Friday, 17 September 2010

Blink Comparator Testing vs Blink Testing

I saw an article on the STC the other day about a competition to go to the ExpoQA in Madrid.

I was immediately interested as I backpacked around Spain years ago and Madrid is a great place to spend time, but I digress...

I scanned the programme (top-down), got to the bottom and started to scan bottom-up and immediately saw an inconsistency. Then I started wondering why/how I'd found it. Was it a particular type of observation, test even?

I realised that when I started scanning upwards I had some extra information and that was why it popped out. I twittered a challenge to see if anyone else spotted anything - at the time I added the hashtag #BlinkTest. After further reflection I wasn't so sure that was correct - I started thinking about the blink comparator technique that astronomers used once upon a time.

It seemed like a blink test (that I've seen described and demonstrated) - but after some thought it seemed closer to a blink comparator test. Was there a distinction, what was lying behind why I'd spotted the inconsistency? Was one a form or subset of the other?

Could Blink Testing be a general form of information intake (and subsequent decoding) and Blink Comparator Testing being a special form where the pattern, legend or syntax is specified and the scanning is aimed at spotting an inconsistency. Maybe. Hmm...

Blink vs Blink Comparator?

I'm going to think of a Blink Comparator Test as being one that takes a particular type of priming (concious) - here's a pattern (or patterns) and this is used to compare against the observation - maybe for the absence or presence of the pattern/s.

I'll think of a Blink Test as being an observation without priming (subconcious) - although there will always be some priming (from your experience), but that's a different philosophical discussion - and it's the subconcious that waves the flag to investigate an anomaly.

Of course, both can be used at the same time.

Why distinguish?

It's about examining why I spot an inconsistency and trying to understand and explain (at least to myself) what process is going on. So, why the interest in understanding the difference? For me, that's what helps get to the bottom of the usefulness and potential application, and indeed recognising the technique 'out in the wild'.

This started out as looking at an anomaly (maybe even a typo) in an article, and now I have an addition to my toolkit - I probably already had it, but now I'm aware of it - and that means I have a better chance of putting it to good use. I can see uses in documentation, test observations and script design (to aid observation/result analysis). Cool!

Oh, the inconsistancy I spotted was the use of (Sp)* in the time schedule, which wasn't in the legend. Simple stuff really producing all that thinking...


  1. Hi Simon,
    Good Post!
    I have been thinking about something similar for a while.
    What I was thinking about was Quick scanning and skimming and how I do this a lot. Are these blinks?, coupled-blinks, integrated blinks? prolonged blinks? or not blinks at all? How about piped blinks?

    I have been thinking about this a lot more and when you first get something to test are you blinking lots loads more data to assimulate? or is that not a blink at all? I have been practising recently with blog posts (Eek...time pressures..)

    Or am I doing blink comparator testing when I get a new version of software to test, i.e. one first look(blink) tells me if something could be different?

    I think there are lots of things happening in a 'blink', can those things be seperated? or is it all about the application of the blink?

    One thing that has been niggling in the back of my mind is the hourglass illusion picture, I'm not sure why or whether it has a bearing on blink* testing..I think that's more to do with perception, and I think blink* tests can be super quick tests where you might be able to spot inconsistencies super quickly..that may give you pointers for investigation...


  2. Hi Peter,

    There's a thread here that I'm going to follow-up - and that's about the difference between subconcious and concious issues that we see. For me the "Blink Comparator" is a concious check or investigation - there's a pattern as a reference/input.

    I'm currently reading "Gut Feelings" and I see many uses for understanding how our intuition works (our in-built heuristics) and by understanding them (the subconcious part) we understand better our concious activity.

    To be continued...

    Separating the things that happen in a 'blink' is an interesting challenge - maybe they collectively form a different pattern.

    I think this mechanism (whether blink or blink comparator testing) happens very quickly - so the inconsitency jumps out quickly or maybe later when your subconcious is doing some post-processing.


  3. Hi Simon,

    You raise a very interesting point about the legend bug in the ExpoQA programme. Like me you would not have consciously gone out of your way to learn the different designations for translation; your subconscious mind would have picked up on it though and without realising it influenced you as you scanned up.

    As testers I think it is really important to remember that sometimes we will 'miss' things because we have scanned down a page so our brains will not have been triggered by extra information further down the page. As soon as we start going back up the page or re-reading the page, almost without realising it, our context has changed.

    I find this stuff fascinating... Great post...