Didn't tethering fees get clobbered by the FCC? The IDE is pretty light on bandwidth, the initial pageload is about 2 MB and it's just shuffling text around during use. It has a keep-alive ping, but otherwise you're only going to use bandwidth while saving changes or using the terminal. How much bandwidth does a terminal use? I recently signed up with PTel, which uses T-mobile's towers and gives you unlimited 3g / 1 GB 4g for $35/mo, no contract. I think a month's worth of coding would run substantially under 1 GB of bandwidth but I don't really have the time to do a rigorous test.

I found interpreters for Python and Brainfuck on the Chrome Store, and of course you have a JS interpreter, and any interpreters written in JS should probably work. There are rather a large number of those for some reason. There's some sort of git app too, FWIW. Beyond that there are a few Android apps that will run natively on ChromeOS without any fussing, and most Android apps can be made to run with minimal effort.

I don't know what you're coding in, but unless it's fairly obscure I'd say it's possible to code and test, offline, using a Chromebook. Either way I hope no one is twisting your arm to get you to buy one.

If a quick Google search can be believed, you can actually get a free VPS. For an IDE I'm using Cloud9, but there are equally good or better alternatives. However, I am already paying for mobile data and a VPS for other reasons. Even so, I'd still probably get out my laptop on a bus only if it was a Google bus. Or maybe Greyhound, if it had wifi. I very rarely find myself without an Internet connection, even in rural Central America. When I don't have an Internet connection, generally I'm not doing anything where I care about having one.

It's not even that I couldn't code without the Internet; there are code editors for Chrome/ChromeOS. However, not having access to API documentation would be a huge issue (for my work), and that would make OS deficiencies a moot point.

I need the Internet for work. Having to have a net connection in order to use a decent IDE is not ideal, but it's a low bar even in rural Central America, or rural Alaska. I don't really understand what it is about the idea of an Internet-only device that bothers you so much, but I am actually pretty sure that you would be less inconvenienced than you imagine.

My first impression was, "WTF?! Why would anyone want to do that?" Keep in mind that not only am I typing this on a Chromebook, I basically live on this thing. For what I use it for, it works well. With a web based IDE and an SSH client, you can accomplish almost anything. Entertainment is not a pleasant situation but that's what we have gaming PCs for, right?

ChromeOS does actually have some nice features. It's nice to have updates that only take fifteen seconds, including a full reboot. The battery life is great, and it's really cool to be able to sit down at a brand-new Chromebook, type your google username and password in, and have all of your bookmarks, apps, and files available within 30 seconds. The thing is, I really don't think you're going to be able to get those same features with any other combination of hardware and software. As you point out, the boot speeds are likely not going to be any faster, and I would be surprised to learn that the non-Google versions of ChromeOS had the same, ah, vendor lock-in.

I'm very ambivalent about ChromeOS. It looks nice, it's very secure, it has a number of good features, and I feel like it is particularly good for schools. I've been able to make my Chromebook do what I want, and having a pair of them was really great for wandering around Central America for a year or so doing freelance web development. They're cheap enough to be more-or-less disposable. On the other hand, it's very much not a replacement for a real operating system. The good thing is that it sounds like the OP doesn't need a real operating system. The bad thing is that he probably isn't going to get what he likes about ChromeOS out of this either, no matter what he does. A stripped-down distro is probably the better option.

As an aside, I also share your sentiments with regards to the swapping issue. I've had a bunch of netbooks in addition to this Chromebook, and I've had real Linux running on this machine via both crouton and a direct install. With ChromeOS, I can only have 30-40 tabs open before it starts killing tabs to free up memory, and fewer than that if the pages are resource-heavy like gmail, disqus threads, or videos. In my experience ChromeOS has far more memory issues than other distos on the same or worse hardware. However, I will say that ChromeOS's failure mode of killing pages early and often works very well to prevent the machine from ever becoming unresponsive due to memory/swap issues. It's kinda hard to pick between those two problems, to be honest.

