What an I waiting for? Can I start tonight?

I guess I’ve started already (I’ve already begun to experiment with hosting #htmx on experimental builds of #fzx) but maybe I can start re-writing jasongullickson.com in htmx right now? 🤔🤔🤔

The path here is to move everything to #htmx & #hyperscript running on #fzx (this will incorporate/depreciate my work on #ApresWeb)

I have a lot to say about why, but its way too much to write in this little box.

But if it does make sense, and if I can come up with a reasonable way to integrate it, this would finally answer two question that have been holding back a lot of work on #fzx and perhaps more important open a wide-range of new applications, including solving some well-known problems in the fediverse.

Now to find the time to pursue this… 🤣

Could #ActivityPub be a standards-based solution to these design challenges?

It’s easy (and unsurprising) to see how it could be applied to the server federation problem (which is essentially keeping inodes in-sync across #fzx instances), but as always performance is a consideration.

Identity is less obvious to me. It seems like an existing ActivityPub identity could be used but I don’t know yet if this can be used for authentication, or if it makes sense.

Two things that I’ve never found a satisfactory design for in #jsfs (now #fzx) are identity/authentication and federation.

I’ve designed a few identity systems that can be implemented on top of #jsfs itself (or perhaps jsfs-x) but I’ve never loved any of them, mostly due to complexity and bootstrap problems.

I’ve drawn hundreds of federation designs (and implemented a few) but they all had limitations I found unsatisfying.

As I read the ActivityPub spec this morning, I got thinking… 🧵

...and with that, #fzx (formerly-known as #jsfs) is running on #solarpotato