Preact: The road to Preact 11

[marvinhagemeister]

With Preact X being a great success there The 10.x release line was all about increasing ecosystem compatibility and the initiative was a complete success. Compared to Preact 8.x there are many more third party libraries that work out of the box with Preact, we saw the addition of Fragments, hooks, a revamped devtools extension, prefresh for native HMR support and much more.

In summary, we learned a ton the past year whilst maintaining Preact X and we want to push Preact even further. So the next question for us is: Where do we go from here? What should a successor look like?

This is a collection of some of our thoughts, but is in no means a concrete feature list for Preact 11. Please keep that in mind.

Size reductions

Our small (smol?) size has always been one of our strong points and the thing I'm personally the most proud of. With an ever growing feature-set we're currently very close to the 4kB mark and would like to bring that number down again. Whilst we may be able to shave of a couple bytes here and there, I think we won't have substantial changes on that front without thinking critical of each feature.

Only include what's actually used

The most common talking point is that many projects (including preact-devtools) don't use a single class component. The natural question that arises here is why we need to still pay for the cost the class API regardless. Ideally only the features that are actually used should be shipped to users.

The same is true for preact/compat which will always include side-effects even though a user may only needs to import to the Portal component for example. There have been some attempts to do that in the past, but to me it seems like some changes in core could make that easier.

The overall motto should be: Only include what's used.

Move IS_NON_DIMENSIONAL to compat

Whilst it sounded cute on paper it has turned into a growing list of properties (see #2608 ) which is something we never wanted to have in core. The reason we kept it over from the 8.x line in core, was that it allowed us to keep old codepen demos working. This was especially useful at a time where we weren't officially committing to Preact X and we were more just playing around with various approaches.

I'd even go so far as to consider it harmful as I've seen it confuse new developers first-hand when switching between writing CSS and declaring inline styles. It's a feature we cannot remove entirely because of React, but moving over to compat should be fine.

Reconciler performance

Whilst the original plan for X was to break from using the DOM for traversal and only use the vnode tree for that, we ended up in the a little bit awkward middle ground. Most operations walk the vnode tree, but there are some remaining ones that still rely on the DOM itself. This is made more difficult by the existence of Fragments.

For our next generation we should cut ties with our past and completely base it on the vnode tree instead. The newly added statistic metrics reveal some great places to look for improvements. Some of the ideas that are currently floating around:

• Add fast path for single child elements
• Store DOM operations in an effect queue (aka centralize paints)
• Add fast path for mounting

The effect queue idea is in particular interesting as it would the browser to batch all paint jobs and process them at once instead of having those intertwined with running JS. Heck we could make a lot of the DOM pointer code easier by applying those changes right-to-left instead of left-to-right.

What's more is that having an effect queue would open up the possibility for custom renderers. It's a long shot and not something we'll focus on in the near term, but we would at least have the possibility to do so nonetheless.

Boosting hydration

We're already in a good position when it comes to hydration performance but we have a lot of ideas what we can do to make it even faster. Any optimizations we'll do on our reconciler will directly benefit hydration, so there is a very close connection between the two.

Besides reconciler performance, there is a need to re-evaluate how we can best boot up from SSR'ed content. Due to us not joining adjacent text nodes anymore we have a mismatch during hydration. SSR'ed HTML will always create a single text node, even if it was created from multiple ones.

// Element with two Text nodes
<div>Hello {name}</div>

An alternative to joining adjacent text nodes is to insert HTML comments as markers in-between them. Not sure which approach is ultimately easier, although my gut tells me that the latter can easily become complex.

Remove the need for forwardRef

The introduction of the forwardRef component is mostly a workaround for not keeping ref in props. If we keep it in there we can make all forwardRef components redundant:

// Current way
const Foo = forwardRef((props, ref) => <div ref={ref}>{props.children}</div>);

// Proposal
const Foo = props => <div ref={props.ref}>{props.children}</div>;

There is a downside to that though in that there may be custom runtime checks in third-party libraries that explicitly check for additional properties in the props object. We ran into some of those in the past if my memory serves me well, and I'm secretly hoping that the increasing adoption of TypeScript has improved the situation compared to a few years ago.

Mark root nodes as roots

Both the devtools and Portal component would benefit from having a way to place sub-trees into existing ones. Currently no tree knows about the others which leads to some weird edge cases with Portals. There has been fantastic work during the 10.x release line to get it stable and I feel like we can make that code easier by having a special branch for root nodes as sub-trees in our reconciler.

