back to article Ever wondered why tech products fail so frequently? No, me neither

It's not working. Sorry, this has never happened to me before. Actually it has, frequently, but let that pass. Can we try again in a few minutes? Foolishly, I agreed to help an ex-colleague with some user acceptability testing this week. It's a chore I swore I'd never do again, such that for my own digital publishing projects …

Page:

      1. Voyna i Mor Silver badge

        Re: UAT Testing

        "My user acceptability testing is done to discover which users are acceptable to the software. That solves a lot of problems."

        Didn't Virgin Mobile start off on that idea? By being more or less online only, they limited themselves to customers who were capable of doing stuff online, and who were therefore less likely to need handholding with technical products. Giffgaff and now Smarty went one better by not supplying phones at all, thus reducing overhead to the bone.

  1. Anonymous Coward
    Anonymous Coward

    Why IT Projects and Products suck

    Being very general here and basing the following points on more than a few decades of experience there are some blindingly obvious things that keep on happenings time and time again.

    1) The "We must be first to market this widget"

    No need for further explanation

    2) The "Make it prettier/sexier" demand.

    Bling over functionality.

    3) Feature creep.

    Especially true for Government projects. The 'Wouldn't it be nice...' things that are added with no respect for budget or timescales.

    4) Lack of testing

    The - you did the unit tests. That's good enough. Ship it!

    5) Documentation - both for users and support.

    No one RTFM's any more so don't spend time doing stuff that isn't wanted.

    There are many, many more that I'm sure the commentards here can add.

    But often it goes down to

    "What will go wrong, will..."

    But PHB's and MBA's don't care.

    1. John G Imrie Silver badge

      Re: Why IT Projects and Products suck

      But PHB's and MBA's don't care.

      If it can be made to work for this quarter the PHB will get his, see the Ex-Presidents Club for why this is usually a 'his', bonus, after that, who cares.

  2. DailyLlama
    FAIL

    Oh, the irony!

    I clicked the link from the email, and Lo! was presented with this error:

    Your connection is not secure

    The owner of www.theregister.co.uk has configured their web site improperly. To protect your information from being stolen, Firefox has not connected to this web site.

    Learn more…

    Report errors like this to help Mozilla identify and block malicious sites

    www.theregister.co.uk uses an invalid security certificate.

    The certificate is not trusted because the issuer certificate is unknown.

    The server might not be sending the appropriate intermediate certificates.

    An additional root certificate may need to be imported.

    Error code: SEC_ERROR_UNKNOWN_ISSUER

    I initially thought it was deliberate, but alas, I can't open any El Reg pages in FF today...

    1. Anonymous Coward
      Anonymous Coward

      Re: ...uses an invalid security certificate.

      I also seem to get this occasionally (also FF), although not (yet) on ElReg. I can't ever seem to work out why, despite looking. However, it usually it goes away after a day or so.

    2. Jos V

      Re: Oh, the irony!

      DailyLlama,

      This is not necessarily caused by mis-configuration on ElReg. There are a number of other reasons for this to happen, like a corporate network monitoring product intercepting secure traffic and replacing certificates, Microsoft safe "Family" settings, which replace the certificates with a MS certificate so they can do monitor/filter traffic, or your local security software doing much the same (ESET, AVAST, BitDefender, etc.)....

  3. Chairman of the Bored Silver badge

    Know thy user

    Sometimes you need a tester who truly understands the environment.

    In meeting in which a vendor showed off a 'improved signal processor line replaceable unit hardened to survive the naval environment'. Program manager handed it to his deputy and said, "You're former navy. Does this look sailor-proof to you?"

    Deputy... "I dunno..." Smashes it into edge of table, kicks it across the floor, slaps it into floor. Stuff rattles inside. "No, sir! Not sailor proof!"

    Vendor ... 'WTF!!!??? That cost ten grand!!!... You're gonna.. "

    PM? "Yeah, ten G. But it's not worth $#!t!"

    1. Voyna i Mor Silver badge

      Re: Know thy user

      Also known as the "Fourteen stone sergeant product stand on test".

  4. Dr_N Silver badge

    Appliance Failures

    I think it's all down to abuse.

    Having two vacuum cleaners failing within a year? Hmmmmm...

    Catch the girlfriend vacuuming the bare, slightly damp concrete floor of the garage because "It's quicker than sweeping up the dust. I've always done it this way. What's the problem?" A real lightbulb moment.

    1. Alistair Dabbs

      Re: Appliance Failures

      >> I think it's all down to abuse.

      As I said, perhaps I broke my spectacles by looking through them too hard.

      1. Anonymous Coward
        Anonymous Coward

        Re: Appliance Failures

        You can never look too hard in spectacles.