Why bother lying about Gitmo? I mean, yes, it's useful as an extraterritorial prison, but attributing its continued existence to Obama is bizarrely counterfactual.

Obama issued orders to close Gitmo in 2009. Congress fought back with appropriations bills. The GOP has been and continues to be hugely critical and combative with Obama's attempts to close the detention camp. Romney was openly supportive of it, and a Republican Senator has said the Gitmo detainees can "rot in hell". Are you just completely ignorant of everything that has happened until this point, or are you arguing the President should just ignore the law, Congress, Republicans, and 53% of the country and close it anyway?

Ironically, the 868L is listed as having the second-highest throughput on the page you linked. It's very strange that mine isn't working correctly. Maybe alternate firmware will help things. The desktop and the ISP-supplied Actiontec get 890 Mbps on speedtest.net, and it's not like PPPoE is computationally expensive. Thanks for the link, it was informative, depressing, and hope-inspiring all at the same time.

I have a DIR-868L, it was cheap(-ish) and reviews suggested it had good (unobstructed) wireless speeds. That may well be the case, but unfortunately it has a more serious flaw, only being able to handle about 350 Mbps of my gigabit connection. I'm pretty sure the hardware is capable, but the firmware is crippled. I've already RMA'd one and got another back with the same symptoms. Apparently D-link engineers are trying to reproduce this issue, but I don't really expect them to do anything about it.

So, I'm looking for a little advice here on one or more of the following topics:

• * Choice of Alternate Firmware
• * Firmware Installation Tips
• * Better Gigabit Routers

Additionally, although too long for a bullet point, I'm interested in the viability of simply getting a wireless adapter for my desktop and just using that as the router. The internet is supplied as a simple PPPoE / CAT6 connection, so it's not exactly hard to set up (how D-Link could screw this up would be mystifying but for things like TFA). There are a handful of other devices on the WLAN but wireless throughput is not really a huge concern; I don't yet have any 802.11ac devices so I'm not going to get full speeds to them in any circumstances.

Your sage advice is greatly appreciated. Thanks in advance.

Would you translate that as "whateverism"? I think it works better in Italian. Good word though.

Because you're not used to thinking about source code the way Git thinks about source code. Git is very much like a database from a usability standpoint, and you will probably get into bad trouble trying to use either without understanding both the problem that they are trying to solve and the implementation. If you do read about these things, you will understand that git's internals make sense, the decisions it makes are logical, and the user interface is (mostly) transparent and simple. Revisions are harder to manipulate than a Word document, though there are plenty of ways to manage them that are conceptually simpler. Git however was made to manage them efficiently. More specifically, it was designed to be efficient for Linus Torvald's workflow. That happens to be very effective for a large number of other software projects, and no worse than any other solution for many others. There are other workflows for which other RCS systems are better (particularly when working with binary files). If you don't need git's features, by all means use something else. However, your decision to use it or not should probably be informed by knowledge of what exactly it does and why: again, this is no different than choosing a database.

It's not theoretically impossible, it's actually impossible. Kernel features equivalent to cgroups do not exist on other kernels. BSD jails are not remotely the same thing. If you can demonstrate otherwise, you'll begin to have a valid point, but you'll still have the question of whether it is practical (your example is too trivial to demonstrate this) and whether anyone actually wants systemd on other platforms.

If you're just going to spout uselessly vague aphorisms though, save yourself the effort of typing a response.

Polemic, I will grant you. I feel you are being extremely vocal and equally dishonest in your criticism of systemd.

Non-portable stuff happens all the time and doesn't hurt anything, and in the long run all software is obsolete—even BSD. Systemd is a service management framework that depends on Linux-specific kernel features. I will grant you that in an ideal world as much of the OS as possible should be system-agnostic, however this is not an ideal world. For one thing, functional equivalents do not exist in other kernels, so portability is a moot point. Also, it would probably be a bigger change for other OSes than just process management; I'm not sure BSD is interested.

