Show newer

here's one:

"Multiplayer Book
Book for two ppl to read at the same time"

Two parallel stories, meant to be read on synchronization. The readers can look at each other, maybe also discuss?

Show thread

🧵: Post an idea or inspiration you noted, but probably be won't ever try out.

playerctl doesn't think my players can withstand such requests

s-ol boosted

the code for my 666RS is on sourcehut, if anyone wants to give it a try.

git.sr.ht/~noizhardware/666cpu

there are examples in the /asm directory, the doc for the 666R is complete, 666RS is still in progress, so no complete doc yet

compiling and running it should be pretty easy, only dependency is SDL2.
If you need help, I'm here :)

There is too many papers about "gesture recognition" that don't acknowledge that "gesture recognition" alone is an interesting task nearly completely devoid of purpose. It's not enough to recognize a gesture once it is completed.

Any interface you build on top of that is *much* worse than a (physical, or touch, or ...) button would have been.

The paragraph below the image, also pictured:

Click-through tools can save steps relative to traditiona linteractors. For example, the tool in figure 2 selects both anoperation to apply (change color) and an object to apply it to(the circle) in a single two-handed gesture. This can save timeand reduce cognitive load, because the user can combine more steps into a single mental “chunk".

Show thread

Couldn't figure out how to add alt-text in Tootle, so it is here instead:

Screenshot of part of the article (from page one).

A diagram shows a circle and a triangle, filled in with gray and white color respectively, and partially overlapping.
A black square frame is drawn on top of both, and a mouse cursor points to a part of the circle that is also inside that frame.

"Figure 2. A click-through button is used to change the area color of a circle."

Show thread

„A Taxonomy of See-Through Tools“, Eric A. Bier, Maureen C. Stone, Ken Fishkin, William Buxton, Thomas Baudel (1994)

lri.fr/perso/~mbl/ENS/DEA-IHM/

s-ol boosted

The industry's current approach to cybersecurity is to ban running non-preapproved EXE files from non-preapproved locations.

This is completely giving up on the idea of a computer as a thing that runs software created by the user. It is also an admission of complete defeat of the idea of an Operating System as a trusted guardian of security (because otherwise it would be safe to run EXEs).

It is likely that this won't be the last of user rights to go. Once gone, they will be hard to get back.

Show thread

hey, can someone *please* come out with a nice set of 2x d16 labelled 0-F so I can start rolling my colors and pointers?

I can only find ones labelled 1-16 like some normal people would use :(

made a SubV / ISA reference sheet for all the common mnemonics and their values:

I almost found this out by debugging in GDB way in the beginning, but I think I broke off my testing 10s before realizing. Could've saved myself a lot of hard thinking, IRCing, code reading etc... Oh well :)

Show thread

So when I was enterng characters, they were arriving in my program, but I wasn't aware that it was my responsibility to send them back out in order to show up on screen.

And when I hit enter to confirm the input, the result also didn't show up, because of the CR/LF thing.

So I assumed input wasn't working at all, when in fact input was working fine the whole time.

Show thread

2) I didn't really understand the I/O convention with serial port communication: with UNIX tools, what you type is echo'ed back by the shell, and the output of the tool is intermixed (unless the tool requests otherwise). I thought this was supposed to work with my bare-metal code through qemu as well.

But actually how PC-serialports/terminals work is you just "fire and forget" your input at it, and the terminal sends back what should be on screen.

Show thread

This was *deep* yak shaving territory.

In the end there was only two small issues/misunderstandings:

1) qemu sets up the terminal to send only CR, not CRLF or LF (which I expected). By piping in from `cat` I was forcing a LF; but when I typed things my code never got to the part where it continues to print the response.

Show thread

Finally got underneath the mental knot/misunderstanding that prevented me from having Serial input in my SubV test program, after three days of a wild goose chase.

First I thought I was doing something wrong with the virtual Serial port hardware; then I thought there was some magic qemu stuff I might need to trigger, or perhaps an ANSI escape sequence... At some point I compiled a RISC-V Linux kernel and busybox initramfs just to see that input works there!

@akkartik what kinds of debugger support have you built for mu and SubX?

I'm looking around and the ELF/DWARF formats are looking pretty annoying tbh.
I'm thinking at this point it would be worth at least to put the labels into the ELF file as symbols, so that one can at least see *where* in the code one is...

Then it would be nice to have the source code mapped using DWARF but maybe that's just not worth the effort.

cc @akkartik

Turns out my last level of comprehension, and also syntax proposal, was still flawed - mainly because you cannot simply bit-slice signed numbers.

The solution was to split the "format" pass into two, and do the earlier half *before* "survey" runs.

This way there is no need for a slice-syntax at all anymore!

git.s-ol.nu/subv/blob/master/R

Show thread
Show older
Merveilles

Revel in the marvels of the universe. We are a collective of forward-thinking individuals who strive to better ourselves and our surroundings through constant creation. We express ourselves through music, art, games, and writing. We also put great value in play. A warm welcome to any like-minded people who feel these ideals resonate with them.