      2. Dr_N Silver badge

        Re: Appliance Failures

        Designer glasses are (pretty much) all made in one factory in Italy, so easy to investigate?

        And they pretty much have a lifetime guarantee. (In France you can get "free" glasses every 2 years under mutual health cover schemes so maybe they are built to last days?.)

    2. bat

      Re: Appliance Failures

      you can see your garage floor, where do you store all the old IT Kit!

      1. Anonymous Custard Silver badge
        Mushroom

        Re: Appliance Failures

        Nah, I think Dabbsy should just give up this pretence of being a mild(ish) mannered journalist and admin that he actually is Superman...

    3. jelabarre59 Silver badge

      Re: Appliance Failures

      We have a Sears washing machine with a lid-switch held in with a u-shaped tab. *When* (not *if* it breaks, and you have to fit in a new one, the piece is so tricky to install properly, the tab on the NEW one breaks in the process of installing it. I can't help but think it was *intentionally* made that way to force you into paying for an (expensive) Sears "Service" Technician to install it for you.

      So instead I just leave the top panel unscrewed, and just reach inside to pull the switch into position to start a load running. feck them.

      1. Shadow Systems Silver badge

        At Jelabarre59, re: that switch.

        I think we either have the same make of machine or else Sears' quality has been flushed down the shitter. Just after the warranty expired on my machine that damned clip started breaking with freakish regularity. After about the 3rd time that damned clip broke I examined the clip, took a thick paperclip & a pair of needle nosed pliers, & recreated the clip shape out of the paperclip.

        Guess what? it's been nearly a year since I did that & I've yet to need to replace that bit o' bent wire. The lid works, the switch works, but that original piece of shitty plastic clip is no longer part of the loop.

        Grab yourself a few paperclips (for practice), a pair of needle nosed pliers, the old clip, & fiddle with the paperclips until you can make a plausable fake. You'll save money & headache meds galore.

        Cheers!

        1. Anonymous Coward
          Anonymous Coward

          Re: At Jelabarre59, re: that switch.

          In the late 1980s I did a re-plumb of the house central heating using two Honeywell motorised three port valves. The motors used to fail quite often. Fortunately you could buy spare motors - although a "real" plumber would have changed the whole expensive valve unit.

          Then they stopped failing. Presumably at some point Honeywell had had enough warranty returns to figure out the problem.

        2. Antron Argaiv Silver badge

          Re: At Jelabarre59, re: that switch.

          Zen and the Art of Washing Machine Maintenance?

  5. allthecoolshortnamesweretaken Silver badge

    They also serve

    who only sit and test.

  6. Bill M

    UAT vs Speadsheet

    UAT: That number is wrong

    Dev: In what way ?

    UAT: Not the same as my spreadsheet

    Dev: Can you email me your spreadsheet so I can cross check

    Next day

    Dev: I found a bug in your spreadsheet, here's a fixed version of it

    UAT: Your spreadsheet is wrong

    Dev: In what way ?

    UAT: Not the same as my spreadsheet

    Dev: You mean the spreadsheet with the bug in it that used local Darwin time for KPI's and compared them to targets based on GMT ?

    I don't jest, I came across this with some metrics peoples' performances / bonuses were being measured against . There had been argument's going on for years - When people get short changed on bonuses are not happy, but strangely keep quiet on receiving a bonus they don't deserve. Somebody had even been wrongly fired. Eventually I got hauled in front of the board to explain. The FD was competent, soon grasped the issue and got the poor chap who had been fired reinstated shortly before the industrial tribunal was due.

    1. Anonymous Coward
      Anonymous Coward

      Re: UAT vs Speadsheet

      Bill M,

      Well Done !!!

      Also well done to your FD who had the integrity to accept the Company was at fault and got someone their job back.

      More usual is the problem is fixed (by you) but the whole episode is swept under the carpet as it shows the Company up and might have an impact on the Directors Bonuses and/or Salary increases, which is a definite No No !!!

      As Industrial Tribunals are a 'Flip of a Coin' most times, particularly when the Company 'sics' their 'pack' of in house lawyers on the poor ex-employee, I have seen Companies continue the Industrial Tribunal process even when they know they are in the wrong. The odds are seen as being worth it versus a possible loss and the subsequent 'Bad Publicity'.

  7. steelpillow Silver badge
    Holmes

    Time to market

    It's because consumers demand tomorrow's fashion yesterday. Creators are forced to bring stuff to the market before it is ready or miss the boat.

  8. macjules Silver badge

    UAT/QA Testing

    Never heard of Selenium for testing?

