Wednesday, February 11, 2015

The World Wants to Replace You With Something Better

In a world where technology is constantly getting better, with fewer jobs and more automation (speaking generally, not 'test automation'), the world appears threatening.  However, the world has always been this way, just less extreme.  Politicians have always looked for better funding, new ways of communicating and ways to look better.  They slowly edged each other out, but in doing so changed the world, each forging more and more ways of getting funding/communication channels/etc.  The world doesn't want you, it wants the next guy's better idea.  I speak of politicians because it is an easy way to see how people 'vote' for what they like (not what they need or should get, but that could be a whole blog post of its own).  Darwin speaks of this in terms of evolution.  Markets speak of this in economic terms using money as a voting mechanism.  One general term that is used is progress.

The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man. ― George Bernard Shaw

Automation of all stripes appears inevitable.  Will all jobs go away?  I doubt that, at least not in my life time, but what will the world look like?  Have you read and considered The New Luddite Challenge?  Like Bill Joy, I have and I still don't have a lot of answers.  I want to automate things because I hate wasting time doing the same boring things over and over again, but I don't think all work is a waste per-se.  Some people really like cooking, and so they are not wasting their time cooking, even if they cook similar dishes over and over again. On the other hand, just surveying some of my co-workers, I am sure cooking is not a pleasure for everyone. Even worse, waste in cooking is not just time, but also in burnt meals and trashed leftovers no one wanted.   In fact, in the last 50 or so years, we have started seeing food automated be it via fast food or food in a freezer.  While we are not forced to have those sorts of meals, automation has left many of us without particularly strong cooking skills.  The world has replaced what might have been superior meals for meals that require little-to-no washing up or cooking at all.

If you look at recent history, we had different people doing whatever it was that they enjoyed.  They would then hire other people to do the things they didn't want to do.  For example, I test software, but I don't much care to do my own garage repair, so I hire someone to do that.  If we automate some jobs or even just sub-tasks in jobs away, what do we do with those who want to still do the job that was then automated?  One of the easy examples is what happens to truckers when we automate driving?  Obviously children that don't know the 'joys of truck driving' won't dream of being truckers, so maybe this is a self-correcting problem, but I don't like saying 'wait until all those old school folks die off'.  I am not here to give answers to this topic, certainly not for such a complex set of questions, but I think if you write automation, you should seriously explore the effects and ethics of the profession and how it changes the world, before you create a world you don't like.

In fact, I would go so far as to say you should frequently examine what you are doing to see if it is what you want to see done.  That is to say, if you work for a company that gets cat pictures to people, is that something you care about?  Do you feel that another cat picture platform is what matters?  Perhaps you'd rather hack together a bunch of low-value apps and make a few hundred per app.

A Real World Example


Ultimately, our automation of simple tasks might eventually cost society some of our jobs.  Yes, test-related jobs can be lost by automation.  To get the regression I have produced with automation would require more people than we could hire, but to even reach something near that level, my company would have to hire 3-5 people with low levels of knowledge ("Beginners").  My wage might meet these theoretical low-level beginners, but they would be simple button pushers.  While there are pros and cons to automation (E.G. Testing vs Checking), I think the coverage could be argued as less for the human side, but to be fair, given enough training, the quality of the testing might go up.

Let us speculate in my circumstances.  There are 50+/- fields that have to be filled out, with different types of allowed values based upon different modes across 12 different version of the product.  These are all API-based, and with very simple text-feedback.  It took me about 6 to 12 months to automate to the level I expect these low-level testers would work at, and maybe 3 of those months were around discovery of how the system worked.  It was all regression work with only 2-3 new features over that time, all of which were covered by a different tester.  In my speculation, I would imagine these beginner testers would only do the regression, not the new features.

So let us consider what would happen if we hired those testers.  The testers would work for ~6 hours a day, and would be typing in input and checking the output.  It would increase headcount, and they might need another manager or at least a lead.  The other pieces of automation wouldn't have been able to leverage the work and we would have had to hire even more people.  Our organization would have ballooned into a factory school testing organization.  It would have been ineffective, but would have had more human-coverage.  I have worked in such an organization 10 or so years ago, we had ~100-150 people banging away at 20+k worth of written tests.

I believe automation in this case was the right choice.  I think it adds serious benefits to the organization and those benefits are long lasting.  I think it is best we never hired those fictional testers.  Yes, I recognize that I have created a false-dichotomy in that there could have been other possible choices in the hiring process, such as 2 mid-level exploratory testers.  I still think automation was a better choice.  Perhaps not testing this system could have been an option, but I think automation was a good idea... or using unit tests... or ....  Perhaps there was a better option, but they got me and my automation.

The Future Is Now

As I have documented before, many organizations are moving towards a No Tester view of the world.  In a very real possibility, some software may not need to be tested in the traditional manner.  The future may not need you. In some ways, I think that is reassuring, but for others it is indeed sad.  Charles de Gaulle popularized the quote, "The graveyards are full of indispensable men."  My efforts, as valuable as they are, are only temporary until the next innovation, but also my mistakes are only temporary until they are patched.

Everyone must leave something behind when he dies ...  It doesn't matter what you do, so long as you change something from the way it was before you touched it into something that's like you after you take your hands away. ― Ray Bradbury, Fahrenheit 451

As we keep moving forward, change will continue to come.  The lesson in this is not to give up.  It is not that you shouldn't go for a testing job.  It is not that you shouldn't bother coding.  Rather, go learn a new skill and tomorrow learn another.  The future wants a better world, and while none of us know exactly what that will ultimately mean, it certainly does not mean you should stagnate and quit learning.  Go learn.  Then find hard problems and work on them.

No comments:

Post a Comment