Martin Splitt (@g33konaut) is a senior software engineer at Archilogic but prefers the term “Open Web Cheerleader”. He contributes to a lot of open source projects. He gives quite a few talks and workshops, writes articles and blogs at 50linesofco.de.
At the Confoo conference in Montreal, Canada, Chris Heilmann took some time to ask Martin a few questions about performance of browser and client side rendering.
If you prefer to have an audio version, you can download it here
Here are the questions we covered:
- What do you consider the main issue with client side rendering? There seems to be enough information about it available and yet people make very obvious mistakes.(00:04)
- Do you think there is an issue that we find fast workaround fixes for performance problems of browsers and become dependent on them? Would it make more sense to wait for browsers to fix those issues? (01:46)
- Developer tools offer a lot of intelligent insights into performance issues and yet we see problems all the time. Are our tools too hard to use? (03:40)
- Is there an issue that tools that give your product scores give you a false sense of accomplishment? Does a high score result in real user needs? (05:10)
- How useful is simulation of different devices and connection speeds in developer tools? (06:49)
- On the one end of the spectrum we talk about low end phones. On the other we talk about AAA gaming in desktop browsers. Do we ask too much of web developers to understand different contexts and learn about how graphics work on a hardware level? (09:12)
- We seem to be very excited about writing polyfills to work around a current state of a browser or OS problem. We’re not as good in removing them. Did we forget about the “stop-gap solution” nature of polyfills and libraries? (12:25)
- A lot of web innovation and performance optimisation in browsers came from experimentation with 3D rendering of games in browsers. Do you think new web developers should understand more of that world? (14:04)
- There seems to be a massive gap between information that is available on performance and what gets applied in real products. What is your process to find great info to apply and discard what it noise? (17:31)