Twitter | Pretraživanje | |
Peter Burns
Googler on the Polymer team. Web Components true believer.
1.256
Tweetovi
246
Pratim
363
Osobe koje vas prate
Tweetovi
Peter Burns 3. velj
Odgovor korisniku/ci @hankchizljaw
Yeah, waiting to give any feedback at all until the data is loaded for the next page is a huge mistake. Feels great with a fast dev server, feels terrible in the real world where things can take time. It's not a necessary part of client-side routing, but it is a common mistake.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Becca 3. velj
We have started to implement web components at NARA and the stoke level is high!
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 3. velj
Odgovor korisniku/ci @pfau @lmthang i 2 ostali
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Deepak Gupta 1. velj
On Monday, I’ll be arguing before the U.S. Court of Appeals for the Federal Circuit, challenging the federal judiciary’s paywall for public court filings (aka PACER fees). This column by gives a good overview.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 2. velj
Odgovor korisniku/ci @qntm
(seriously though, the Idris book is really great: )
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 2. velj
Odgovor korisniku/ci @qntm
Ah, but with dependent types you could have the compiler prove that x = 7! And all it costs is five to six times longer to make or change anything!
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Sam Saccone 28. sij
With launching to the world today I would love to pull back the covers and tell you a little of how we made it possible to update your Android Phone's operation system through the browser! 📲
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 28. sij
Odgovor korisniku/ci @Slashgear_ @justinfagnani @k33g_org
One way of thinking about it: If you vend a framework's component model, someone using your component needs to know about that framework. If you vend a web component (written with a library or framework), someone using your component doesn't need to know how it was written.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 24. sij
Odgovor korisniku/ci @rauschma
Wild guess: string valued enums were added later, so if you'd realized strict was better here but you hadn't changed numeric enums because it wasn't worth the disruption...
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Zach Klein 15. sij
My 5yo asked, “How big is a wolf?” So I Google it... Google’s first result is the option (in browser!) to place a realistic wolf in the room with us so we can walk around it and see for ourselves. Magical. The closest I have felt to a Young Lady’s Illustrated Primer.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 23. sij
Odgovor korisniku/ci @Rich_Harris
It is not at all hard to change the API without breaking existing users. At the limit, just rename the api. LitElement at least does not rely on its existence, just uses it as a performance enhancement if available.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Robin Ricard 21. sij
My colleague is being awesome, again: is a playground letting you toy with our stage 1 proposal: Record and Tuple. Please try it and give us a lot of feedback but keep in mind it's a stage 1 proposal: it will change!
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Polymer Project 15. sij
An effort 9 years in the making, but we’ve finally made it. WebComponents are supported natively in every major browser
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Stencil 15. sij
Microsoft Edge now supports Custom Elements natively!!
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Sergio Contreras 15. sij
Web Components are supported in all modern browsers: ✅ Chrome ✅ Firefox ✅ Safari ✅ Edge Use a standards-based approach is the natural progression of the web platform 🌐
Reply Retweet Označi sa "sviđa mi se"
Peter Burns proslijedio/la je tweet
Anders Hejlsberg 15. sij
Proud to be one of them too!
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 14. sij
Odgovor korisniku/ci @justinfagnani @slightlylate @lgarron
Seems like that would be better covered by a build-time transform. You'd want configuration params (e.g. only run-time type check a subset of the code, for performance), and you'd want to match the semantics up with your build-time type checker, which means versioning.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 14. sij
Odgovor korisniku/ci @justinfagnani @slightlylate @lgarron
Dart can make breaking changes, the web can't. Leave type semantics to userland.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 14. sij
Odgovor korisniku/ci @justinfagnani @slightlylate @lgarron
This gives the type system(s) time to grow and mature (there's still so much really great type system space to explore!), and there just isn't much value in the browser doing type checking. This would mean that types would be advisory, and people would ship bad types. That's ok.
Reply Retweet Označi sa "sviđa mi se"
Peter Burns 14. sij
Odgovor korisniku/ci @justinfagnani @slightlylate @lgarron
If TS-in-browser happens (and it has merit), the browser should parse the syntax and completely ignore all type information, semantics should be identical to compiling TypeScript to esnext today (with type errors ignored). Advantages: zero build step; debugger can surface types
Reply Retweet Označi sa "sviđa mi se"