Presentation is loading. Please wait.

Presentation is loading. Please wait.

An empirical process to get an assessment of the quality of the quality of a system Never exhaustive  one can never prove the system will work Comparison.

Similar presentations


Presentation on theme: "An empirical process to get an assessment of the quality of the quality of a system Never exhaustive  one can never prove the system will work Comparison."— Presentation transcript:

1

2 An empirical process to get an assessment of the quality of the quality of a system Never exhaustive  one can never prove the system will work Comparison against expected results  requires correct preparation of expected results Assumes a sequence of events  what if this sequence is not always followed ?

3  Forms Static : review, inspection,… Dynamic : execution  Functional Unit – examples : input validation, calculation, formatting System – examples : website with interfaces stubbed out, website running against test customer database Integration – full functioning system – as close as possible to production Conversion – when you upgrade Acceptance – the customer signs off  Non functional Scaling Duration Security Performance

4 Project Initiation4% Project Management10% Training12% Communications4% Analysis9% Design10% Build25% Test17% Roll Out8% Evaluation1% For larger – 10 + people projects For smaller projects – testing cost increase

5

6

7 Unstated mental models It compiled so it will work I hate MS$&^*T stuff It is written in PHP so it works Show that the website works Everyone has broadband …  Imperative that you break this mental block

8 Repeat/recreate the error Other browsers, OS-s,… Different sequence Different data – input and DB-s Different HW (video, memory, disk,..), SW Different connection speeds Different connectivity paths …  First assumption has to be: Your application has an error

9 The “owner’ is way to invested into the system to truly test it Main focus will be to prove the system works – not on trying to break the system To aware of the many assumptions that went in the design/implementation Quite often the system is too large to be tested by an individual; there are too many variations

10  Create unit tests for complex lower design issues  Test driven development Helps clarify the requirements/design Helps automate  Document larger tests in a form that works for you scenarios, use cases in word document, excel,… Document data, steps and outcomes If possible automate Do not forget the exceptions, errors,..  Execute these tests in an environment that is not your development environment; maybe everything you execute them use a different browser, OS  VM-s are your friend  If none of your tests fail: find someone else to test

11  Focus first on the common scenarios  Then focus on common exceptions  Then focus on rare exceptions  Then…. MONKEY test

12

13  Continuously test – if you wait till the end – it will be very, very, very tedious  If you find lots of errors in a certain subsystem, and they keep coming back (ie you can not seem to fix it), toss it and rewrite it  It is probably the most undervalued activity to ensure you site works  There are truly testing geniuses – if you find them try and hire them.. Able to perform repetitive tasks with high focus Have fingers that make code break Have an incredible attention for detail Are very productive in finding high quality errors

14 Questions…..


Download ppt "An empirical process to get an assessment of the quality of the quality of a system Never exhaustive  one can never prove the system will work Comparison."

Similar presentations


Ads by Google