Is there anyway to tag label tokanize

there is minWidth and maxWith, these seem to be perfectly identifying the breakpoints

you can label them based on any heuristic

you can predefine your own names with ranges for example and match those

is that an option?

You can make it work, but it is not ideal

manually assigning ref ids is better?

fundamentally what is this name: max-w-screen-xl, who is defining that name? the code?

It is, you do not need to care about the values, values that may be more volatile, or sometimes, do not matter.
For example, if you change max-w-screen-xl values underneath, my codegen dont care, whatever value is, it should work

the name is defined in Webstudio primarily, I see a world where programmers are invited into the "designer" space, and both roles collaborate to generate artifacts that streamline the development

So the conversation happens in Webstudio ideally, yes

I am still trying to better understand the problem

again, what is this value? max-w-screen-xl

its a set of predefined names for breakpoitns

right?

Yes

so you are having a set of predefined names you want to use in code

and there are a set of predefined names in builder

code tries to be the source of truth here

but in reality the builder is

Right, but how? Are you talking about the "Breakpoint name"?

I think there is a fundamental problem here

the problem is that only one should be the source of truth

not 2

by manually setting the id in builder you are trying to bring that id from code into builder

this is fragile

I agree with you, I want webstudio to be the source of truth

instead we should be able to use in code what comes from builder, not telling builder how to name that thing

the problem is that I have two world I live in: everything in the Webstudio land, and things that are manually done outside the codegen

I can not make the assumptions that absolutely everything is done in webstudio

thats right

I don't have an answer right now, but I don't think puttung ref ids inside builder is the answer

Me neither, just a thought

it seems to me that you want to drive certain configs from code and make them immutable in builder

I am not sure what is the right answer neither, hopefully it could be figured it out

e.g. passing breakpoints config to builder from outside and disable changing them

well, even better if you get to enforce constratins to designers

or maybe passing initial breakpoints config from outside where refid is provided

so they get to work with tokens and rarely with "Magic Values"

actually there is no need for refid, it could be just the id we have provided from outside along with config instead of generated one

As long as if you destroy specific object with an ID, you could create it back again with the same ID I guess

basically you want to configure initialBreakpoints and make adding breakpoints disabled

if you search you will find its a static config initialBreakpoints

Well, same for gap , margin, padding, fonts

like imagine the entire tailwind config as tokens, limiting the webstudio design file

yeah, you want to provide a theme config to builder that would enforce a bunch of constraints it seems

unless you force the tool to exit out of the limitaitons of course, but it would be everywhere

you got it

I don't want to discourage you, but this is beyond what we need right now in its current stage

your use case is outside of my comfort zone right now

still, I would like to work with the underline data from the webstudio to see how much it could be map to code, and how

all good, I just want the tool to get there at least

I am sure this all can be solved in time

I am not expecting stuff like this be a priority but I would like to settle and invest my resources into a tool that will get there eventually

but right now adding such things to the core before we figured out a massivve amount of other questions is not practical to majority

Plasmic was a really good idea, but closed-source and the business model wasn't align with me generating code and hosting things myself and controling things way too much

I get that entirely

I couldn't afford such blackbox magical thing that it will fail at some point, just a matter of time (or why I avoid webflow for that matter), they all have their values, 100000%, but I want to take the tools a bit further and solve way more complex problems and push the boundaries

sure, but then you want me to modify core and spend engineering ressources on things we are not ready for

All I ask is to be patience with me, and please dont discourage me from making a tiny progress. Or upfront if you are not gonna there :hugging_love:

Well, hopefully I can help little by little, the more I learn, the more I get to code and execute on things

I am actually encouraging you to focus on the short-term goals, we are overwhelmed with the amount of things we need to do

the thing you want will come, no doubt

and when it comes, there will be a deep understanding of the ux and use cases

yep

e.g. when I decide on features, I always analyze who is this for and what kind of company and would they be an important user group

@jkrishna has spent tons of time building for that use case at teleporthq and this went nowhere afaik, because it is so fragile and hard to use that in the end devs won't use it

so I need to question deeply the scenario in which it would be used and how

definitely want you to win here, I do not want to be coding boxes for a living anymore 🙂 (In a good way I say it)

we finally have the capabilities on the web to move such things out of the code to some extent, and culturally in the designer space as well

jfyi JK is https://github.com/JayaKrishnaNamburu

he maintains https://github.com/teleporthq/teleport-code-generators

just that low-code-no-code-trust-me-our-platform-is-the-way-we-host-it-you-dont-need-to-code is definitely not the way and no way I wil bet on it

I have no idea what you are talking about, is this because the code generated from the platform won't fit developers needs?

we are launching the CLI soon, so you host it yourself if you want to

but the generated components for now is what we need them to be

consumer is using them as a component

customizing the generated output is very complex and fragile topic

it opens up the API for huge amount of bugs

keeping the implementation details of generated components as a "system" thing and letting consumers use them as components are, just like any react library you would install from npm

you don't go into ast to generate a diferent code when installing a lib from npm, right? but that's what you seem to be trying to do here

I can see this as being useful, but I don't want to deal with the complexities of it as we are a few developers, not until there is a robust ecosystem and more funding.

I understand, we get there, eventually, at some point 🙂

exactly, its not trivial

its the hardest problem in the nocode realm

in fact so hard most attempts failed so far

if we do it, we do it right

something to checkout: mitosis

Incentives***, well try to fix another bug tomorrow, I am disconnecting for tonight. Thank you so much for steaking around and to be patient. Appreciate! Good night :Sleepge:

Thank you for investing so much here, I appreciate the effort

lets focus this energy on becoming a great visual tool first and a platform for integrations

insanely important to fix all the ux problems and start building out developer apis

Our roadmap is:

• CLI launching in a week, publish to vercel
• Fixing a number of UX problems and things that are preventing people from using it right now (e.g. custom favicon, SEO settings and more)
• finally creating a way for connecting to external APIs and CMSs

To become a platform we need to create apis and appstore like system where 3rd party developers can submit an app that can use those apis.

This is the most important thing to achieve as a core and from there we can see how these will shape the ecosystem

I can imagine such apps would be the right place to do crazy stuff with the data and provide additional UI for configuring things from builder, so that builder can stay the source of truth while giving on the other side the right output.

Just curious to know what are you referring to when you say codegen tools?

Are you building something on top of WS or using the output of WS in some other tool?

That is the intent, yes, using the WS specification/document

@Yordis Prieto sent you a friend request, your DMs are protected

I need to fix that