Testbash 2017 – part 2

Testbash 2017 – part 2

Ahhhh caffeine. ūüôā

Professor Harry Collins up next, this will be very interesting

Artificial intelligence, language and the net

This already feels like I’m back at university listening to a lecture by one of the professors.. and that’s a good thing!

There are two elements to testing. Element 1: the clockwork, and element 2: computers as social prosthesis. Clockwork operations give exact answers/results, but this is not always useful, helpful or right. The interface, the product as a social prosthesis does not have to do everything that people do, they can do more or even less; leaving more up to the human!

But most of the time we are trying to make computers act the same as humans do, but they can’t. They can’t infer meaning, context.

“Locus of legitimate interpretation” – basically people’s perspectives affect how they view and evaluate something.. this is leading towards talking about AI, and how people assess how successful a computer or machine is at ¬†performing a task.

“Today’s big problem is deep learning, or more exactly, the handling of huge databases on the internet”; ¬†there is a perception that the more data the better, but this is retrospective at best, and is unlikely to lead to ‘complete knowledge’ being available on the internet.

Ok so got slightly lost.. but I think the gist was that so much data is available on the internet, available online that it’s so easy to trust it. Yet how has the data earnt that trust, it does not represent human knowledge, or if it does its retrospective at best.

Rediscovering test strategy -Mike Talks

¬†He’s started off celebrating 20 years in IT, starting in 1997…. ¬†starting as a tester in 1996 myself this could be interesting.

Video of 1997… now we have a very much younger mike talks! He describes what 1997 mike did testing on windows 95 using floppy disks etc. I remember it well!!

Trying to do testing we did in 1997 in 2017, it just seems wrong. Very wrong. Who in their right mind would insist on such archaic practices in this day and age. Systems move so quickly, technology is changing all the time so why do we keep the same processes!

Test strategy should be concerned with having some idea of how the various parts of the product can be tested. Its not a fine detailed plan of everything. It’s about the general areas we are going to cover, the types of testing we are going to cover, and having a conversation with the customer about this.

“Everyone has a plan until they get punched in the mouth” mike Tyson! ¬†Having a detailed plan is fine, until something happens that invalidates the plan!

Test ideas come as a deluge, and it’s often very difficult to make sense of them. But the best thing to do is write them all down, mindmap them, postit notes whatever, then just run with them. Use your imagination. Don’t just come up with test ideas that you know you can do, that’s safe and will not help you grow. ¬†Cluster those ideas by product areas or some other logical grouping, which will help identify holes. Get others to review, team reviewing, get input and criticism from others. This will make the final result far better.

Mike introduces his test game “test oblique”, must inquire about that. I like games to play with our test group.

Testing at its simplest is we control something and observe how the system responds. This is often thought of as the functional testing, data entry, bva etc. This is where a test strategy comes in. It covers all sorts of things from platforms (cloud, mobile, os) to methods, practices.

Identifying these things  again helps to identify holes and missing things. Read about ideas from blogs, websites (ministry of testing), ask for help, speak to people, go to meetups, chats over coffee.

I totally agree with this. I learn so much from conversations with peers especially at conferences and meet ups.

I also totally agree with bringing in experts to help with transitions or stuff that you are not that comfortable to knowledgable in. The problem is that company/team/people need to be receptive to what the experts have to say. I’ve found some environments that have a “not invented here” culture and think they already know best so it’s not worth paying a large amount of money on experts until they will be listened to!

Future challenges include AI. This is a something I’ve had a bit of exposure to with machine learning algorithms, and these are just hard to test, but something that is so very interesting and challenging, an likely we will have to do more of.

Let’s talk about ethics and software testing – Tobias Geyer

Firstly, he wins for the most amazing suit at Testbash award!
I’ve known Tobias for a few years now, so I’m looking forward to this.¬†

Moving forward do we, as testers, not only have to consider the functional an non functional, but do we also need to  consider the ethicality of features?

It does raise the question whether companies that have had very public ethical issues recently were aware of what they were doing? (Probably) and whether the testers raise these ethical concerns (perhaps). But it is entirely possible (no proof) that everyone was  aware but chose to ignore them.

Gosh, this is a serious talk… but interesting!

Oh and Alex (@alex_schl) gets a shout out!

He gives some time to looking at the IEEE, ACM and ISTQB codes of ethics which are interesting and seem pretty much common sense to me. Then introduces James Bachs paper on ethics as a great alternative to there.

Giving examples of autonomous driving, emissions, Facebook and dating chatbots, not only are most of those issues ethically bad, they are also illegal! I find it hard to believe that the various companies and all involved were not aware of the implications, and often at the higher levels of a company. Testers (and devs etc), can raise issues, and most likely did, but if they want to keep their job, were probably told to be quiet or where actually happy to ignore them!

Thought provoking and we would like to think we would always do the right thing, but until you are in such a situation, you just don’t know what you would do….


Ok, just before we go to lunch, Dan ashby and mark winteringham give a quick talk about the software testing clinic they’ve been running for the past year. Awesome job guys!!!! ¬† “One of the hardest things is talking about testing” sounds like some sort of addiction support group… hello my name’s Chris, and I’m a tester!

Great morning of talks, and some stuff to think about.

Leave a Reply

Your email address will not be published. Required fields are marked *