UI performance optimisation plan

[dkhurshudian]

I've done some optimization in #544 but there is always something that could be done better. This issue will serve as a place for performance-related suggestions and discussions.

[dkhurshudian]

• blueprint icons has font-based and svg-based implementation each is ~100kb gzipped, our project uses both, we need to get a rid of font-based one. Additional keep an eye to this issue A way to only bundle icons you actually use palantir/blueprint#2193 to load only required icons.
• Implement dynamic import for translations
• Embed metadata to html

[pudo]

Good stuff. I've spent a little bit of time thinking about (3) and short of introducing a node server process I can't really think of a good way for doing it. It's also locale-dependent, so it would need to be able to either do content negotiation, or get much larger.

[dkhurshudian]

Good stuff. I've spent a little bit of time thinking about (3) and short of introducing a node server process I can't really think of a good way for doing it. It's also locale-dependent, so it would need to be able to either do content negotiation, or get much larger.

Actually I don't think embedding is the best solution, this is kind of reminder for us to keep in mind that we have to do something with this request.

[pudo]

Aye. In general I feel like the bulk of our performance issues is now back-end related, so I'd propose we focus on doing speed-ups there (already started), and do other work on the front-end.