Quality Assurance And Software Testing

Etin,

Glory Hallelujah!

At first it was a call for graphic designers and now you are calling for testers.  Hopefully, this is a sign of good things to come with respect to software development in naija, in general.

With respect to the Northern Rock (NR) incident, I will beg to differ though, in terms of definition of quality assurance.  Allow me to indulge you please:

I have been following the NR things for 3 days now.  To cut to the chase, customers made a run on the bank for their money.  In the real world, this was seen in the long lines outside the banks.  Online, this was seen in terms of jamming the site with excessive hits which resulted in the site hanging/crashing.

Now, sure, it could b argued that lack of testing was a cause of this.  I would beg to differ.  When it comes to money and life, the software that supports such systems are referred to as "mission critical".  This means that there are many fallbacks and redundancies put in place to ensure that if something goes wrong, something is in place to detect, catch and alert to the situation.  Also, such systems have a less tolerable level of defects.  In line with this reasoning, I will say that I am sure that the developers carried out the necessary levels of stress and performance testing before launching the site.  This of course would not take into consideration a run on the bank.  Why?  because this seldom happens.  If you beg to differ, I will then ask this question: would you also blame the civil engineers who built the bank building for not conducting enough testing of that structure to ensure that it would handle an event where all customers lined up outside the bank for the sole purpose of asking for their money back?  I don't think you should.  There is an acceptable level of failure built into all things.

Look at it another way - the day the Lewinsky report was published, CNN's website also crashed.  Why? because it was handling more requests than it could bear.  Of course,  one could argue that they could have bought more servers in anticipation, but truly though, how many servers are enough?  That's just like saying, "let's build more roads to help ease traffic".  We all know that once an additional lane is put in place, those who stopped driving will end up driving, thus clogging up the roads yet again.

To me, quality assurance/testing has to do with the functionality of software.  So, given the First Bank website, if I click on "sign up for internet banking", i expect to be taken to a page that allows me to do so.  Currently, I get a "page not found". In this case, enough testing wasn't done on the site and consequently, the quality of the software is very poor.

I am not an expert in the art of testing, but have been dealing with testers for a long time and also took a few courses so as to understand and appreciate what they do.  What I have written, is based on my experience dealing with those in this discipline.

As always, my 2 cents.

Etin,

Your points are very well noted.  However, IMHO and with all due respect, I think you have missed some points in that article.  Let me point them out, if I may:


Here is what Mr. Ilube says about the above:
"Well, I can say from experience that building computer systems that can cope suddenly with a completely unexpected burst of perhaps 10 or 20 times their normal processing volume is notoriously difficult to do, but by no means impossible."

Impossible, no.  "Notoriously" difficult, yes.

I seriously doubt the guys at NR didn't have anything in place.  If they didn't, they wouldn't have been that successful.  Perhaps the process they had in place was inadequate.  Again, Mr. Ilube says:
"There are sound economic reasons for taking this third approach as major spikes happen rarely, if ever."  Note "as major spikes happen rarely, if ever".  If there was no precedence, how would they know?  Now that there is a precedence, other banks will take note and hopefully do something about it.

My last statement above is supported by Mr. Ilube parting statement:
"That Northern Rock IT director is the only IT executive in the UK online bank world today who has been at the sharp end of a digital run on the bank.

That direct experience is invaluable - and whatever happens to the rest of Northern Rock's executives, he will be in high demand to share his experience across the finance sector and make the whole online banking world safer."

Again, note the first paragraph.

I have indeed read the entire article, which was well-written by the way and by no means am I cherry-picking to support my statements.  I am happy that you dug up this article and shared it with us.  I have indeed learnt from it myself.

Banks, like other businesses, are around to make a profit.  In this age where investors demand continuous and consistent growth, businesses have to cut the fat in order to squeeze out an extra penny.  The result is that unexpected things like a run on a bank will probably not be addressed until it happens.  If they do decide to do it, who do you think will pay for it? The  bank? Mba.  It will be you via higher "service charges".  In the end, no business pays for anything, including taxes.  We as consumers pay.

Mr.  Ilube wrote well.  However, he is engaged in what we in the states call, "Monday Morning Quarter-backing".  In Nigeria,we would say, "Medicine after death".  Either way, again, well-written article.  Let's hope that other online businesses heed the warnings in general and that nigerian-based companies learn that quality assurance and testing are perhaps the bed-rock of any successful software application.

Just imagine if a plane were the first bank of Nigeria website - I click the "internet banking" link and it doesn't work.  Replace this with the pilot engaging the wheels in order to land the plane.  Now, imagine the wheels not popping out of their compartments.  Close your eyes and imagine what would happen then 

As always, my 2 cents.