// This is not possible currently
<TreeA>
  <Foo />
  <root>
    <TreeB />
  </root>
</TreeA>

If we follow that thought further we could theoretically even look into switching renderes on the fly if there is any attached to the root node. A use case for that would be to switch to rendering into a canvas element somewhere in the middle of the tree.

What else can we do?

The above list is already a lot of work and will keep us busy for months but there may be stuff I've missed. Again, the points mentioned here is a collection of ideas and not a definitive feature set for Preact 11.

[38elements]

I think that it is better that Preact 11 stops support IE11 and Edge which is not based Chromium.
Preact X will continue to support these brower.

[marvinhagemeister]

I think that it is better that Preact 11 stops support IE11 and Edge which is not based Chromium.
Preact X will continue to support these brower.

Good idea! We should definitely drop the requirement to work out of the box with IE11. This would allow us to use Map or Set objects. IE11 users can always polyfill those. Regarding Edge: Admittedly we never had any bigger problems with it if I recall correctly. Ultimately we need to see what kind of code base we'll end up with, but I have a feeling that supporting old Edge browser doesn't require any changes on our part.

[38elements]

Regarding Edge: Admittedly we never had any bigger problems with it if I recall correctly.

It may be unimportant thing, there is #2331.

[developit]

2kb 2kb 2kb

[rauschma]

@developit mentioned plans for better supporting server-side-rendered HTML that stays at least partially functional even without JS (I found HEY inspiring here). Any details on those plans yet?

[developit]

@rauschma The main things we have explored related to that are actually implemented using Preact X as it exists today, so I'd put them on an independent timeline from a Preact 11. The reason for this is that anything related to HTML ends up requiring that folks use (more) specific server-side techniques, versus the current simplified renderToString() approach.

I'll try to prioritize at least publishing some of the exploratory work and demos we have. It's less impactful than providing an off-the-shelf fullstack framework, but that's probably fine given the highly variable nature of Preact usage.

[lukejacksonn]

Some very interesting ideas here Marvin, thank you for airing them and thanks to the preact team generally for all their hard work on 10.x. I have committed this year to building preact apps almost exclusively and it has been a pleasure; the tools, the community, everything is like a breath of fresh air.

One thing I am wondering about the future of preact however, is, will it always try adhere to the react API? Being ~ its namesake and given all the effort that went into increasing ecosystem compatibility in X, I assume the answer to this is yes. But I am curious, if you happened to stumbled across a potentially useful new primitive or feature outside of the react API.. would it be considered?

I've been inspired recently by the likes of hyperapp and svelte with their effect and subscriptions APIs respectively. Do you see any promise in these kind of interfaces specifically, or are the team content that these kind of behaviours can be encapsulated by hooks?

[marvinhagemeister]

Some very interesting ideas here Marvin, thank you for airing them and thanks to the preact team generally for all their hard work on 10.x. I have committed this year to building preact apps almost exclusively and it has been a pleasure; the tools, the community, everything is like a breath of fresh air.

Thank you so much for the kind words! It's comments like these that keep us motiviated! :heart:

One thing I am wondering about the future of preact however, is, will it always try adhere to the react API? Being ~ its namesake and given all the effort that went into increasing ecosystem compatibility in X, I assume the answer to this is yes. But I am curious, if you happened to stumbled across a potentially useful new primitive or feature outside of the react API.. would it be considered?

We will keep compatibility with the React API for the foreseeable future, but it's very likely that we'll put our own spin on thinks more in the future. Keeping ref in props and thereby removing the need for forwardRef is one small step in that direction.

I've been inspired recently by the likes of hyperapp and svelte with their effect and subscriptions APIs respectively. Do you see any promise in these kind of interfaces specifically, or are the team content that these kind of behaviours can be encapsulated by hooks?

Personally, I fully agree that a simple subscription primitive is very appealing and something I've played a lot in the past and shipped a few projects with. We see a lot of code in the wild where re-renders are triggered for the sole sake of triggering effects. Not saying that this bad code, but rather that hooks tend to lend itself to that usage and I'd love to welcome different ideas to Preact. So far reactivity is a leading contender to do that in my opinion.