    1) Developer writes script for testing page

    2) QA/UAT tester runs script in browser with minimal effort required

    3) Test fails - push it back to developer

    4) Test passes - move to done

    1. Anonymous Coward
      Anonymous Coward

      Re: UAT/QA Testing

      > "Never heard of Selenium for testing?"

      Selenium tests are OK for basic checking (does element get rendered? can I do X?), but they have their limits (can I easily see button Y?). They also tend to be slow and brittle (I've been subjected to a project where the Selenium test suite took over 8 hours to run, and have seen /all/ the tests start failing because Chrome/Firefox autoupdated and broke the driver [and there wasn't an updated version available yet]).

    2. tiggity Silver badge

      Re: UAT/QA Testing

      And test scripts should ideally be written by a different dev / tester to the code being tested.

      As if the original dev had missed something important in implementation, chances are would not suddenly remember to add that to the tests of their own code.

      1. Ken Hagan Gold badge

        Re: original devs

        You might be surprised. Testing (or even writing a user guide) can often be enough to make the original dev think about edge cases and sometimes they are the only people who know where the edges are. Or to put it in the language of testing, you need white-box testing as well as black-box testing.

        Writing a user manual is another activity that can make the original dev consider their work from a new angle. I know no-one reads them anymore, but that doesn't mean there is no value in writing them.

        1. Doctor Syntax Silver badge

          Re: original devs

          "I know no-one reads them anymore"

          That's because nobody writes them any more.

          1. Charles 9 Silver badge

            Re: original devs

            No, it's because no one expects to need to read them anymore. If a device isn't pick-up-and-play intuitive without directions, it's considered too complicated.

            1. dajames Silver badge

              User Manuals

              No, it's because no one expects to need to read them anymore. If a device isn't pick-up-and-play intuitive without directions, it's considered too complicated.

              Nevertheless, I know people who complain that they can't use modern software because they haven't got a manual to read. The idea of FWITIW (Eff With I Till It Works) doesn't seem to appeal to them.

              Some modern software just is user-hostile crap (and a manual would make that failing more obvious) but some users do seem to need a manual even to use the best-designed programs.

              1. EarthDog

                Re: User Manuals

                When you are on 2 week sprints FWITIW doesn'tt work.

          2. Terry 6 Silver badge

            Re: original devs

            Somewhere about 1990 manuals started to explain in excessive detail anything that might be blindingly obvious - while omitting to explain anything that was complicated or hidden. At which point they became a bit useless.

            As in a full explanation about how to use the on/off button, but not a word about how the switch works that somehow managed both to open one end of the dust collector and also release the other end from the device body for fuller cleaning.

        2. DJO Silver badge

          Re: original devs

          Writing a user manual is another activity that can make the original dev consider their work from a new angle

          Devs are generally the worst people to write manuals, you end up with Microsoft style documentation which goes into extreme detail telling you everything except the things you actually need to know. This might one of the reasons that hardly anybody looks at them.

          Get a proper technical author and have the dev provide copious notes, of course that note creation can have the same serendipitous benefit for the devs you suggested.

          1. Terry 6 Silver badge

            Re: original devs

            Yeah in my post an hour or so before DJO's I didn't attribute this to Microsoft in particular, because it's so common, but their "Help" is a particularly egregious example of explaining the obvious while ignoring the stuff that isn't apparent. In my opinion the nadir was Win 8.x because they just wanted users to click on the big icons and run the programmes like phone apps- no user interaction required.

            1. Antron Argaiv Silver badge

              Re: original devs

              @Terry6

              Re Microsoft "Help"

              I thought I was the only one who'd noticed this. There also seem to be points awarded for how many clicks they can get out of you while continuing to avoid providing any useful information.

    3. Anonymous Coward
      Anonymous Coward

      Re: UAT/QA Testing

      RE: "Never heard of Selenium for testing?

      1) Developer writes script for testing page

      2) QA/UAT tester runs script in browser with minimal effort required

      3) Test fails - push it back to developer

      4) Test passes - move to done"

      Not according to our QA dept.

      Devs write junit tests. QA writes Selenium tests, and ensures you can only build them in the special QA environment based on special complete builds of everything in the world. Devs can only be told summaries, like today's testing degraded from 1500 passes to 1400, with problems in some broad area. QA then writes up one or two of the worst as detailed manual reproduction steps, and files bugs. They don't see the point in "letting" developers run selenium.

      Of course, it also took 6 months, and detailed traces of build times with and without my personal SSD, along with cost analysis to approve getting those, and close to 2 years to get some Devs to use VMs instead of uninstall versions of multiple apps, install old versions, do patch, repeat.

      Anon for some reason...

      1. macjules Silver badge

        Re: UAT/QA Testing

        Yep. Know the feeling.

