What would be required to build patcher for browser

tekknovator 2021-12-15 05:03:33 UTC #1

The shower thought of the morning is inspired by the excellent M8 display PWA for Dirtywaves M8 tracker.
Imagine you could fire up Axoloti patcher in a browser.
Or just a gui, with a low level usb handled server side. Imagine axoloti running connected to a Norns as host. Just open the patcher GUI on your desktop machine etc, then send to axo without the need to unplug the usb.

What would be necessary?
Could emscripten compile the existing codebase? Did anyone try?
Would the chromium USB/serial API be sufficient for client side connection?
Etc.

janvantomme 2021-12-15 11:49:01 UTC #2

Building something from scratch is probably the only option.
Current patcher is JAVA.
All objects are XML files with some C code, those are concatenated into a CPP file in the build directory, compiled and uploaded to the board.

tekknovator 2021-12-15 15:30:18 UTC #3

So just to make sure I get this right. Everything the patcher does is take the graph and c extensions described in the xml, create a temporary c file, runs the compiler on it and then uploads the binary?

Edit: just realized that it is exactly paraphrasing what you wrote @janvantomme

Edit2: did anyone ever try it? Rewrite?

janvantomme 2021-12-15 16:34:06 UTC #4

I think someone has tried to make the patcher look better, but gave up after realizing how much effort it would be. There are some GUI mockups on the forum somewhere.

SirSickSik 2021-12-15 22:19:35 UTC #5

gui mockups? sounds interesting.. do you know where/how I could find these?

janvantomme 2021-12-16 12:47:22 UTC #6

Can't seem to find them anymore.

DIYDSP 2021-12-22 15:47:56 UTC #7

I don't know the exact details but this resonates with some of my other thoughts:
I don't think Axoloti should be forgotten! It is way too huge and powerful of a step in the right direction.
I would love to see it made a little more open and flexible, e.g. system diagrams drawn, some basic API documentation, etc., so that it can be continuously ported into new projects, such as this one.

tekknovator 2021-12-22 18:54:36 UTC #8

This is not even a project yet
It really was just a thought. I could pull it off, but lack the time.
The original project was a students research, from what I get. It is perfectly understandable that such a project is hard to maintain, once life kicks in.
With a lack of hardware, it is hard to keep community momentum going. It is a 1 2 punch for Axoloti.

The lack of hardware could be addressed by porting Axoloti to more platforms. Where there any ports to other, more generic, boards?

tele_player 2021-12-22 22:43:35 UTC #9

Years ago, there was a port which could run with STM32 Discovery board, and about 2 years ago, new hardware was designed by a user here (urklang on this forum), using a newer, faster STM H7 chip and different codec. This was called Akso, and resulted from discussions here lamenting Axoloti unavailability.

Unfortunately, the Akso hardware never became reliably available, and the port of the software and firmware needed a lot more polishing, Nick needed to get a real job

https://zrna.org/akso

Then COVID…

I think the Axoloti system is brilliant, but for various reasons, it fell short of its potential.

janvantomme 2021-12-23 15:35:11 UTC #10

At the moment it's mainly a shortage in semiconductors. There are new prototypes of future Axoloti versions, but it's hard right now to do production runs.

tele_player 2021-12-23 15:55:34 UTC #11

There are new prototypes? From Johannes?

janvantomme 2021-12-30 16:26:11 UTC #12

Yes. -> http://community.axoloti.com/t/is-axoloti-over/7592/2