That said we'll always try to keep that part pluggable like now as there never is a silver bullet approach and different applications require different trade-offs. One neat library that uses RxJS as a first class option to manager state for Preact components is bassdrum. There is also #1923 which adds a vue-like composition API on top of Preact as another alternative to hooks.

So yeah I see a lot of value in these concepts and we do want to make it easier for those systems to directly plug into Preact in the future. Hooks are a great approach for local state but they're by far not the only one. Our devtools extension barely uses hooks or classes for that matter for example.

[rauschma]

@developit Having something lightweight in that space would be very welcome!

[yisar]

I see that the new size calculation standard is to calculate the size of Hello world project. After tree-shaking, it seems that there is no need to pay for the class API.

In addition, does the diff algorithm, which is detached from DOM, mean that it is completely detached from DOM, and vdom does not contain DOM information? If so, it would be great. I'm looking forward to it～

[marvinhagemeister]

In addition, does the diff algorithm, which is detached from DOM, mean that it is completely detached from DOM, and vdom does not contain DOM information? If so, it would be great. I'm looking forward to it

The way vdom works is that it always needs a reference to the underlying element. There is nothing that says that this _has_ to be a DOM node though and that can be leveraged for different renders. There is more work to it than that on the road to support custom renderers, so it may be that we lay down the groundwork in v11 and add offical support for it in v12.

[yisar]

There is a way that no need underlying element, such as https://github.com/Matt-Esch/virtual-dom,
It's very primitive, but it can be used as a reference. It's a completely DOM independent approach, and is well suited for cross platforms.

[marvinhagemeister]

Good point, I stand corrected! Creating "patch" objects and push them in an effect queue is something we thought about too. It's something where we will likely experiment more with it after Preact 11 or when we rewrite our reconciler.

[lukejacksonn]

Thanks for your response above @marvinhagemeister 🙇

We will keep compatibility with the React API for the foreseeable future, but it's very likely that we'll put our own spin on things more in the future

This sounds sensible and I can't wait to see what _our own spin_ ends up looking like!

I had never heard of bassdrum so will have to check it out, also https://github.com/preactjs/preact/pull/1923 looks very interesting. Keeping things pluggable seems like a good idea too. Very much looking forward to v11 now! Keep up the good work and thanks again for your insight and explanation.

[sauravkhdoolia]

@marvinhagemeister I have never worked with Preact, But i happen to pass by this repo. And you can already see what I'm doing!

This is fresh, This is oxygen. I know React is mainstream, but only because it came up with the idea of components i guess.

You guys are doing awesome work. I found React very FB sided, instead of being a true OS project. They work on what FB wants(i mean that's ok) but still, thats not how OS should work.

I love React by all means, but that direction ain't fair. You guys will rock soon.

Mark my words.

Keep at it. ❤️

[marvinhagemeister]

@sauravkhdoolia Thanks for the kind words :heart: We share your enthusiasm and can't wait to see what's next :+1:

[jeremy-coleman]

I think preact needs a reverse complement of preact/compat, either at runtime or as a codemod. Right now, there's little incentive to use the preact namespace instead of react. As evidence of this, i think the stars to dependent projects ratio is a good indicator, maybe think of it as an adopt-ability ratio. Lower is better.
Preact 1.03 @ 26.7k/23k
Vue 0.15 @ 169k/1.15m
React 0.04 @ 153k/4.10m
Angular 0.45 @ 62.5k /1.39m
Svelte 1.66 @ 35.7k / 21.5k

If developers could use preact without fear of not being able to switch to react, i think adoption rates would be much higher, which will allow preact to not only compete for a larger chunk of the open-source mindshare , but also to allow people to actually use deviations from the react api.

[38elements]

Since most browsers support Brotli, I think it may be better to focus size compressed by Brotli, not gzip at Preact 11.
https://caniuse.com/#feat=brotli

[zgoda]

For me the top priority is performance, I don't care about React compatibility at all. Most used mode is drop in bundle with HTM as I am doing embedded apps on really small devices with just about 4 MB storage. Preact 10 is already good but if it can be better with 11 then I'm all in.

[klh]

I think that it is better that Preact 11 stops support IE11 and Edge which is not based Chromium.
Preact X will continue to support these brower.

