Hacker Timesnew | past | comments | ask | show | jobs | submit | viktorcode's commentslogin

I use several non-fullscreen windows over desktop. Stage manager makes switching between them very convenient. But I do use full screen windows, they live in their separate spaces. I see no reason whatsoever to maximise any window without it going full screen mode

It will. The web chat has censorship features, but the model you can download doesn't.

The way I understand it, it's a way of compressing vectors by switching from their per-component representation to polar coordinates representation, where the nearby vectors are clumped together to a single line, allowing to describe them by different lengths

Judging by the comments sideloading plays a major part in everyone's life. What apps do you sideload guys? Why those apps are not in a store?

F-Droid. And also by Google's definition, everything I install from F-Droid. So Antennapod (Podcasts), ConnectBot, DAVx (sync my Fastmail calendar to my phone), Etar (Calendar app), Jellyfin (media player), Jiten (JP dictionary), KOReader (ebook reader), OsmAnd~ (Maps), VLC.

Meanwhile from the Play Store I have Bitwarden, Firefox, 2 banking apps, a few airline apps, Wireguard and Whatsapp. So I actually have more from F-Droid than the Play Store from what I regularly use.


Why not grab Fennec from f-droid as well? It used to also have more features, I'm not sure if that's still the case but might as well go with the open source build

> What apps do you sideload guys?

I sideload no apps. I install most apps from either F-Droid main, or an other repo.

> Why those apps are not in a store?

All of them are in a repository. Just only the state sponsored ID-app is only available via the ad-infected Google RAT delivery service, also known as Google Play.


Every non-stock app on my phone was installed from an APK directly downloaded from the manufacturer or open source developer's site / Github releases. I've never had a Google Play account and have never used any Android "app store".

I switched from iOS to Android about three years ago. I saved all the APKs for everything I installed (or updated). When I got a new phone last fall it was pleasantly like geting a new PC. I imported my SMS and contacts from my last backup, then installed all the apps I use and imported or manually set any settings I wanted to customize.

The biggest pain was having to manually logon the couple of sites I allow to keep persistent cookies since device owners aren't allowed to just import/export cookies from mobile Chrome.


I _install_ apps through F-Droid, because on average, they are much less user-hostile. Less tracking, less accounts, less shenanigans. Built for usefulness rather than profit extraction. Which apps it shows is also 0% influenced by ads and other commercial value, whereas on Google's store, it's the opposite as it's the biggest factor.

> Why those apps are not in a store?

Why'd I put my app into their store if I don't agree with the store owner's policies?


I primarily go for apps via obtainium and fdroid. I go to Aura if I have to. GPlay if absolutely required (and I actually have to have the app)

Would Obtainium continue to work? I like the freedom of entrusting developers I know and installing APKs from repositories instead of restricting myself to app stores whose publishers have to be identified and approved by an advertising company.

Can I keep this freedom?


Even if all my apps were from Google Play, it's not up to Google to remotely decide what code I can and cannot run on my device. Especially important when talking about whole population.

I install from F-Droid when possible. It has less noise, and all apps are free as in software.

There are some true gems such as:


- NewPipe

(I'm not sure if you wanted to edit in entries or if this was our cue :D)


Apart from why "those apps are not in a store", there's very good reason to want to use an alternative source for your applications. F-droid is a far safer source than google play is, because they actually vet the source code and project and build it themselves. You are far more likely to download malware from google's official 'safe' sources than from F-droid, and hence it's my first option when searching for simple utility applications because the top results on google play will be utterly infested with ads and tracking at minimum.

I was hoping for Mojo to appear as optimisation strategy


I have a few questions.

In that system does the age verification result come with some sort of ID linked to my government issued ID card? Say, if I delete my account on a platform after verifying and then create a new one, will the platform get the same ID in the second verification, allowing it to connect the two and track me? Or is this ID global, potentially allowing to track me through all platforms I verified my age on?

What a verification process looks like from the user perspective? Do I have to, as it happens now, pull out my phone, use it as a card reader (because I don't have a dedicated NFC device on my computer), enter the pin, and then I'll be verified on my computer so I can start browsing social media feed? Or, perhaps, you guys have come up with a simpler mechanism?


The wallet ecosystem is still really varied at the moment. Our implementation is unlinkable. So an issuer cannot track where you use the attribute. And a verifier cannot see that you've used the same attribute multiple times with their system. This is great for privacy and tracking protection, but not so great for other things. For example, people sending their QR codes to other people with the correct attribute (like maybe an underage person sending an 18+ check to an adult), is hard to solve for because they are unlinkable.

