HNHacker News
TopNewBestAskShowJobs

smasher164

1,521 karma · joined August 13, 2015

akhil.cc
submissionscomments
smasher164··on New x86 micro-op vulnerability breaks all known Spectre defenses
The ML folks are pulling themselves out of that rut now. There’s lots of interesting work going on for the next generation of compilers.
smasher164··on New x86 micro-op vulnerability breaks all known Spectre defenses
Maybe EPIC [1] architectures need a revival. Rely on compilers to take advantage of explicit instruction-level parallelism, and keep the CPU dumb.

[1] https://en.wikipedia.org/wiki/Explicitly_parallel_instructio...

smasher164··on Wasm3 compiles itself (using LLVM/Clang compiled to WASM)
If I were writing a sandboxing or instrumentation tool, operating on WASM would be much simpler than dealing with the quirks of assembly, syscalls, or some higher-level IR.
smasher164··on Wasm3 compiles itself (using LLVM/Clang compiled to WASM)
An AOT compiler would do a one-time translation to a platform-specific binary.

If the environment you run on doesn’t support JIT compilation (iOS for example), AOT compiling WASM is useful.

smasher164··on iPad Pro M1
It's be pretty great to see Final Cut Pro on the iPad Pro. I'm surprised it doesn't already exist. Maybe the workflows there are incompatible with the iPad UX?
smasher164··on iPad Pro M1
> as a developer you can put anything you want on an external screen

as a developer, not a user. I can’t have two different apps running on different displays, nor can I take advantage of the display’s resolution. In practice, everything is mirrored.

> Almost all apps obviously take advantage of the keyboard

When I say take advantage of the keyboard and mouse, I mean to offer shortcuts/navigation akin to what you’d find on a desktop.

I don’t mean to say that apps should be exactly like desktop apps, but the iPad UX is making it harder to use the current iteration of document/spreadsheet apps.

There are things the iPad does strictly better, of course. But I don’t see these experiences as mutually exclusive.

smasher164··on iPad Pro M1
The whole "this device isn't meant for programmers" argument is a red herring. I don't like how the discussion ultimately boils down to complete control over the OS. That's not going happen, and likely requires a different mental model around OS abstractions, so we should put that aside.

My original point was that productivity apps, whether its photo/video/audio editing, word processing, or 3D modeling, all categorically benefit from features like better file management, IPC, background processing, working with external displays, easier keyboard/trackpad navigation, etc...

I saw some things in the presentation around external storage and thunderbolt devices which look like steps in the right direction, but the OS is really not doing it any favors. I would guess that the iPad is successful among this demographic despite iPadOS, not because of it.

smasher164··on iPad Pro M1
If you look at it in the abstract, using a computer is equivalent to programming it (ad-hoc), whether the computation and memory live in your head, on a clipboard, or in some app. If the whole "low-code"/visual programming trend is any indication, we need to collectively raise the power of tools and narrow the gap between usage and programming.

Apple should be trying to empower users with safe abstractions, not lock them down even more.

smasher164··on iPad Pro M1
I agree that the target market might be different. However, even though multimedia workflows work pretty well on the iPad, some of the same restrictions that apply to development apply there.

For example, there's no notion of rendering a video in the background, organizing project files independent of the application, a plugin ecosystem for creative software, and proper IPC between these apps. So even though the market is different, there's still a lot to be done to come close to desktop editing workflows.

smasher164··on iPad Pro M1
I should mention that my frustration only came out of trying to switch my parents over from an old 32-bit Windows laptop to the iPad. They're already iPhone users and most of their time is either spent in a web browser or in document processing. So I figured that an iPad Pro with a magic keyboard would be perfect for them.

Even still, the level of finesse they have with Excel and productivity software on a desktop OS was just not replicable on the iPad. It pained me to see my Dad work on his spreadsheets on the old laptop again.

The silver lining to this is at least the iPad is good for watching videos and pretty decent at web browsing. They still use the device casually.

smasher164··on iPad Pro M1
To me, it doesn't matter how powerful the iPad is if the software experience is subpar. Multitasking is cumbrous enough that it might as well not exist, multi-display setups can only mirror at a fixed aspect ratio, it lacks multiple user profiles requiring a family to be under a single user, and the few apps that take advantage of the keyboard/trackpad are nowhere similar to their desktop counterparts.

Take Excel for example. Try right-clicking on a cell or dragging down a formula, and you'll see that it only picks it up a fraction of the time, and you have to click multiple times with gesture-like latencies in between to get it to work.

smasher164··on Sublime Text 4 [video]
I used Sublime as my only editor for many years on end. But I could never quite get a handle on the plugins and editor configuration, experimenting with options that didn't exist or affect the editor in any way. With the introduction of LSP, like many non-VSCode editors, integration lagged behind, and older language-specific plugins were no longer seeing much maintenance.

I do still prefer Sublime for the specific use-case of opening a random text file. Its startup time is far faster than any other editor.

smasher164··on BPF for storage: an exokernel-inspired approach
uh..yeah. The halting problem doesn’t mean you can’t do termination checking. It just means there’s no general algorithm for deciding if a machine halts.