SysV init was portable because it was trivial. It was also deeply (but subtly) flawed. You could be very nearly sure that a process started or stopped when it should, but pidfiles have always been a bad hack around missing kernel functionality. Which do you want, working process management or BSD compatibility?

You're mischaracterizing his remarks. He's not going to try to find functional equivalents for cgroups on other kernels. Please explain the problem with that, and note that while "I would like all kernels to have the same (or equivalent) feature sets" is a problem, it is mostly your own mental problem. Also, a justification of why an OS plumbing layer should be OS-independent would be nice.

There are better ones. Why do we care about this project, as opposed to the dozens of other (better) NES emulators?

Users may or may not want stable software. Developers want to be able to use new CSS features. As for the Firefox developers, I'm not sure what their UI goals are, but presumably they have them. That's two strikes against stability.

I suspect you have no idea how difficult it is to support older browsers, as a web developer. The development workflow usually goes like this: first, you code to the spec, and test in a browser that reproduces the specification well. Then you start trying to find out why it doesn't render correctly in Internet Explorer. For each incompatibility, you have to add a workaround. Then continue this process for as many old browser versions as your client has money for.

It is not free to continue supporting old browsers. It's usually shockingly expensive, frustrating, and very limiting for the developer. If you don't like something about your browser, please become involved in its development — file a bug report if nothing else. Please do not use old, incompatible, insecure browsers.

However, if you ever want to induce a mental meltdown in a web developer, tell them their new job is making sure all the sites are IE6 compatible. Even just saying "IE6" a few times can cause convulsions in susceptible developers. I'm not brave enough to try the same trick with earlier IE versions though: Cthulhu alone knows what kind of eldritch horror might be summoned.

Do regale me with your interpretation of what "appeal to emotion" means. I didn't address most of what you wrote, you're right. You don't really have a clue what you're writing about and I would rather have you shut up (or at least tone down the paranoia) instead of taking the time to educate you. It's not your field, you don't understand what is happening or why, it's better if you just stick to your expertise.

If there is one point of confusion that I do want to clear up, it would be that systemd is not a simple replacement for init, and was never intended to be. Or, if you rather, the job of init was not simple and it failed spectacularly to accomplish it. With SysV init, there was no effective service management. You could start and restart services, but imperfectly, and I don't know a nice way to talk about double-forking and writing to a pidfile. That it works most of the time does not mean that it's not deeply flawed.

The fundamental conflict here is between old-school UNIX admins who think that the OS just exists to launch user-defined scripts, and the requirements of a modern OS. No one is trying to say that having scripting access to the OS is not useful, although you Windows users seem to do without it, but the interpreter is not the OS, and Bash makes a shitty replacement for C. Even OpenRC recognizes this. It's kinda mystifying that you've jumped on the opposite bandwagon; Bash scripting is about as useful to you and your customers as tits on a boar. Can you imagine if the solution to some Windows issue was to fire up a command-line text editor and edit an init script?

Linux needs to grow up, and so do systemd detractors. Doing things in Bash is the UNIX way, but so is cleaning those scripts up and rewriting them in C. Try some Bash scripting on for size if you think you disagree, and don't forget to double fork and write that pidfile!

The NSA comments are baseless ad hominem. You're better than that.

P.S. By the way, I think I've asked before, but where is your store located? Our disagreements aside, it'd be nice to know if I were ever in the area to be able to pay my respects.

You don't know anything about what you're talking about, per the usual. It looks like you've been listening to gweihir too much. There were at least two other projects looking to replace init, and half a dozen more providing some form of process tracking. Lumping two things that you hate and fear together must feel good, but systemd's development and adoptation has been a public process, and both systemd and pulseaudio offer important technical improvements (that apparently you don't understand). I realize all you do here is go off on ill-informed rants, but this one is only tangentially connected to reality; you might want to dial back the frothing a bit. I mean, you don't even use Linux — what's it to you?

P.S. That bit about "by the time he's done the kernel will be just a VM running on top of systemd!" was hyperbole, right? Because otherwise it's so far beyond crazy that it makes me wonder if you know what those words mean.