Most systems right now have you load data in your phone. Then when a check happens, you scan a QR code. You then get a screen on your phone saying X wants to know Y and Z about you, do you want to share this information? Then you just choose yes or no.

For your social media example. You would just get a QR code on your pc, then pull out your phone, scan and verify, then start browsing social media on your pc.


In the Swiss system, it depends on what they verified. If they required your full ID, that has a document number like a passport and they could track that.

If they did the right thing and only asked for the over 18 bit, then they wouldn't have a trackable identifier.


You are describing a situation where a pairwise pseudonymous identifier is generated. I don't think any real system does this with government IDs, but it might be possible.


> We previously explored Swift, but the C++ interop never quite got there

But Rust doesn't have C++ interop at all?


You can do it via the C ABI, and use opaque pointers to represent higher-level Rust/C++ concepts if you want to.

Firefox is a mixed C++ / Rust codebase with a relatively close coupling between Rust and C++ components in places (layout/dom/script are in C++ while style is in Rust, and a mix of WebRender (Rust) and Skia (C++) are used for rendering with C++ glue code)


> You can do it via the C ABI, and use opaque pointers to represent higher-level Rust/C++ concepts

Yeah but, you can do the same in Swift


My understanding from a brief read of the Swift issue is that they kept running into bugs in the Swift compiler which, in practice, prevented them from doing the things that they ought to be do in theory. This went on for long enough, that they got fed up and abandoned Swift.

The Rust compiler is incredibly solid (across all target platforms), and while it's C/C++ interop is relatively simplistic, what does exist is extensively battle tested in production codebases.


>But Rust doesn't have C++ interop at all?

It also doesn't have the disadvantages of Swift. Once the promise of Swift/C++ interop is gone there isn't enough left to recommend it.


I’m curious what issues people were running into with Swift’s built in C++ interop? I haven’t had the chance to use it myself, but it seemed reasonable to me at a surface level.


There's a list of unsolved problems in this Ladybird issue, now closed because they dropped Swift: https://github.com/LadybirdBrowser/ladybird/issues/933

for example: "Swift fails to import clang modules with #include <math.h> with libstdc++-15 installed. Workaround: None (!!)"


Yeah, that part doesn't make much sense to me. IMO, Swift has reasonably good C++ interop[1] and Swift's C interop has also significantly improved[2] since Swift 6.2.

[1]: https://www.swift.org/documentation/cxx-interop/

[2]: https://www.swift.org/blog/improving-usability-of-c-librarie...


It may have in the future. Crubit is one effort in this direction: https://crubit.rs/


There is also cxx.rs, which is quite nice, albeit you have to struggle sending `std` types back and forth a bit


> albeit you have to struggle sending `std` types back and forth a bit

Firefox solves this partly by not using `std` types.

For example, https://github.com/mozilla/thin-vec exists in large part because it's compatible with Firefox's existing C++ Vec/Array implementation (with the bonus that it's only 8 bytes on the stack compared to 24 for the std Vec).


Luckily, ladybird also does not use `std` types


It's technically Rust -> C -> C++ as it stands right now


Rust has cxx which I would argue is "good enough" for most use cases. At least all C++ use cases I have. Not perfect, but pretty damn reasonable.


So nice to see another person being that enthusiastic about ray tracing! I didn't do a comparable level of work in this field, but as a hobby this fascinates me a lot!

One common misconception is that ray tracing is computationally prohibitive. It was, but no longer so; it's a target within our reach, especially so when there's GPU with hardware acceleration for ray casting.

Many games use ray tracing for partial scene processing, and of course they all work in real time. My favourite example is Metro Exodus with ray traced global illumination, which works on last gen graphics hardware pretty well. Not all games use the technology efficiently, but the trend is already obvious: with accessible real time ray tracing rendering the scene will become a much easier task.

P.S. I used "ray tracing" when more accurately I should have used "path tracing", but I prefer to use a single term to encompass the whole technology with all its variants.


LoC is a good code quality metric, only it has to be inverted. Not "it wrote C compiler in 100 000 lines of code", but "in just 2000 lines of code". Now that is impressive and deserves praise.


I don't want to maintain an information system written in the style of code golf competitions, either.

I like the author's proposed "Comprehension coverage" metric. It aligns well with Naur's Programming as Theory Building.


Not necessarily.

If I minimize my project and get everything on one line, is that good? I think not.

Measuring success based on how may or how few lines there are is a bad idea, I think.


this is also easily gameable. Any language can easily be converted into a single LOC


As someone from game development, not supporting Vulkan on Android and sticking with OpenGL ES instead is a safer bet. There is always some device(s) that bug out on Vulkan badly. Nobody wants to sit and find workarounds for that obscure vendor.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: