Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

System76 Pangolin Linux-first laptop with AMD internals now in stock

system76.com

Post reply on HN
Print view View original

441–448 of 448 posts

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#441

Post by CRConrad »

occoder wrote: post #426
CRConrad wrote: post #418

Earlier quoted context omitted.

Why on earth should they make the numpad keys double as arrow keys?!? Then you couldn't navigate and enter numbers in, say, a spreadsheet without futzing around with NumLock all the time! Utterly ridiculous idea. That keyboard in your pic looks pretty perfect as it is.

I hate to break it to you, but the numeric keys already double as arrow keys on that "perfect" keyboard. Besides, this is a laptop running Linux, what do you think is the more likely target audience? People entering numbers in a spreadsheet, or people doing software development? Most of the comments here are people complaining that the numeric keypad is there at all .

> I hate to break it to you, but the numeric keys already double as arrow keys on that "perfect" keyboard.

I hate to break it to you, but they can't be both simultaneously. They only "double" if you have a mind-reading genie hovering above your shoulder and pressing NumLock for you whenever you need to switch from arrow-mode to number-mode and vice versa. I don't have one of those; do you?

> Besides, this is a laptop running Linux, what do you think is the more likely target audience? People entering numbers in a spreadsheet, or people doing software development?

A) This Clevo-clone keyboard is exactly the same as on Clevo-clone Windows boxes. Windows spreadsheet users outnumber Linux software developers substantially, so that's who the market produces hardware for.

A) Personally, I kind of do both; I often (well OK... But not exactly rarely) use a spreadsheet full of numbers to generate code.

C) One gets the feeling that the main driver behind this particular comment was contempt for the ordinary spreadsheet-using "sheeple". At least, condescension seems to come naturally to you, Mr. I-hate-to-break-it-to-you.

> Most of the comments here are people complaining that the numeric keypad is there at all.

Sure. And they are (quite rightfully) clamouring for keyboards without one... But what does that have to do with this? Not wanting a numpad at all is a very different (and far less misguided) idea than severely crippling the usability of both the separate arrow keys and numpad on keyboards that have one, which is what your suggestion boils down to.

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#442

Post by CRConrad »

ncmncm wrote: post #427
CRConrad wrote: post #416

Earlier quoted context omitted.

I thought it was from boxing?

I am corrected. The usual story is that is is about pitchers, but the timeline is wrong. The first official baseball game is cited as in 1846, with "southpaw" occurring in 1848 already referring to boxing.

Cool, thanks for confirming (i.e. doing my research for me :-).

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#443

Post by burntsushi »

jcelerier wrote: post #421
jcelerier wrote: post #415

Earlier quoted context omitted.

I was on my phone and trusted the Google results of "ripgrep sloc", thanks for the correction ! What takes so much code in the ripgrep codebase ? I am developing a cross-platform GUI app with audio, video and network features and it's not even 500k lines

hmm, but I just cloned the ripgrep repo and it tells me 28k loc ? $ cloc . 193 text files. 170 unique files. 114 files ignored. github.com/AlDanial/cloc v 1.88 T=1.88 s (73.3 files/s, 25634.5 lines/s) -------------------------------------------------------------------------------- Language files blank comment code -------------------------------------------------------------------------------- Rust 85 3984 9128 28104…

I provided precisely the commands you need to run to get the output I got. Not only did you not run 'cargo vendor', but you're using 'cloc' instead of 'tokei', which has a less precise line counting algorithm.

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#444

Post by burntsushi »

jcelerier wrote: post #415
burntsushi wrote: post #412

Earlier quoted context omitted.

Uh... No. You're off by an order of magnitude. $ cargo vendor $ tokei vendor/ crates/ -t Rust -s code --exclude winapi ------------------------------------------------------------------------------- Language Files Lines Code Comments Blanks ------------------------------------------------------------------------------- Rust 1248 615700 493777 83201 38722 And 8 minutes for compiling ripgrep sounds crazy. My i3 Mac Min…

I was on my phone and trusted the Google results of "ripgrep sloc", thanks for the correction ! What takes so much code in the ripgrep codebase ? I am developing a cross-platform GUI app with audio, video and network features and it's not even 500k lines

I am counting the number of lines that are being compiled when you run 'cargo build' on Linux. This includes things like huge Unicode tables and ffi bindings (for the entirety of libc, for example).

