Originally posted by westtester
View Post
Project was doomed from the start. Most of the tests were written in Spain by Sabis, many in Spanglish and some were just requirements pasted into excel sheets and passed off as test cases. Majority of the testers were long-suffering branch staff pressed into service with the promise of boozy nights away on company expenses, plus a slew of contract testers from KPMG. Incentives included test team of the week ( £25 Sainsbury's voucher) and test case prize draw (£5). Souvenir pencils emblazoned with the slogan 'Be a test ninja' were considerably less well received.
Project was Agile in name only, waterfall wedged into sprints. Daily progress meetings demanded x number of tests to be executed every day regardless of availability of functionality. Planning went out of the window as testing was focused on whatever happened to be working that day.
Quality of code from Sabis was dreadful, defects long thought fixed would reappear unexpectedly causing days of delays as it killed important processes stone dead. Documentation was lacking and there were far too few business experts and analysts to cope with demand. An in house test case management system built on a single licence HP ALM database was flaky too. Cheap though I guess.
Senior management would visit on occasion and be shown happy-path demos whilst whingers were shooed away. Contractors seeing the coming storm began to quit before TSB panicked and ordered KPMG to up the notice period from one week to four mid contract.
We all used to joke that there was no way this would be ready in 2019 let alone April, but somehow a decision was made to go ahead anyway. Never though things would go quite so badly. Scouring Twitter, you can see screenshots of numerous Java exceptions, references to an open source Netflix load balancer and all manner of bizzare messages which never showed up during testing. It'll take months to sort out I reckon, how IBM are going to unpick it I can't imagine.
Project was Agile in name only, waterfall wedged into sprints. Daily progress meetings demanded x number of tests to be executed every day regardless of availability of functionality. Planning went out of the window as testing was focused on whatever happened to be working that day.
Quality of code from Sabis was dreadful, defects long thought fixed would reappear unexpectedly causing days of delays as it killed important processes stone dead. Documentation was lacking and there were far too few business experts and analysts to cope with demand. An in house test case management system built on a single licence HP ALM database was flaky too. Cheap though I guess.
Senior management would visit on occasion and be shown happy-path demos whilst whingers were shooed away. Contractors seeing the coming storm began to quit before TSB panicked and ordered KPMG to up the notice period from one week to four mid contract.
We all used to joke that there was no way this would be ready in 2019 let alone April, but somehow a decision was made to go ahead anyway. Never though things would go quite so badly. Scouring Twitter, you can see screenshots of numerous Java exceptions, references to an open source Netflix load balancer and all manner of bizzare messages which never showed up during testing. It'll take months to sort out I reckon, how IBM are going to unpick it I can't imagine.
Comment