Talk in English - US at PHP TEK 2025
Track Name:
Wrigley Field
View Slides: https://presentations.garfieldtech.com/slides-serialization/phptek2025/
Short URL: https://joind.in/talk/f6c46
(QR-Code (opens in new window))
In 2021, I was tasked with building a new serialization system for TYPO3. The result was a trio of new, general-purpose libraries that demonstrate the power and flexibility of the modern PHP language: FP, AttributeUtils, and Serde.
In this session, we'll start with an overview of the plethora of potential PHP processes for handling serialized data, and show their pros and cons. We'll then transition into talking about Serde and what goes into a modern PHP 8-centric serialization framework, before showing novel ways to use serialization tools.
Along the way, we'll touch on the power of PHP attributes, and how AttributeUtils makes them even more powerful.
Comments
Please login to leave a comment
Good talk that was very technical about the implementation of Serde and how serialization/unserialization works internally. Larry is extremely knowledgeable about the topic, which showed in his expert subject matter talk. The code snippet slides were informative and helpful, and Larry was well prepared for the talk.
I would have liked to have had more context on when I should use serialization and real world use cases. The mention about API versioning was a good real world use case when mixed with Attribute scopes.
Overall a great talk and I learned new things for me to take away and try out.
this talk was very information dense and got that information across well
Really interesting talk! I appreciated both the overview of different business/operational requirements which call for particular solutions when it comes to serialization. The level of technical depth was good while still being grounded in practical examples and scenarios. I haven't used Serde but I left thinking about the different implementations where it would've come in handy. Admired the API design which revolved significantly around attributes which help position definitions in PHP, etc. as the "source of truth" for serialization rather than maintaining multiple schema definitions which must be maintained in lock-step manually.