years ago I had a T-Mobile phone with a little plastic antenna that didn't seem to make any difference to the signal.  I ended up just ignoring it was there.  Wonder if it was the same situation.  My current phone doesn't have a visible antenna.

    The real WTF is the upgraded forum software just traded one set of WTF's for another.

  • LC (unregistered) in reply to Jeremy D. Pavleck

    ...Printed out.

    then faxed to a colleague...

    who scans it in...

    converts it to a pdf...

    then emails its to his boss...

    who prints it out...

    places it on a wooden table...

    takes a photo...

    scans it back in...

    and finally posts it onto the website for the developers to download and type in.

  • (cs) in reply to LC
    Anonymous:
    ...Printed out.

    then faxed to a colleague...

    who scans it in...

    converts it to a pdf...

    then emails its to his boss...

    who prints it out...

    places it on a wooden table...

    takes a photo...

    scans it back in...

    and finally posts it onto the website for the developers to download and type in.


    AH, I have pretty good OCR software on my computer. Foiled!
  • (cs) in reply to Jeremy D. Pavleck
    Jeremy D. Pavleck:
    Oh, here's 2 of the biggest WTFs in my company:

    One of the net engineers, the Principal one, had written scripts that would build MRTG config files (For windows MRTG). His initial scripts were written in AWK.
    People complained about the lack of awk in the company and told him to rewrite it - so he did.
    Now it's an excel spreadsheet that's dumped to a word doc, then edited by hand, that queries 3 SQL servers, then runs a mailmerge in Word that spits out the config - which he then has to go through and remove all the WIncentric CRs.

    The reports on how we are doing, our metrics, is created by several Excell spreadsheets, which talk to an access DB, which talks to several production SQL DBs, which is then cut and pasted to a new excel, which is then... well, it's nasty - let's just say that.


    How selfish of you to keep that to yourself.  You should share these WTFs.

    Sincerely,

    Gene Wirchenko

  • (cs) in reply to kipthegreat
    kipthegreat:
    Anonymous:
    This isn't limited to software...

    In fact, I had read somewhere that those making vacuum cleaners can tone down the motor drone quite a bit, but won't do so because their preliminary customer reviews showed that the users felt that the motor was underpowered because of the lower sound!

    So there, that's market dictating the law. And no harm so long as you can profit from it!



    About 4 years ago, I had an internship at a cell-phone company, and they were trying to strike a deal with Verizon to support their phones.  These phones had an internal antenna, but Verizon made us put an antenna on there because "people feel like they are getting a better signal if they raise an antenna."  The antenna was not connected to any internal electronics--just a placebo!

    As I heard it once (and I haven't dug enough to confirm or deny this, but the Cisco documentation seems to support it...), in the early days of VoIP, there was no dial tone when you picked up a handset; in a VoIP system, it's unnecessary.  Early testers didn't like not hearing a dial tone, so the phones were changed to play a .wav file at the user.  I'm out of the phone business now, but supposedly, it's just a 'standard' .wav file, and can theoretically be set up to play just about anything you want at your users.  (The potential for BOFH-ness is amazing ...)

  • (cs) in reply to Gene Wirchenko

    Good point Gene, I think next week as I'm recovering from my Memorial Day drink-a-thon I'll write these two up and share the tails

Leave a comment on “It Must Be Broken”

Log In or post as a guest

Replying to comment #:

« Return to Article