Hi,
Interested in what automation strategies / approaches people have come across, specifically in larger programmes of work, for example;
- a dedicated environment for running automation in (environment gets restored / reset to a baseline "gold" data set prior to an automation pack being run) or run against the QA system / seperate client
- use of "time travel" software in running the pack (eg Time Machine) to do time based testing
- in terms of the actual implementation of the automation of test cases - the development of generic re-usable components (say generic vbscript test cases in the case of QTP -v- specific vbscript test cases) against specific components
- how often is the pack run? Once a week, only at specific points in the project?
- run the whole pack or selected parts of the pack?
- primarily a project tool and only secondarily a support tool?
- pack maintenance strategy for programmes / waves (BAU and project packs (maintenance of the BAU pack)) - goes live with the project
- expectations around automation (accelerator to reduce test cycle time / get wider coverage in a shorter period of time) - did it work?
- develop and stabilise / baseline manual test cases or straight to automation?
What sort of problems have people come across where automation has been implemented and what has been the root cause of these issues eg....
- data – coping with transactional data burn
- ring fencing of infrastructure
- over-complication of scripts
- ring fencing of resources to the automation operation
What has been the set up where you have seen this run / work well? So, for example, I have seen this run well in where there is dedicated infrastructure reset with a standard data set and automation scripts are developed / enhanced as part of projects, but realise this is probably at the extreme end...how little can you get away with and for it to work well and add value?
Thanks