We only care about the specific subset of programs that the compiler is able to prove properties about. https://en.m.wikipedia.org/wiki/Termination_analysis

smasher164··on BPF for storage: an exokernel-inspired approach
I'd like to see more proof-carrying code techniques extended to WASM. Some of the work could be handed off to the compiler, embedding the proof in the binary. This could make expensive proofs, like termination checking, more tractable.
smasher164··on OpenSSL will release a high-severity issue fix on 25th
Looking forward to the day that Evercrypt[1] becomes a production-ready replacement for libcrypto.

[1] https://hacl-star.github.io/HaclValeEverCrypt.html

smasher164··on How Safe Is Zig?
I think if Zig wanted to, it could introduce lightweight linear types using the concept of proof variables and interleaving from ATS. Since resource management is explicit in Zig anyways, there's not much additional overhead in "consuming" proof values to signal that you've dealt with a resource.
smasher164··on Goroutines are not significantly lighter than threads
I thought this article would talk about overcommit and demand paging, but it just counted the RSS between two processes and called it a day.

They don't even attempt to distinguish the size of the stack from the rest of the runtime.

smasher164··on Apple now lets you transfer your iCloud Photos to Google Photos
If iOS supported Time Machine backups, my NAS backup solution would be complete. It's crazy that we're either locked into iCloud or have to sync with iTunes.

Yes, I know about iTunes Wifi backup, but I'd rather not be on my home network for this.

smasher164··on Compensation as a Reflection of Values
The question could be flipped around as "why should they be paid more/less?"

The goal is to treat the salary as a necessary (evil?) for "what you need to feed your family, buy a house, etc..." and not as the primary motive for work.

An employee who makes a valuable impact does so because they want to see the company/product succeed.

smasher164··on Compensation as a Reflection of Values
I really want to see this model scale somehow. I've spoken with friends about this model, and it's hard for them to get past the idea of compensation as a proxy for value. The discussion eventually devolves into

    - X type of employee is more valuable than Y type of employee, therefore they should be paid more.
    - There is no mechanism or incentive for growth (a.k.a the career ladder argument).
smasher164··on Git's list of banned C functions
From the commit message:

If you're thinking about using it, consider instead:

  - strlcpy() if you really just need a truncated but
    NUL-terminated string (we provide a compat version, so
    it's always available)

  - xsnprintf() if you're sure that what you're copying
    should fit

  - strbuf or xstrfmt() if you need to handle
    arbitrary-length heap-allocated strings
smasher164··on Parsing can become accidentally quadratic because of sscanf
Sounds like a new entry for https://accidentallyquadratic.tumblr.com/?
smasher164··on Teaching Compilers Backward
I think if schools want to give a taste of the parsing->execution flow, they should have students implement an NFA based regular expression engine. No fancy features like captures and named character classes, just the primitive operations will do.

Regular expression syntax is context-free anyways, so they could use grammars if they wanted. Then again, you have to wonder if all of this should be covered by a theory or PL class instead.

smasher164··on Teaching Compilers Backward
Yes and no. The traditional goal of a parser that accepts valid inputs, and rejects invalid inputs is well understood.

A more relevant goal of error-tolerant parsers that can respond quickly to incremental changes is still a wide research problem.

smasher164··on Teaching Compilers Backward
Much to the chagrin of a lot of educators, I think this approach is the way to go. Too many compilers classes get bogged down in grammar classifications and parsing.

Some argue that parsing is a microcosm of the rest of the compiler, since it requires one to transform a program from one representation to another. However, this message doesn't really come across when you're operating on highly unstructured input, and not doing any simplification passes.

I think execution is really important for visualizing the effects of your work.

smasher164··on Tauri: An Electron alternative written in Rust
People act as if they haven’t dealt browser incompatibilities in the past. Just feature-detect/polyfill and move on.
smasher164··on Go 1.16 Release Notes
I do care about categorization though. Here’s an identifier parsing library of mine (https://github.com/smasher164/xid) where I can’t test that codepoints are in some category without knowing the version.

If unicode was an external library, it would be my responsibility to update this identifier parsing library.

smasher164··on Go 1.16 Release Notes
Excited for io/fs. The design work that went into it is hopefully going to lead to the level of interop that we see with io.Reader and io.Writer.

It's nice to see the float parsing speedups too.

I do wish the Unicode package wasn't tied to the Go release cycle. Unicode 13 was released in March of last year.

smasher164··on Golang generics proposal has been accepted
This is a significant milestone for Go, and I'm extremely happy for the community. I didn't imagine getting here when I first started using the language, and yet here we are.

Special props to Ian Lance Taylor and Robert Griesemer for their continued revisions of drafts, and exemplary discussion with the community in implementing feedback.

smasher164··on Making Win32 APIs More Accessible to More Languages
Now if MacOS followed suit, all would be right in the universe.
← PreviousPage 8 of 9Next →