You are browsing a read-only backup copy of Wikitech. The primary site can be found at

User:Pablo Grass (WMDE)/conferences/IJS2020 Munich

From Wikitech-static
Jump to navigation Jump to search

IJS2020 Munich was a JavaScript conference in October 2020, run in shared setup with IPC2020 (PHP). Due to the COVID-19 crisis the conference was conducted online (using swapcard).

The conference was structured into two workshop days (Mon, Fri) and three days (Tue - Thu) of talks given in up to three tracks. I attended Tuesday through Thursday.


The program was listed well ahead of time and building a schedule was easy because of a convenient online tool.

This time the International PHP Conference was not only co-hosted but, maybe due to how the digital conference software works, felt all integrated. One could pick talks from either subject in one program, which inspired openness for "front end" and "back end" aspects in discussions.

Topics reached from agile workflows and domain driven design to things more specific to the front end layer like logical CSS properties and PWAs. There were quite a few talks centered around the topic of micro front ends. The security-centered talk around CSP did showed how the layers of the stack are connected and how decision in one can influence the others. The odd "watch me run composer install on that cool library I just found" talk was rare.


The conference was quite international in both speakers and audience but with a significant change from last year: ten native German speakers this time vs five in the previous year (granted, in the talks I respectively selected). Talks were conducted in English and German (very few). Promotion, due to the remote setup, almost did not exist if one did not seek it - comfortable but it remains to be seen if sponsors will be able to afford such setup in the long run. In this vein, the audience count is also noteworthy - swapcard videos offered a viewer count (of unknown precision) which never, even in keynotes, exceeded 70 people. I'll leave it to the reader to draw conclusions from that.

The familiarity of many topics acted as a good indication with respect to "where I (and my employer) stand". The next best thing seems to not have passed us by. The repeated mention of micro front ends, and some valid cases built around them, may even indicate we are slightly ahead of the curve on that. The influx of impulses seemed a bit reduced compared to last year and I would have wished to hear some more experiences with respect to methodology. Maybe the latter is too much asked for a JavaScript-branded conference and a learning is to add a dedicated conference in this area to next year's schedule.

Key takeaways

  • micro front ends have arrived in general awareness
    • the "front ends" terminology is deceptive - they are really meant as the tip of an iceberg which extends across layers of a bounded context
  • dependency-cruiser looks like a great addition to the QA tool box (incl. in TypeScript projects) and should find its way into our default setup
  • invest in CSS logical properties - educate people, don't write any more opinionated (physical) code, especially not in green field projects
  • webpack 5 module federation, with time, will replace homebrew answers in applications consisting of multiple micro front ends
  • still have not found a team where product people actively co-own BDD scenarios in a repository
  • GraphQL will not replace REST APIs (shocker)


In this Userspace:

User talk: