Here’s a short (<10 min) YouTube clip taken out of National Geographics channel "Megastructures" series about building of Burj Al Arab in Dubai (a great piece of architecture BTW, but this time it’s not Santiago Calatrava – the architect’s name is Tom Wright).
One interesting subject here is problem of installation of very bit steal-tubes made “trusses” ends of which didn’t fit exactly the junction points because of thermal expansion – the distance from one point to another changed in +/- 5 cm in 24h.
Eventually one of the architects proposed simple yet effective solution – have a bigger “hole” at one junction in which the round steel bulk was placed, with eccentric “hole” of needed size in that bulk. So that when the bulk was rotated inside it’s “bigger hole”, the “smaller hole” used for top junction of truss was like changing it’s distance from the fixed bottom junction.
The funny moment is at 4:30: “Engineers thought about it, the structural fabricators, the steel fabricators, the architects, everybody thought about it, and it was actually one of the architects who came up with this great solution”.
When I heard this first thing I thought was… have they ever heard of division of labor? (-:
Well, in this case it actually is somewhat logical and forgivable – at least for engineers and architects it’s quite Ok IMO.
But severe cases sometimes (luckily not often) happen in software development… (don’t read further if you are one of my current or past supervisors)
Isn’t it ridiculous, idiotic and highly unprofessional when system administrators, tech leaders, managers and other kinds of “expensive” and just unsuitable personnel do manual testing? When technical leaders do teaching of product usage? When developers do project status reports? When technical writes do planning? And so on and so on.
I now learned that there are people in this world who participate in software development at leading roles, but just don’t know what division of labor is. Some of them seem to imagine work as a big pile of $#it “we all” must eat (and this attitude is repeatedly showing up for some people in some details), so dividing it into pieces per person with sizes roughly corresponding to salary sizes is a division of labor. But it is not!
The division of labor is based on kind – not amount – of work to be done.
Division of labor is not some fancy $#it invented by crazy purists and inapplicable to real life – it’s actually working way to gain effectiveness, which means to have better results with less expense. Looks like far not all people (especially in Western Europe, including Switzerland and other *lands, and also some particular individuals in Ukraine) are aware of it.
Even if there is a manual testing in progress and there are not enough testers but some “free” developers or managers it doesn’t mean they should go do testing. It means you should have better planning to keep everyone equally occupied over time. And possibly means you should hire some testers (and maybe fire some devs, or better managers because how there can be a “free” expensive manager? especially regarding the fact the planning flows made some departments overloaded while others are “free”).
Involving non-testers in testing will give low addition in quantity, but will result in significantly lower quality – it is practically quantity over quality, and not so much quantity gain actually (although it depends on particular numbers of course).
BTW, I’m a bit out of latest data, but manual testing here in Lviv seems to be in a “sky is falling” state to me. Except for EPAM (which BTW is relatively new in Lviv) that AFAIK has some special tests for testers, but I don’t know if they do any teaching, nobody nowhere does teaching of testers how manual software testing should really be performed. Nobody teaches them how bugs should be reported and in which way the reproduction steps/conditions may be determined (and what to do when you cannot determine them). No university and no company training center is preparing people to be real testers AFAIK (SoftServe’s training program for testers is very very very weak AFAIK, so I don’t count it at all).
Or did I miss something?