        Moved Dev testing to CI integrated Sonar testing only to get bitched at by devs when their ‘code’ fails to pass unit tests and Jenkins spits the errors back at them. We now have to write out full QA tests, the Loadtests and just about everything except send them loo rolls to do the obvious.

    4. EarthDog

      Re: UAT/QA Testing

      No! No! No! No! Developers never write Selenium (or related products) tests. Modern QA practice require Qas to have scripting or flat out developer skills with the devious mind of a QA and the ability to remain ignorant of any but the functional requirements.

  9. WibbleMe

    Real people want humanity enhancing devices, not crap Tec that's sold in the local pound shop

    1. Anonymous Coward
      Anonymous Coward

      Re: Real people want humanity enhancing devices, not crap Tec that's sold in the local pound shop

      You mean there are real people out there?

      Can't be. Look at the local pound shop...

  10. Chasola

    Mark E Smith

    I'm glad to see that you share the grief of the North.

    Victoria will be very upset.

    I'll raise a glass to him tonight at the Hex Induction Hour.

    Sorry for the irrelevant comment but some things are more important than tech.

    chas

    1. Charlie Clark Silver badge
      Pint

      Re: Mark E Smith

      Hit the North! Nice of Dabbsy to sneak the tribute in.

      Pint, for Mark E Smith, of course.

  11. Prst. V.Jeltz Silver badge

    my retail goods usually fail when I lend them to someone else . Strange that.

  12. Prst. V.Jeltz Silver badge

    Has anyone noticed that software testing jobs appear to pay almost as much as software developer jobs , and much more than most I.T support jobs? They seem to be much higher up the tree than you'd have thought.

    ..and yes I know there will be certain skillset , and certain tools to be familiar with - but same and more so for all the other IT jobs.

    1. Anonymous Coward
      Anonymous Coward

      You think we do this for the money? Hell no. Nothing beats the soft blubbing of a pm as they slowly fall apart, or the dull, beaten, lifeless gaze of a dev you twisted into knots with his own logic? Money? That is for mortals.

    2. EarthDog

      I put this in another post but will place it here too.

      A good QA should have a devious mind, be good at scripting or full scale software development, good at following the functional design of a product, the ability to communicate with both managers and developers, and able write good tests (as opposed to writing scripts if you catch the difference). Programming is a much narrower pursuit.It is not a position to dump your worst developers into.

      1. Helldesk Dogsbody

        One group you can also throw at QA, particularly for regression testing, are your support staff. Been there, done that. When told "But NOBODY would ever do that! It's not a valid test!" all you need to do is pull up the previous weeks first line tickets...

  13. Fading Silver badge

    UAT....

    There are five stages of UAT;

    Denial: - this is not what we specified

    Anger: - No really this is not what we asked for the damn thing doesn't even do what the old system did!

    Bargaining: - if you can just make it do the original thing we might be able to sign off the first phase?

    Depression: - I can't believe we paid so much and we will have to use this.

    and Acceptance.

    UAT - testing the users until they accept what you have developed.

  14. W4YBO

    Emitting a loud "crack!...

    "Emitting a loud "crack!" they actually exploded into several plastic artefacts while sitting on my face as I was reading. I dunno, maybe I was looking through them too hard."

    I had a freshly ground prescription lens do something similar. New wire-rimmed spectacles, just a couple of hours after picking them up, and I noticed a curved line emanating from a corner of the right lens. Just as I was removing them, POP!, and the right lens was in several pieces on my desk. LensCrafters technician told me the grinding lab probably forgot to stress relieve the lens, and pressure from being mounted in the rim drove the stresses past the point of self-disassembly.

    1. Doctor Syntax Silver badge

      Re: Emitting a loud "crack!...

      "the grinding lab probably forgot to stress relieve the lens"

      Glass can have lots of internal stresses. For my degree project I devised an experiment which required a number of stands made out of glass rod with 6 bends in each (think of those Bauhaus tubular frames chairs and you get the idea). Over a few weeks each of the initial bunch made out of soda glass developed cracks at at least one joint. They had to be remade in Pyrex.

      1. Voyna i Mor Silver badge

        Re: Emitting a loud "crack!...

        "Over a few weeks each of the initial bunch made out of soda glass developed cracks at at least one joint. They had to be remade in Pyrex."

        Soda glass needs careful annealing; borosilicate has a much lower coefficient of thermal expansion and doesn't.

        I've watched the (amazing) machine that makes laboratory glass tubing. It's enormously long, so as the molten glass is extruded and drawn to correct diameter it cools slowly enough not to lock in stress. But having said that, if you have a lot of bends that need careful cooling, it will probably save a lot of time if you use borosilicate glass in the first place.

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019