Care to share the name of that mysterious ROM?

Sure thing!

Liquidsmooth. Running on a VZW Motorola Droid Bionic. Beautiful, simple. Fast. Easy.

Yeah, as if Cyanogenmod is ever going to release anything other than an Milestone release.

Srsly. They ditched 10.1 for 11, when 10.1 wasn't even yet stable? It's now most of a year later, and we still don't have a properly-stable, just-works release that doesn't change once a month?

And I say this as someone who actually likes Cyanogenmod, but found another AOSP 4.4.4 build that actually lets me use my multi-core >1GHz pocket computer with more than a thousand megabytes of RAM as it should be.

Nay, I have plenty of clue. But the private commercial networks I manage do not benefit from QoS, as all packets that transverse these networks are equally important and congestion is -- by design -- not an issue.

You really don't have a clue that it's even possible to solve this problem with a home gateway. Your perception clouds your vision.

Your loss, friend.

Now get off my lawn.

Again, you haven't used it. You're working with a theory, and really have no idea what you're going on about.

Latency increases somewhat under load (as it must), though not appreciably enough to affect any of the things we do with it. Jitter is very low as well. No matter how hard people or things hit the network, the user experience remains very responsive for interactive tasks...perceptibly the same as it is with an unladen connection.

This, as opposed to hundreds of torrent peers hammering away, one or more Netflix stream soaking up as much as it can get its hands on, and et cetera: Without QoS (and I didn't name it that, such terminology has been in place for quite a long time as relating to this sort of technique), this network was essentially unusable.

And now, it works reasonably. Individual TCP or UDP sessions are placed into groups with other similar sessions, and those groups have their own assigned priorities. This can be done by port, IP/MAC address, or deep packet inspection, or the amount of data the session has used, or even DiffServ flags.

It even has a fancy GUI that actually works.

You sound a lot like people used to sound back in the day, proclaiming that NAT (or ipmasq as it was more-commonly known at the time) could never successfully allow FTP, ping, or traceroute to seamlessly work. They'd list a lot of seemingly-logical reasons as to why it can't work and never will work, and then go on a long-winded rant about why either proxy servers or public IP assignments or at least one-to-one NAT is the only way.

Fast forward, and those people have STFU because -- gosh -- NAT works and does these things. They were ignorant of the possibilities of creative people making creative solutions.

I mean, sure: "Proper" QoS (ie: DiffServ and sensible routers with sensible queues and routes from end to end) might be nice. Maybe it even works on a private network. It doesn't work on the greater Internet, though, as you yourself say.

So rather than say "fuck it, I give up, there's nothing to do," I've simply solved the contention issues of my own grossly-overburdened last mile. And I've done it all from one side of the pipe.

If that seems impossible, then you're the ignorant one. There is a world of things that you did not learn in school, and some of them actually solve real problems that people experience. Nothing of this universe is so rigidly-defined that it cannot be adjusted in some useful way.

If you want to learn about it, though, download Tomato and spend an hour or two playing with it on a relatively-saturated network. Then read the source code if you still think it can't work that way.

Nothing I'm doing with QoS involves DiffServ -- at all.

That you proclaim otherwise shows that you haven't used the QoS features of Shibby's version of Tomato-USB.

And until you do, we don't really have much to talk about here.

The fact remains that I can rate-limit specific ingress UDP and TCP streams based on a number of parameters, leaving room in the otherwise-saturated pipe for other packets, using nothing more than an ancient freebie WRT54G and a small Shibby build.

How does this all work behind the scenes? I really don't care -- it's not my primary field of study. All I care is that it accomplishes everything that I said it does.

Apparently nobody here has used the QoS features in Shibby's version of Tomato-USB.

Torrenting and Netflix and gaming and multiple kids playing with Facebook and Youtube, all at once, all on 2Mbps of downstream while maintaining sufficient low-latency that interactive tasks and VoIP work fine?

Why not?

I did this just last night, as I have many nights before.

I'm perfectly capable of prioritizing my own bandwidth, thanks. I don't want my ISP prioritizing it on my behalf. Ever. At all. Not even a little bit. Not my wired ISP, and especially not my cellphone ISP where I pay by the gigabyte.

I'm not the person you're replying to, but: My (almost 20-year-old) car tells me when I have a brake light bulb out, and/or a failure in the brake light wiring and/or the mechanical switch that activates said lights.

When it tells me that something is wrong with the brake lights, I stop where it is safe to do so and fix the problem. I do this because I recognize the importance of showing the people behind me that I am, in fact, slowing down. (I also tend to have spare bulbs and a proper toolkit.)

I'm not suggesting that everyone would be so anal about their brake lights, but they should at least be informed.

With modern automobiles, where a network of microcontrollers is cheaper than a maze copper wires and multi-purpose LCD displays are standard fare, much of this functionality could probably be implemented in software alone.

And it should be. So people are at least informed that they're being jerks.

What sometimes confuses people is that Google also sells IT services

They also sell books, movies, TV shows, software, and hardware.

Which part of this very conventional process of exchanging money for goods and services involves me being a product instead of a customer?

I've been spending a lot of time away from my momma's basement* and have been mostly been hanging out at my special lady-friend's place.

She has a 2Mbps connection, and depending on who is visiting, there can be a half-dozen people actively using it. With her old router (a D-Link box that only supports stock firmware and DD-WRT), everyone hated the Internet here: It barely worked. There were nonsensical discussions about "how many people were using the Internet" when things would slow down.

