Interview Question (1 Viewer)

A

AnthonyGerrard

Registered User.

Local time

Today, 08:14

Joined

Jun 11, 2004

Messages

1,069

• Oct 10, 2017

• #1

QA and UAT?

I work in a small team. I test my development.
It goes onto a master copy. Someoen else tests the development manually.
It then goes out as a test version to the users.

It gets signed off.

What is best practice in access development - what do others do?

 

Ranman256

Well-known member

Local time

Today, 03:14

Joined

Apr 9, 2015

Messages

4,339

• Oct 10, 2017

• #2

same.
test yourself
then others test
QA test
deploy

 

C

CJ_London

Super Moderator

Staff member

Local time

Today, 08:14

Joined

Feb 19, 2013

Messages

16,607

• Oct 10, 2017

• #3

I agree with Ranman, that is the order of things for any software development. But depending on the environment in which you are working but you might want to google about test scripting - an overview here

https://en.wikipedia.org/wiki/Test_script

point is the testing should not be random but proceed in an organised way to ensure the new app/amendment meets the original design specification together with any change requests - whether this is done manually or using software is up to you.

 

scott-atkinson

I'm with the Witch.......

Local time

Today, 08:14

Joined

Aug 31, 2006

Messages

1,622

• Oct 10, 2017

• #4

Everything that I develop, I stress test myself first, then I give it over to a selection of people to UAT, on in particular always breaks the first draft as they want to try an do something that the system isn't designed to do..

I then incorporate or fix the broken functions and re-issue..

For me the UAT is pretty continuous, I'm always changing and amending the operation of software as new business needs arise for it's use..

 

Minty

AWF VIP

Local time

Today, 08:14

Joined

Jul 26, 2013

Messages

10,366

• Oct 10, 2017

• #5

Ditto with Scott. Develop, step away, then self test, hopefully trying to do things in a "strange way" that makes no sense. Then UAT.

I too have 1 or 2 users that can break anything given 10 minutes of "fiddling" time.

 

M

Mark_

Longboard on the internet

Local time

Today, 00:14

Joined

Sep 12, 2017

Messages

2,111

• Oct 10, 2017

• #6

Step one is to determine boundary conditions that you need to test for:
This includes things like "What happens if a NULL gets into the data even though this should never be NULL" or "What happens if I have a negative in a field that should never be negative".

Making a list of these types of issues can help you eliminate some issues (or know what is causing them) before your users EVER get a chance to ruin your work. This will also help when you are displaying results that, in theory, shouldn't be happening if things are working correctly, such as when your sales people decide they will take a loss on a sale "Because they are a good customer".

 

A

AnthonyGerrard

Registered User.

Local time

Today, 08:14

Joined

Jun 11, 2004

Messages

1,069

• Oct 23, 2017

• #7

Cheers All.