WebCamp is technology oriented conference for developers and designers. I would like to add software testers into their pitch. Testival community is one of the supporting organizers and I saw yesterday a lot of familiar software testing faces.
As every professional conference, we got at opening word all important informations and reminders. New things is GDPR notice: You are informed that you will be photographed and filmed. If you do not agree, do not attend the WebCamp (In cooperation with EU bureaucrats, in Croatian Uhljebi). There is UnConference track that failed on first day. I will provide suggestion how simply improve that track.
Keynote talk: Taking CI and automated testing tests seriously by Alen Ladavac
This is not best put title for the talk, because it was ok 10 years ago, but we are living in 2018, 95% of audience is working for privately owned tech companies that generate income with taking CI and automated testing seriously. Maybe not all of those companies do not put exploratory (aka just testing) software testing seriously, but this is also changing giving 60+ attendees at Testival 2018.
Alen is CTO for CroTeam, Croatian game development studio (Serious Sam). Question is how to compete with game quality of bigger international studios?
They have unit tests in c++ where they test low level code features: strings, math, memory utilization and object serialization. Test suite is triggered in Jenkins. Trouble is that Jenkins workspace size is 200GB, so they can not use clean workspace feature in timely manner. They have 90 jobs for different target platforms. With such big workspaces, cloud infrastructure is not an option ( bandwidth price), so they created local environment based on cheap hardware with fast interconnections.
Then we came to part they call game “Functional Testing”, which is actually exploratory testing. They use gamers for that purpose. They play games and is expected from them to report issues. Trouble is that they are gamers and they do not know how to do that and they do not care about it :) So croteam created in game bug reporting tool. When gamers spot an issue, they “Marked it” and describe it. E.g., they mark a stone that is missing a shadow. Mark is important, because it is directly connected with code or code asset (e.g. 3d model).
There is a Bot with purpose to complete the game in new release. Bot has fast forward option and in that mode can complete The Talos Principle in 10 minutes (versus human gameplay of 10 hours). It is important to state that croteam is aware that Bot can only confirm game completeness, not absence of game bugs.
In Q&A I asked about their Quality Criteria. Here it is:
- platform requirements
- game completeness
- gamer opinion
- break earth physics is low priority risk
Here are my notes:
Originally published at tentamen blog.