This is wonderful. I can't tell you the many times I've edited multiple files just to fetch one new piece of data for a view. When you do something manually as a programmer instead of the computer automatically doing it for you, you know something is wrong with your architecture. Facebook has offered us a better architecture.
Curious to how messy/efficient your code gets with GraphQL at scale? (besides at FB, I'm sure they have a highly specialized system). I just imagine shit getting crazy with lots of views and data dependencies.
GraphQL looks great. That's a problem that always bug me. That makes totally sense now. Not sure about the function style for params. But the idea is simply great.
I like the way the data dependency of a component is colocated with the UI description. However, it's still repetition. DRY: couldn't the framework *infer* the data dependency by traversing the node tree generated by the JSX in render()?
I just hope it will work seamlessly when rendering on server with renderToString as well. Otherwise it would be a huge upset. Also the speakers haven't said a word about how server side that actually would be making queries to a real database would be implemented. That's also very interesting.
Haven't yet looked at the Relay part, but The graphQL seems a bit odd... To specify the attributes you want in your object you still need to know which attributes are available to you (aka look at the docs). So why nitpick the attributes, just pass the whole object. And let the children take what they need to.
I swear, it's like nobody notice history is currently in the making. Everyday people have the power to just simply look in the blockchain & most people aren't even paying attention. It's just simply crazy 😅 I'll be shock if nobody takes notes of this years from now.
Interesting idea, but it would have been much nicer if GraphQL would have been in pure json. Now we have javascript, jsx and graphql all in the same file. gets messy.
This is wonderful. I can't tell you the many times I've edited multiple files just to fetch one new piece of data for a view. When you do something manually as a programmer instead of the computer automatically doing it for you, you know something is wrong with your architecture. Facebook has offered us a better architecture.
Developer1 : "Are your updates optimistic?"
Developer2 : "I hope so!"
Take the principle of Relay to your children! Children declare what they need and often not what they parents think they do :)
Curious to how messy/efficient your code gets with GraphQL at scale? (besides at FB, I'm sure they have a highly specialized system). I just imagine shit getting crazy with lots of views and data dependencies.
GraphQL looks great. That's a problem that always bug me. That makes totally sense now. Not sure about the function style for params. But the idea is simply great.
I like the way the data dependency of a component is colocated with the UI description.
However, it's still repetition. DRY: couldn't the framework *infer* the data dependency by traversing the node tree generated by the JSX in render()?
can't wait for GraphQL and Relay surely the best thing
Happy birthday :) 9:15
I just hope it will work seamlessly when rendering on server with renderToString as well. Otherwise it would be a huge upset.
Also the speakers haven't said a word about how server side that actually would be making queries to a real database would be implemented. That's also very interesting.
Looks like there is some "juggling" pain managing the queries, but the query validation feature should help.
Haven't yet looked at the Relay part, but The graphQL seems a bit odd...
To specify the attributes you want in your object you still need to know which attributes are available to you (aka look at the docs). So why nitpick the attributes, just pass the whole object. And let the children take what they need to.
not clear whether this is server-side or client-side rendering from the start
I wonder if they knew this would be used with something like web3 back then?
I swear, it's like nobody notice history is currently in the making. Everyday people have the power to just simply look in the blockchain & most people aren't even paying attention. It's just simply crazy 😅
I'll be shock if nobody takes notes of this years from now.
thank you for the awesome video , how can i get a full reference for all the fields and edges ..etc
We gotta talk. Who is the real superman?
Interesting idea, but it would have been much nicer if GraphQL would have been in pure json. Now we have javascript, jsx and graphql all in the same file. gets messy.
Anders Gunnarsson Oops, looks like they are already making this better adding GraphQL.js.
7:20 _Woo_