Reading time approx. 3 minutes

The plan

Around 14 hard-working helpers gathered at our premises to carry out maintenance and update work. Among other things, the automatic provisioning of the servers was to be revised, a development environment for typo3.org was to be created and the typo3.org website was to be updated. A special task fell to Michael Lihs and me: to evaluate a commenting system that has similar functions to disqus but can be self-hosted.

The challenges

Each of the teams faced a major challenge: the TYPO3 server provisioning was implemented by people who were no longer actively involved. The server team therefore had to familiarize themselves with the existing development and decide whether it should or could continue to be used or whether a new basis should be set up that could be more intuitive for new team members to use. This discussion was started several times over the weekend until we finally decided on a mixed solution.

The typo3.org team had two major questions: How to provide developers with an up-to-date version of the website without transferring all live users and all TYPO3 extensions? And how to split a website with older technologies so that one part (which is to be updated later) still runs on the existing system and the other part is available on newer technologies. During these discussions, we were able to provide some input based on our experience from customer projects and help to plan the corresponding concepts. The implementation will take place in the next step.

The work on the commenting system had to deal with completely different problems. The plan was to set up a server with the discourse tool. Michael and I then wanted to add features such as a login via typo3.org and adapt the layout to TYPO3 colors. Unfortunately, we realized that some crucial features could not be implemented without further ado. Although we contacted the developers behind discourse so that we could get these features implemented, we also evaluated other similar commenting systems. Unfortunately, we noticed that although many systems do implement some functions from disqus and implement them well, they never fulfill all the requirements needed on typo3.org (single sign-on of typo3.org users, comment field under a post (no redirection to the comment tool), customizable layout of the integrated comments). There are now three options for how to proceed - either we set up our own commenting system, we use the architecture of discourse and set up our own frontend based on it, or the TYPO3 community is once again faced with the question of whether disqus might be the best solution - despite the third-party hosting. Because since the question was last asked, new tools and processes have been established in the community, which mean that the problems with disqus are simply no longer a problem.

Conclusion

Despite all the work we had to do, there was still enough time for good discussions. This resulted in some knowledge transfer, which (hopefully ) allowed everyone to take away something useful. For me personally, a highlight was a conversation about workflows for the Mac task starter Alfred, which prompted me to search the corresponding repository for useful extensions. For me, this sprint was the first time I participated in an open source community outside of simple code changes. It was definitely a lot of fun and despite all the challenges, there was a relaxed atmosphere the whole time. You could feel that people were very motivated to advance "their" systems and TYPO3 - and this motivation was very contagious. I am glad that I was able to be there and will definitely take part in further events.