Blog

The Prediction Problem

On January 17, 2013, in News, Newsletter, by Association for Software Testing
0

By Matthew Heusser

In December and January, a tester’s mind is likely to turn to the future – to what we might accomplish and how testing might change.  We make lists; we get excited. Being critical thinkers, you might ask yourself “how can I check the great theories I just came up with?”

I wonder the same questions myself, and, while it might not be possible to check the future, it usually is possible to check the
past.

It being December 29th, I did a little fun:  Googling “Future of Testing” and a year. I also googled “Future of Software Testing” and looked for popular, but old, articles. Here are a few of the more popular gems:

* We’ll get away from testing and move toward quality assurance!
* All tests will be automated
* Testers will become programming-testers

And my Personal Favorite:

* Tests will become re-usable across projects. Instead of writing your own tests, you will buy and integrate existing, off the shelf
test automation.

Yes, someone actually claimed this.

At a keynote speech.

At Google.

Half a decade ago.

It’s not just testing; we have the same problems with development and management literature. If you check those, you will see the same insistence that “tomorrow’s CIO/CTO” will need to have “both tech and business skills” and be “as comfortable in the suit as in the server room.”

The first time I read one of those articles, it was 1997. I think Larry and Sergy, founders of Google, were in grad school in Ann Arbor, clearly more comfortable in the server room than the board room.

Somehow, they did okay.

As you can see, I’m reluctant to make predictions.

What I can do is look back and spot trends, trends that haven’t been named yet but are clearly coming. For example, a talk on the future of testing about social media back when it was friendster, crowd sourcing when uTest was looking for series A funding, mobile applications when the blackberry was the new hotness, or the potential for virtualization back when it was something sysAdmins were doing to save hard drive cost – those all would have been real, valuable insights.

Do I have anything like that up my sleeve?

Not earth-shattering, but I think I have a thing or two.

My friend Scott Barber gave a talk at Agile Testing Days this November, and claimed that we would eventually come to think of 2011 as the time of a great split between devops/cloudy/agile and traditional, regulated, perhaps scripted testing. I think he had something there – and that many of the skills that made testers successful in traditional, regulated, perhaps scripted-land are less appreciated and less valuable in devops/cloudy/agile-land. So if you are in a traditional/regulated environment and not having fun (or vice versa), it might be time to pay attention to the market and do a little re-skilling.

In addition, I think it’s not going to be a two-way split.  It will be a continuum. There will be a lot of different ways to develop software, and a lot of different ways for testers to add value.

I’m about to make a prediction, here, and I’m going to be accountable for it.

In my mind, the “future of testing” folks tend to make the same mistake: They assume testing will converge onto one right way to do things. I think a million test methodologies will bloom, and we’ll pick the ones that are right for this team, this project, at this point in time.

I think the future of testing is its divergence – at least for the next five years or so. Yes, in general, I agree with Scott, there is going to be a rift between the guys who build an app over the weekend using Amazon Web Services, and the folks who have a separate and distinct operations group that checks a change control checklist before moving code to production — but there will be a million steps in-between, not to mention other dimensions, like the risk tolerance of the team, the ability to rollback, the pressure between “gotta have new features now” and efficiencies from doing regression testing a little less often … it’s enough different dimensions to make a run-on sentence and just scratch the surface.

So there, I made a prediction. I intend to be accountable for it. Ask me again in five more years if I was right about testing diverging, and why.

I’ll be here, in this test community.

It is an honor, by the way, to be in the test community, and on this board of directors.  Thank you all for electing me to serve.

Tagged with:
 

Comments are closed.


Looking for something?

Use the form below to search the site:


Still not finding what you're looking for? Drop a comment on a post or contact us so we can take care of it!