With Shibby's version of Tomato USB, I set up some QoS rules on an old WRT54G. I gave her own laptop a slight preference, but really: With QoS, multiple independent Netflix streams are working OK, even though the boy streams Youtube almost continuously. The Sonos plays music from Spotify perfectly. Interacting with the Web is fast and responsive from all devices. Torrents don't bog the connection. Interactive ssh, RDP, and VNC are very usable.

Nobody complains now. Sure, downloads are slow, but downloads will always be slow here compared to at my own place in Mom's basement. Things just work, and the streaming stuff (except Sonos, which gets a high priority, because buffering audio is teh suck) scales to available bandwidth on a minute-by-minute basis, and it all seems to work fine.

Key points:

Prioritize small streams of all types: Who cares what they are, they'll be gone almost instantly anyway and their impact is therefore small. These are things like NTP, DNS, the initial handshake of every HTTP connection, and other little stuff.

Prioritize important stuff: The HTPC is probably a better streaming target than some stranger's iPod Touch who you gave your WPA passphrase to just to be a nice guy.

Progressively penalize larger transfers: A lot of loading a modern Web page is the initial HTTP handshake, and a mile of CSS includes. Getting these done quickly is not so expensive in terms of absolute bandwidth availability, but really does improve the user experience.

Eventually, the classes go to a "bulk" category: Something over a few megabytes, or somesuch: If the game/windows/whatever update takes forever, so be it, as long as you can still use the rest of the internet while that transpires.

One can also do the "deep packet inspection" game, which is well-supported in Shibby, and gain a little bit more control. But that's decreasingly useful as more and more connections are encrypted by default (and one cannot inspect an SSL-ish packet without also performing a MITM attack upon the whole connection).

Rant: This is what IP TOS flags are for, but they're almost universally useless because end-user programs STILL do not (or cannot) use them properly. But if they were used properly, I could totally accommodate that low-latency VoIP or interactive SSH session, at almost-zero expense to Netflix streams.

720p is HD. It's a perfectly cromulent resolution, is defined by ATSC (the HDTV specification for the US), is way better than the 480i we had for over half a century, and can be argued to be superior to 1080i for some content types.

There isn't anything wrong with 720p. Most modern flat-screen televisions (yep, I wrote that correctly: Most people don't have a behemoth TV) do native 720p at best, and it's actually just fine for the viewing distances and screen sizes involved.

That said, Netflix as of a year or so started doing (what they call) SuperHD. To my eyes (and no, I haven't done frame grabs to verify), it's 1080p24, and it uses almost all of a 12Mbps pipe (which I did measure).

You don't Lynx on a DOS machine with Telix.

You load a TSR to provide TCP/IP, whether via Ethernet or phone lines or whatever. And then you run a web browser from the command line.

Telix? No. Telix needs a host to do the heavy lifting.

Disclaimer: 15 years or so ago, I acquired a very odd XT with a floppy drive, integrated 10base2 Ethernet, monochrome (not HGA) video, and no hard drive. I tied that thing to the Internet directly with my always-on ISDN connection. It did telnet, ssh, FTP (along with a resident FTP server - yep, multitasking(ish) under DOS) and web browsing just fine, using native, local programs. With 640k of RAM. Booting from a pair of low-density 5.25" floppy disks, manually loaded one after the other . . .

It wasn't actually all that hard to figure out, and it was ridiculously reliable.

And it doesn't 'run out', it just gets slower at the shared wire level for the user. Which is why netflix looks like crap at 7PM every night.

No. Netflix looks like crap at 7PM every night because they ditched Akamai and started their own CDN which is typically backhauled by Cogent, and Cogent tends to have terrible connectivity.

AFAICT, the format of Mythbusters hasn't changed in a very long time.

There is a certain cadence of it which has not, AFAICT, varied since the show included "the kids."

One of the producers of Mythbusters is said to listen to it in her car, and if she can't follow and understand the episode by voice alone, it gets redone.

(As your attorney, I think you're trying too hard to coalesce your own aging mindset with the continually-renewed world around you. Give it a rest. Things move on.)

I'm actually rather OK with this. Though Kari is fun to watch and has certain skills (particularly welding and being hot), and Grant is very talented with nuts and bolts and software and robotics, I actually like Mythbusters mostly for the hard science (even if wrong) of Jamie, and the manic presentations of Adam.

Who was the other one again? Oh, yeah, that other guy.

Anyway, I remember Mythbusters with just Jamie and Adam. I miss it: Two well-versed, very smart people arguing against each other but toward a common goal is a win every day.

As I recall from my understanding at the time, BP has privately-owned and operated franchises. Like any franchise, the franchise license comes with certain contractual obligations.

One of the obligations for the first year or two of a BP franchise is buying BP gasoline.

After that, and again IIRC, it's open market: The BP (or AM/PM, or whatever) station can totally buy Marathon gas from the Marathon distribution point across town, instead of BP gas from the next state.

There is nothing wrong with this. Gasoline is generally a commodity, and about the only thing that keeps it brand-centric is the additive package which is (or may be) mixed differently for Shell or BP or Mobil or whatever.

So no, I didn't avoid BP stations after the BP gulf oil spill, because: Meh. I already knew better: Chances are, the owner was already buying whatever gas he wanted, according to market demands, whether sourced from BP, or Shell, or Marathon, or Exxon, some other such entity.

Punishing a BP franchisee for an oil spill is a cause which is based on a red herring, and is therefore nonsensical.