State of the Market

[madbloke]

Surely there will always be the initial run through of the application to record the testing steps before the robot takes over...

Not necessarily. For web consumer, the cutting edge companies are doing continuous delivery - they've defined what the KPIs are important - so when they push to a canary node, if the KPIs flatline they kill the canary. If they don't its rolled out to everyone.

Harder for some to adopt this (financial services, safety critical), and there will always be laggards, but its where its going.

[SussexSeagull]

Not necessarily. For web consumer, the cutting edge companies are doing continuous delivery - they've defined what the KPIs are important - so when they push to a canary node, if the KPIs flatline they kill the canary. If they don't its rolled out to everyone.

Harder for some to adopt this (financial services, safety critical), and there will always be laggards, but its where its going.

For the foreseeable future the vast majority of organisations will have a person look at it first (granted this could be someone in the business and/or a developer as opposed to a specialist tester).

Manual Testing has supposed to be have been replaced by automated testing for the last 30 odd years (there were automated test tools on mainframes) and while the technology has come on leaps and bounds the thing stopping it is often management only think short term and don't want to pay for it.

[eek]

For the foreseeable future the vast majority of organisations will have a person look at it first (granted this could be someone in the business and/or a developer as opposed to a specialist tester).

Manual Testing has supposed to be have been replaced by automated testing for the last 30 odd years (there were automated test tools on mainframes) and while the technology has come on leaps and bounds the thing stopping it is often management only think short term and don't want to pay for it.

+1. automated testing works if you start doing it on day 1 and add to it as the system is developed further. I've never seen someone suggest retrospectively adding automated testing and not running away when they see the price....

[nomadd]

+1. automated testing works... ...running away when they see the price.

Not worked on a project for at least 10 years that doesn't use automated testing.

And all of the tools are open-source, so the cost inconsequential.

And the jobs that are sitting in my inbox include lots of "Test Automation Engineer Contract" (even though I'm not a tester.)

[MrMarkyMark]

Not worked on a project for at least 10 years that doesn't use automated testing.

And all of the tools are open-source, so the cost inconsequential.

And the jobs that are sitting in my inbox include lots of "Test Automation Engineer Contract" (even though I'm not a tester.)

+1 the Mrs is a Test Manager, now successfully contracting .

She says it is all about automation these days and she does know her stuff

[eek]

Not worked on a project for at least 10 years that doesn't use automated testing.

And all of the tools are open-source, so the cost inconsequential.

And the jobs that are sitting in my inbox include lots of "Test Automation Engineer Contract" (even though I'm not a tester.)

The tools are, the cost of writing the script the robot follows isn't....

[MrMarkyMark]

The tools are, the cost of writing the script the robot follows isn't....

Its like all things, if set up right from the start and the value is recognised, happy days.

Unfortunately, trying to get most clients to think horizontally and vertically is a massive problem, i.e. if you do it this way you will save £££££, in the future.
You can wax lyrical about stucture, process, graceful cascading changes through systems and sub sytems etc. etc. and usually all you will get is "How quickly and cheaply can it (the immediate box checked requirement) be done"

[LondonManc]

Its like all things, if set up right from the start and the value is recognised, happy days.

Unfortunately, trying to get most clients to think horizontally and vertically is a massive problem, i.e. if you do it this way you will save £££££, in the future.
You can wax lyrical about stucture, process, graceful cascading changes through systems and sub sytems etc. etc. and usually all you will get is "How quickly and cheaply can it (the immediate box checked requirement) be done"

That's partly down to the budgetary silo mentality that comes from on high, leading to the thinking that they won't want to finance someone else's future gain if it looks like they're overpaying for their own project.

[MrMarkyMark]

That's partly down to the budgetary silo mentality that comes from on high, leading to the thinking that they won't want to finance someone else's future gain if it looks like they're overpaying for their own project.

Exactly, which is why in IB they would rather spend their time having budget wars and trying to kill each other rather than collaborate on a wider project

[SussexSeagull]

I am the first to agree that more companies SHOULD develop automated testing from the get go in a project.

In reality I imagine if your CV suggests automated testing you are more likely to end up on one of those projects and certainly there does seem to be a lot of of automated roles going but it still needs to be manually tested before automating and a lot of companies don't want to pay for it.

Some companies still use spreadsheets for test planning and issue management.