Last year I have created a few Meteor apps and was quite happy with my productivity while using Meteor and a few community packages (like aldeed:autoform, yogiben:admin, peerlibrary:blaze-components, 'orionjs` and others).
Now I’m starting one more project and because of Blaze become deprecated (according @gschmidt in Next steps on Blaze and the view layer it seems that I should use React or any other UI framework, but not Blaze. The problem is that there are no good admin panel and forms package for React or Angular2, so I decided to look what above mentioned package authors are doing. I was suprised that they are still working on their packages and are releasing new versions!
@mitar, @aldeed, @nicolaslopezj, @yogiben and other authors of Blaze related packages, what is your vision of migration out of Blaze, or maybe you still have some hope that Blaze will be not deprecated?
For me, I do not care what is the rendering engine below. I like ease of use of Blaze and API of Blaze Components. So I am planing to continue to use (and develop) that, and then when there will more serious React support, allow rendering with React inside with same API from the outside.
@mitar, as far as I understood from @evanyou posts, MDG is not going to build any king of project like ‘Blaze2’. They will probably build some stuff which helps to migrate from Blaze into pure React but not more. So you will have no better React support to hook into that you have now. @evanyou am I right?
@chompomonim That’s a very good question, but this issue is not unique.
Many of the older packages (including yogiben:admin) were hardwired to use iron-router. When flow-router became the plat du jour, it seemed like a lot of our hard work went down the toilet.
We at Meteor Factory are rebuilding yogiben:admin to be more modular, as a series of packages that work together.
I think a really good example of a modular set of packages is useraccounts, which allows you to use a combination of frontend frameworks and routers.
IMO Blaze will never die. I’ve taught Meteor to relative newbs and React is a far steeper learning curve. I believe that Blaze will still be used for prototyping.
We’re currently rebuilding admin to work in React. I guess most other authors will have to do the same.
I think we wanted to announce that we were going to switch to React as soon as possible (before having a concrete plan) so that people could get in that mindset as soon as possible. I’m sure that the final plan for migrating will include working with authors of popular packages to make sure everything can move over smoothly. For now, I think it is a totally fine choice to continue building apps in Blaze to take advantage of the many great packages for it.
For autoform, I plan to keep it based on Blaze on the rendering side, but the submission and validation pieces really have nothing to do with rendering engine or templating language, so those will probably become their own packages that you can use with forms rendered by any means, even static. Then there will probably be alternative versions of the remaining auto-rendering package, though I don’t know yet whether angular2 or react will be first. Still experimenting with both. It’s worth noting that some folks have had success mixing blaze-based autoform into React apps, so really you can probably mix n match with little effort, but for best performance we’ll eventually want something like autoform-react and autoform-angular2.
I am currently focused on finishing new major versions of collection2 and simple-schema, so any work on autoform is still on hold for a few more weeks.
Yeah I mean at the end of the day anyone can write a library to render HTML from a SimpleSchema definition, I’d definitely agree that your focus is 100% correct.