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#445

Post by jcelerier »

burntsushi wrote: post #443
jcelerier wrote: post #421

Earlier quoted context omitted.

hmm, but I just cloned the ripgrep repo and it tells me 28k loc ? $ cloc . 193 text files. 170 unique files. 114 files ignored. github.com/AlDanial/cloc v 1.88 T=1.88 s (73.3 files/s, 25634.5 lines/s) -------------------------------------------------------------------------------- Language files blank comment code -------------------------------------------------------------------------------- Rust 85 3984 9128 28104…

I provided precisely the commands you need to run to get the output I got. Not only did you not run 'cargo vendor', but you're using 'cloc' instead of 'tokei', which has a less precise line counting algorithm.

but, cargo vendor will fetch the dependencies, which should definitely not be counted as part of the source code of a given program ? else every program should also count the kernel, graphics driver, etc.. which would make it a pretty useless metric

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#446

Post by LocalPCGuy »

occoder wrote: post #407
LocalPCGuy wrote: post #399

Earlier quoted context omitted.

I will never buy another laptop with a "short" right-shift key and the up-arrow next to it. For coding, it SUCKS to miss hit that and I never got use to it on the one laptop I've had with a keyboard like that.

I must admit that I've never used such a keyboard before, but it seems better to me than half-height arrow keys, or heaven forbid, half-height Up and Down arrows but full-height Left and Right arrows, what an abomination! What I don't understand, if they're going to include a numeric keypad on the right, is why they didn't make the bottom of the numeric keypad the arrow keys. Looking at [1], 2 should be Up, 0 should…

I have had both the short right shift and the half-height arrow keys on different laptops, and the half-height arrow keys are by FAR preferable for me. Much less mistyping that way. Granted, I don't do much gaming on a laptop (although I doubt I'd use the arrow keys that much for that anyways, probably ASDW).

But I totally agree that the people putting keyboards together don't seem to actually spend much thought and definitely don't user test it with representative users, including keyboard-centric users like coders.

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#447

Post by burntsushi »

jcelerier wrote: post #445
burntsushi wrote: post #443

Earlier quoted context omitted.

I provided precisely the commands you need to run to get the output I got. Not only did you not run 'cargo vendor', but you're using 'cloc' instead of 'tokei', which has a less precise line counting algorithm.

but, cargo vendor will fetch the dependencies, which should definitely not be counted as part of the source code of a given program ? else every program should also count the kernel, graphics driver, etc.. which would make it a pretty useless metric

This conversation is super frustrating. This is occurring in the context of compilation speed. So if you're measuring compilation speed by LoC (which is itself just a heuristic), then you should, at minimum, count as many lines of code that are actually being compiled. That's what `cargo vendor` does: it fetches the source code of all dependencies that are needed to build the project.

You are losing sight of the forest through the trees here. This isn't an abstract argument about what is or isn't part of a program. This is a concrete measurement heuristic that has precisely defined inputs.

Re: System76 Pangolin Linux-first laptop with AMD internals now in stock

#448

Post by jcelerier »

burntsushi wrote: post #447
jcelerier wrote: post #445

Earlier quoted context omitted.

but, cargo vendor will fetch the dependencies, which should definitely not be counted as part of the source code of a given program ? else every program should also count the kernel, graphics driver, etc.. which would make it a pretty useless metric

This conversation is super frustrating. This is occurring in the context of compilation speed . So if you're measuring compilation speed by LoC (which is itself just a heuristic), then you should, at minimum, count as many lines of code that are actually being compiled. That's what `cargo vendor` does: it fetches the source code of all dependencies that are needed to build the project. You are losing sight of the for…

.. you mean that the dependencies aren't downloaded as binaries but compiled from source every time ? damn
Post reply on HN