Good idea! We should definitely drop the requirement to work out of the box with IE11. This would allow us to use Map or Set objects. IE11 users can always polyfill those. Regarding Edge: Admittedly we never had any bigger problems with it if I recall correctly. Ultimately we need to see what kind of code base we'll end up with, but I have a feeling that supporting old Edge browser doesn't require any changes on our part.

My two bits here:

I for one appreciate the ie11 support,
if you remove this, it will be harder to argument for Preact vs others due to lack of browser support (especially in bigCorp settings where ie11 is A LOT more common than the whole www).

if you remove it, you'll have to do some heavy lifting to assure that simple polyfilling makes preact ie11 compatible with end-2-end test suites and how-to's that work.

[klh]

could we move to htm and have it compiled out on build instead of vhtml by default (and save a few kb on the post build size) ?

[porfirioribeiro]

About IE11, dropping it will allow the codebase to evolve, ship less and better code for modern browsers.

Maybe it could be possible to create a IE11 compat layer? I think Vue was heading in that direction.

[marvinhagemeister]

Now that Microsoft has dropped IE11 support officially, I'm very much in favor of doing so too :+1:

[o-alexandrov]

To be precise, Microsoft also drops Edge (legacy).

[klh]

wish:
optional DeepEqual comparison in useState, keep the default Shallow Eval, and an optional Deep Eval.

const [ getter, setter ] = useState({DeepObject}, {deepEval:true}).

[remcohaszing]

I’d like to see hooks exported from the Preact core module. For most node modules, sub modules are not supposed to be used directly.

import {
  h,
  useCallback,
  useEffect,
  useErrorBoundary,
  useLayoutEffect,
  useReducer,
  useMemo,
  useRef,
  useState,
} from 'preact';

[marvinhagemeister]

@remcohaszing Really curious about where it's not supported. Can you share more details?

[developit]

It's really unlikely that we'd ship hooks in core, much more likely that we will move Component out of core.

Subpackages are going to become commonplace over the next couple of years, since they're now specifically supported and recommended by Node.

[osdevisnot]

If I may add my question here. Preact has been phenomenal at React feature parity and I assume same for Concurrent Mode.

So, should we anticipate CM in Preact 11?

[marvinhagemeister]

@osdevisnot No, we currently have no plans to support Concurrent Mode. It's still up in the air whether CM is something that benefits apps at all, even for React. The current implementation in React breaks most state management libraries: https://github.com/facebook/react/issues/15317#issuecomment-716554673

Most of our current focus is on progressive hydration, faster SSR and lazy loading things.

[ConradSollitt]

I read a number of comments on supporting IE 11 and I think it might be a good to consider having two production builds for browser usage using type=module | nomodule. For example:

<script type="module" src="[email protected]/dist/preact.min.js"></script>
<script nomodule src="[email protected]/dist/preact.nomodule.min.js"></script>

Recently I have done this a lot more myself in some projects and and just minimizing the source files using terser and then creating the legacy browser build using Babel/webpack. When I've done this the code size on the compressed module builds is smaller and in my opinion much cleaner because classes and other modern syntax are kept.

In the past there was an issue with browsers fetching both module and nomodules and with a specific version of Safari 10 (widely used work around available) and Legacy Edge but as of December 2020 the percentage of users using browsers that double fetch looks to be less 0.05%; granted for a high traffic site that result with large number of users but I've been testing double fetching of recent nomodule|module code I've written and am finding it to not really be much of an issue.

• https://gist.github.com/jakub-g/5fc11af85a061ca29cc84892f1059fec _Testing double-fetching of module/nomodule JS code_
• https://gist.github.com/samthor/64b114e4a4f539915a95b91ffd340acc _Safari 10.1 nomodule support_
• https://caniuse.com/usage-table
• https://caniuse.com/es6-module

If considering this though, one important item to be aware of is there are a still a lot of widely used older modern browsers that support ES6 and type="module" but do not support newer syntax such as the the Spread Syntax .... Example include UC Browser wikipedia link which is widely used in Asia and older versions of Samsung Internet (widely used everywhere).

• https://caniuse.com/?search=javascript%20spread

If you consider dropping IE 11 you would be in good company as Bootstrap 5 (currently in Beta) is dropping IE. Additionally Vue 3 shipped without IE 11 support (it's planned) but after 3 months it's not ready and made a lower priority.

• https://blog.getbootstrap.com/2020/06/16/bootstrap-5-alpha/
• https://github.com/vuejs/vue-next/releases/tag/v3.0.0