Amazing, so it would like to be able to run effect on heterogeneous environment, for exemple in backend/orm/postgres we work with fastapi, but effect can't work on python ?
Hmmm, I wonder ... while Effect doesn't (yet?) run in Python, you could at least get some of the benefits of unifying the schema across front-end and back-end by serializing the schema to JSON Schema and sharing that schema with your Python backend. You'd need some method for going from JSON Schema to your database schema. Ultimately, this problem is begging for some unified tooling.
1. what happens if the software on one of the devices hasn't been updated yet and the new schema is not supported on that device? 2. how does effect/schema compares with things like Lenses from Ink&Switch or grafana/thema?
> what happens if the software on one of the devices hasn't been updated yet and the new schema is not supported on that device? basically, it's up to the application to handle that case. In Composer, it shows a "Data unsupported" message. however, the schema *is* serialized to the database, so it can be looked up and perhaps the older device will know how to handle it, similar to 28:40.
> how does effect/schema compares with things like Lenses from Ink&Switch or grafana/thema good question! As I mention at the end of the talk, we're not actually handling migration yet. We're talking with Geoff from I&S about adding some sort of lensing system on top of Effect/Schema. I'm not familiar with grafana/thema, but I'll take a look!
So basically, it is a create-t3-app with NextJS Prisma and tRPC + some additional stuff that adds your dta into your localstorage and schema inside your db?
Not at all, actually! I'm actually building a comparison of a create-t3-app with Prisma + tRPC to a DXOS-based app. I'll share a link here when it's done. Good eye, though! That's definitely the stack I would choose to enable this sort of behavior if I wasn't using DXOS.
One of the benefits of frameworks is when it handles things for us so that we can move up the ladder of abstraction. But if you want to build it yourself, you can definitely assemble something DXOS-shaped from existing pieces! I'm writing an article that expands on this topic.
13:07 - wooowwwwwww yesssss! 🔥🔥🔥 Somebody get this video a Z-Pak - because it smashes - really brinin the hammer, don'cha'know - haters r like "it stings when you number 1"
Amazing, so it would like to be able to run effect on heterogeneous environment, for exemple in backend/orm/postgres we work with fastapi, but effect can't work on python ?
Hmmm, I wonder ... while Effect doesn't (yet?) run in Python, you could at least get some of the benefits of unifying the schema across front-end and back-end by serializing the schema to JSON Schema and sharing that schema with your Python backend. You'd need some method for going from JSON Schema to your database schema.
Ultimately, this problem is begging for some unified tooling.
1. what happens if the software on one of the devices hasn't been updated yet and the new schema is not supported on that device?
2. how does effect/schema compares with things like Lenses from Ink&Switch or grafana/thema?
> what happens if the software on one of the devices hasn't been updated yet and the new schema is not supported on that device?
basically, it's up to the application to handle that case. In Composer, it shows a "Data unsupported" message. however, the schema *is* serialized to the database, so it can be looked up and perhaps the older device will know how to handle it, similar to 28:40.
> how does effect/schema compares with things like Lenses from Ink&Switch or grafana/thema
good question! As I mention at the end of the talk, we're not actually handling migration yet. We're talking with Geoff from I&S about adding some sort of lensing system on top of Effect/Schema. I'm not familiar with grafana/thema, but I'll take a look!
@@jessamartin got it, thanks! Thema has similar goals as I&S lenses, but uses Cue as the schema definition language and it’s for the Go ecosystem.
Great talk. Would love if you could post all the talks at once, drip feeding a talk a week is too slow haha
Glad you enjoyed it!
Can this be used with Graphql? Can we define graphql schema with effect/schema?
So basically, it is a create-t3-app with NextJS Prisma and tRPC + some additional stuff that adds your dta into your localstorage and schema inside your db?
Not at all, actually! I'm actually building a comparison of a create-t3-app with Prisma + tRPC to a DXOS-based app. I'll share a link here when it's done.
Good eye, though! That's definitely the stack I would choose to enable this sort of behavior if I wasn't using DXOS.
@@jessamartin is it ready?)
Very surprised by the emphasis on magic, we need less magic, not more
One of the benefits of frameworks is when it handles things for us so that we can move up the ladder of abstraction. But if you want to build it yourself, you can definitely assemble something DXOS-shaped from existing pieces!
I'm writing an article that expands on this topic.
13:07 - wooowwwwwww yesssss! 🔥🔥🔥
Somebody get this video a Z-Pak - because it smashes - really brinin the hammer, don'cha'know - haters r like "it stings when you number 1"
I'm sure people running the sound love it when you call them 'yo'. :(
So… Basically Zod
zod + drizzle orm actually. But zod only does validation in one direction. Saying this as a zod power user.
key difference is that schema is ast based so it allows you to generate whatever you want (incl encoding and decoding) from the ast
Cool talk, but feels like I just watched an ad mostly
why?