DrupalCon Seattle Day 4 Recap: Amazee Sessions

DrupalCon been a very productive conference so far. The first two days of pre-conference sprinting resulted in fixing the testing pipeline of the webpack module, a prototype for using Drupal as a datasource for Gatsby using GraphQL instead of JSON:API and even solved an unexpected issue for the devel module to provide a way to load dump an entity with all its references embedded inside. You can read more about these solutions here.

After a good Wednesday afternoon, breathing in the calming air of Seattle, Thursday came and it started for me with a breakfast with Victor. He ordered his favourite meal - the Fresh French Croissant and after a short meal, we headed to the venue.

I decided to start the day with the Web components BoF led by Salem Ghoweri. Pega systems use web components a lot and it was interesting to learn about the advantages and pitfalls. IE 11 seems to be the biggest obstacle, especially when it comes to the shadow dom. Polyfilling it is very expensive computationally, so what they did is to actually ditch the shadow dow in browsers that don’t support it natively. In general, looks like web components are getting traction and IE is the main problem, so same as 5 years ago.

The next spot was the GraphQL 101: What, Why, How session by my friend Maria Comas. It started with a brief history of the query language followed by its definition.

I learned that the reason to build the GraphQL spec at Facebook was because of the need to find a tool that is powerful enough to handle everything Facebook does while staying simple enough so that it’s easy to comprehend by the product developers. The thing Maria likes most about GraphQL is that it is a tool that changes human behaviour.

The session finished on time and there were hardly any questions so we had time to get the best seats for the next Amazee session:  CSS-in-JS and Drupal sitting in a tree... by John Albin Wilkins. On our way there Victor, grave as always decided to make use of his hipster camera and take this photo of me. No comment.

John is a natural so his sessions are always entertaining and packed with great content. In this one, he compared 3 different ways of doing CSS-in-js which are:

The session summarized the two years of our adventures with the topic while doing both fully and progressively decoupled projects. TLDR; John recommends CSS modules, mostly because it’s the only tool that makes it possible to share the styles between javascript and Drupal. If you’re interested in this topic I would encourage you to check out the recording for the reasoning and lots of interesting details.

After that, I headed to the Considerations of Federated Search and Drupal session by Adam Bergstein. The ability to find content that originates from many different websites is a hard topic which is required by the enterprise clients quite often, so I thought it might be interesting.

Nerdstein started with a high level, generic overview of the system. The structure is similar to what we have in Drupal migrations. He recommended using scrapy. It’s a tool from the python ecosystem which is great because there are many great data manipulation and natural language processing packages. Scrapy also has many destination plugins, e.g. for elastic search, so it’s easy to insert data directly into the search index.

Next, there was lunch and an unexpected booth on the way there - a box with cute, fluffy creatures.

I’m not really sure how they ended up there but they definitely made lots of people happy. Here are some photos. Hopefully, they will make you happy as well.