jfries

If you have a guarantee for the worst case of generating a pixel (which you indicate by saying that you never drop frames), couldn't you get rid of the framebuffer? Schedule pixel generation so that they complete just in time for when they're needed for output.

This would save up RAM for other things (and be a fun exercice to get right).

Yes, that would give some peace of mind, right? Unfortunately for us, that's not the case. The only platform specific code is the 8 instructions on top of "Code of framework" on http://www.sizecoding.org/wiki/Memories

First to set the video mode, and then to set up a timer used to progress time.

Not sure if this app supports it or not, but it would be very useful with a 5 second warning before being thrown into a video call. I could see all kinds of embarrassing moments happening otherwise.

Since propagation of information is limited to one cell per step, it's possible to speed this up by splitting the image into smaller subimages and solving those first, and then solve the borders between such subimages.

This technique can be used to solve multiple steps as well, the only change is that the border between subimages becomes thicker.

Could progressive mode be used to serve thumbnails by just truncating the image at a suitable point, or does the spec (and so, decoders) expect the whole image to eventually arrive?

What are the implications of this?

Surprised by the negativity in the comments. This is an extremely impressive presentation of ability to put together current technologies and get the thing working front start to finish.

A+, would hire.

That's a tempting theory, but the increased ratio of white vs blue collar jobs isn't enough to explain it. See graph on https://www.businessinsider.com/great-news-weve-become-a-whi...

Would that be legal, or counted as insider trading? Since you're trading on relevant information not available to the public.

Interestingly (or perhaps annoyingly) enough, the placement of the 0 was different in different countries. At least in Sweden the first digit was 0, and not 1 as in this implementation. This was reflected in the emergency number which was 90000 at the time, which is easy to dial without mistakes even in a stressful situation.

Is there a need for localization?

It's disingenuous to compare a single year vs an average like this, without going into what the distribution looks like.

Maybe years are either very calm or have lots of fires? That would result in some years being much worse than the average.

Yes, I think it's frustrating that this is not widely adopted. The upsides are really obvious and the only downside (I guess?) is that it feels a bit awkward to do it before society has gotten used to it.

Another thing with obvious benefits I wish we would adopt from east Asia is wearing a face mask when having a cold to prevent infection those around you.

Aren't these large numbers just the sum of the maximum jail time for each of the charges? How are multiple charges handled in other countries?

Cool! -betterC looks well worth exploring. Also, pure and safe are fantastic features, I just wish the default was the other way around, so you need to specify impure and unsafe.

I've heard before that "NAT is not intended as security", but isn't the effect still the same, that an external device can't connect to a device behind NAT without explicit configuration allowing it?

I'm probably missing something, but I think that, for home networks at least, NAT is wonderful because of how it requires some effort to make devices exposed on the external network. If we were given an unlimited supply of IP addresses from the ISP and all devices were accessible externally, it seems security issues in would be a much larger problem.

This article may be interesting as trivia, but the nuance of many of the words mentioned is so different from the literal translation that it doesn't make much sense.

I would be super interested in more details on how that works. What do you actually pass to z3 to solve?

The user id can be used to seed the random number algorithm to achieve that.

I have used them to prove equivalence between pseudo code and optimized production code. Several rounding-related tricks and a too large search space made it very difficult to prove with traditional testing.

Isn't this exactly when active noise cancelling works the best?

Very interesting article! It answered some questions I've had for a long time.

I'm curious about how this works in practice. Is it always good enough to take the outputs of the next-to-last layer as features? When doing quick iterations, I assume the images in the data set have been run through the big net as a preparation step? And the inputs to the net you're training is the features? Does the new net always only need 1 layer?

What are some examples of where this worked well (except for the flowers mentioned in the article)?

In Sweden, you need to tell the employer about the plans to take time off for starting a company 3 months before.

That's on par with parental leave, where it's not uncommon to tell the employer 3 months before the delivery.

(Hi Pedro!)

Maybe nitpicking, but languages like Chisel and MyHDL aren't really HLS. Here there is a straight-forward mapping between the written language and the rendered result, and there should be little surprise in what logic is actually generated.

I am convinced that some specimen of this class of languages will eventually overtake verilog. One feature I'm eagerly waiting for is an equivalent of Option/Maybe types, which makes it impossible to access some signals unless they are signaled as valid by a qualifier signal.

I'm curious about what improvements you would like to see in SystemVerilog?

I really like that trick. It's very powerful. But I wonder what the difference is between this and protothreads?

Speaking of IDEs, a cool feature would be automatic read-only inlining for short functions. Could probably help understanding what's going on in code with lots of calls.

Formal verification is alive and well in ASIC design land. I suppose it's especially suitable here because of the mix of late bugs being very expensive and that the designs are fundamentally restricted to not too many fancy tricks (for example, hardware can't be monkey patched in runtime).

But where I think formal really shines and saves a lot of time is in bring-up of new code. Using formal, it's possible to write a simple unit-test style test environment, and rely on the formal tool to insert all tricky stall conditions, etc. It's then possible to verify things that are normally tricky to both design for correctly, and also to test.

This is very interesting.

I remember reading about pilots placing an electrode on the tongue and eventually learning to intuitively feel what the external sensor was telling them.

Or another one about a person wearing a belt which always vibrated in the spot facing north. This helped them navigate cities more efficiently after getting accustomed to the vibrations.

Do you have other examples?

Imo, it's a bit unfortunate to drill the curiousity out of kids. I'm very thankful for being allowed to mess around unsupervised on my father's work computer as a 7 y/o. Pressed a lot of stuff I didn't know what it did and learned in the process.

This was very cool idea and is one of the things that made me start playing. In the Official Guide book to UO you could find for each creature what it desired. From what I understand that was lifted from the database guiding the ecology system.

A great source for more information on the design of UO (including some bits on the ecology system if you dig deep enough) is Raph Koster's blog: https://www.raphkoster.com/tag/ultima-online/