It could do with more polishing up but if you're like me and want to see more experimentation with new browser engines, I've just curated a great place for you to start:

Make sure to read the README (though I should inline that):

I'd love to hear back from some of you with more projects to add to this list, however small!

@yaaps Please read over this list/notes, give me your thoughts, and share it around.

I hope this makes browser engine experimentation more feasable!

@alcinnz @SoniEx2
My top priority for 3-6 months is this:

I'm turning a wiki engine into an ActivityPub service. It's pretty clean considering that you can control all user facing elements from the web UI and hadn't been updated in 7 years. That's how I'm delivering games, social, and publishing. I can target this browser.

@alcinnz @SoniEx2
Specifically addressing the areas of having content that would utilize the capabilities of the browser and producing a framework for testing it

@yaaps That would be great!

I'm definitely keen to see how it sounds in the voice-browser I'll be starting soon...

I think you're mostly looking at html and css properties, but there's also this:

State control is less attractive than would appear at first blush, mostly because it gets used to force people to traverse menu trees when random access would be easier for everyone who doesn't have an investment in artificial hierarchies

If a Turing complete SGML becomes a w3c standard, web apps are going to become more hellish than ever

Every web publisher making a statement with their design, most saying, "Hold my beer"

@yaaps I'm not aware of anything this way, links please?

Though I'm definitely not keen on Turing Completeness, our vulnerable CPUs can't handle it well not to mention the surveillance it enables.

SCXML is state *control* extensible markup. It defines conditionals

So this isn't just a nudge to turn on audio and mike, it's Turing completeness in the output code of JS and Wasm that enables yet another layer of indirection in code that's already nearly impossible to audit

The main risk is that information not available to the remote server can be used to decide what code to run, which could be used in a variety of ways to defeat the purpose of VPNs

@yaaps Ah! I was thrown off by "SGML" which I understand to be an XML precursor.

Well, what I'm planning to implement is to have the computer read the page out to you with a certain beep indicating a link (defined very generally to include e.g. <audio> tags) for you to repeat one back in order to navigate.

Forms would be interesting to support too, but that'd be quite different experience.

I use SGML as a general term for XML and HTML because I'm old 😂

The web UIs for PBEM adjacent games rely a lot on table data. I'll update formatting to grid model so that block elements consistently have semantic relationships and tables are *always* data

What's the best way to make relationships implied by placement in a grid layout explicit for speech rendering?

@yaaps Good question! I'm very new to audio design and would appreciate any guidance.

But three things come to mind:

1) Arrow key navigation could help, and would be useful anyways for longer documents.
2) I need to extend my styling engine to support the selectors which would support that communication.
3) The best way I can think of is to inline the headers into each of the corresponding table cells.

For TTS, there are properties I can use to frame or replace the contents of an element to create an appropriate aural presentation for the content:

Application design for voice menus is much the same as drop down menus, but you want to strive for an illusion of open-endedness by providing unique primary names for each node, recognizing synonyms generously, and using the tree to guide prompts rather than constrain inputs

The Alexa dev materials are probably the best source for designing a VUI

My involvement in VUI is as a hobbyist going back to the late 1990s when Diplomacy and other PBEM games were feeling pressured to build web UIs. My argument was that CLIs need to be maintained because page mode UIs wouldn't achieve feature parity before they gave way to hands-free patterns. (I hate being right. Good thing I preserved the code). Speech input is linear like a CLI

I thought that was your source for discontinued CSS for speech 😂

@yaaps I meant I saved the Alexa link for guidance.

And yes, I found the right link for the CSS properties I'll support.

The CSS speech page I linked was a guess. That draft is discontinued, so whenever you come across your actual source next, please give a shout so I can cycle it into my reading. Tx!

Sign in to participate in the conversation

For people who care about, support, or build Free, Libre, and Open Source Software (FLOSS).