Speed depends on the gear that the merchant has, and the process that the consumer uses.

At Kroger, my debit-based magstripe credit card transactions happen nearly instantly -- less than a second -- and I don't have to sign anything if it's under $50.

It goes like this: Swipe card, begin to think about putting it back in my wallet, and the receipt printer goes *whoosh* with a flurry of thermal paper. Put card in wallet, and the clerk is handing my receipt to me by the time my wallet is away.

It takes longer than that less-than-a-second to physically open and close the til to be able to handle cash, let alone actually transact with it.

And if I've got my wits about me, it's even faster: Before the items are even done being rung up, I've generally already got my credit card swiped and selected and put away.

Same with the phone and Google Wallet, at the few places that actually accept it near me: I've generally already done the tap-to-pay thing before the clerk is done ringing things up.

Hmm. You know, I've never had an old, proper WRT54G/S (or the current GL model) die from heat death. I've got dozens of them scattered around. Radios get weak or strange after awhile (electron migration of somesuch), and maaaaybe I remember some swollen filter caps on one (which got repaired), but I don't consider any of that heat-death (and it's not like bad caps weren't ridiculously common for a time from almost every manufacturer of almost anything).

I've had the power supplies dive on me, which is problematic. I find that the old linear supplies are far more reliable than the new switch-mode ones, so I tend to install them with overkill power supplies. (Asus, my current go-to cheap router-wifi-box maker, is no better when it comes to just plain garbage for wall-warts.)

The modded GS I have, I did attach a heatsink to the CPU because I was overclocking it for fun. But that doesn't count. :)

By early routers that didn't route, I mean the BEFSR-whatever-it-was style of garbage that reared its ugly head back when I was still using a *nix PC for routing at home. Grossly inadequate and broken, like a SCSI adapter that never quite works right (even with active termination, new cabling, and the goat blood). Or a client of mine that had a fancy metal-boxed Linksys wired router with many ports and some sort of VPN functionality: It was wonky from day 1, from the complaints. I replaced it with a random (but non-Linksys) switch and a WRT54GL running Tomato, and never had to troubleshoot that side of things ever again.

By switches that suck, I mean blocking 10/100 switches sold for a premium in the day of cheap non-blocking 100mbps with auto MDI/MDI-X. A then-cow-orker bought a bunch of them and scattered them in the field, and they all got replaced with something (anything!) different within a year or two.

Hasn't that always been the case?

They struck gold with the WRT54G and WRT54GS (I still have a modded GS as a spare). But everything before or since has been garbage.

Their nics are garbage, their switches always suck, and their early routers largely didn't route.

Just sayin'.

The Internet of rushed to market, horribly secured, never updated, easily pwn3d things.

Is that a new problem?

(To answer my own question: No, it's not.

DD-WRT seems so splintered: A million different builds, of a million different versions, for a million different things.

For comparison, Tomato is more monolithic. When a new version is prepared for release, all of the different builds are updated to that version. The builds themselves are genericized as much as possible: All old Broadcom-based MIPS routers (think WRT54G) get the MIPSR1 release, for instance.

For everything else, there's OpenWRT.

For my own purposes, I'm sticking with Asus routers. It seems like solid kit, and they sell the same hardware for years and years without the sneakiness that Linksys and Netgear do with routinely completely changing the underlying hardware while keeping the same model number.

(Oh, and Belkin has owned Linksys for almost 2 years now.)

Indeed.

I didn't learn much math in school, but I caught up on everything I needed as an adult -- as I needed it.

I didn't learn any programming in school, but I caught up on everything I needed as an adult -- as I needed it.

I know as much math -- and programming -- as I need to always accomplish the things that I'm trying to do, whether calculating volume displacements, de-rating wiring, or hacking up a strange combination of awk, sed, and perl to homogenize a random dataset into some higher-level program that expects it to be formatted -- just so -- in CSV.

Problem solved, I then move on to the next challenge....learning as I go.

Mostly what I learned as a kid was how to learn. The rest? An academic would say I fly by the seat of my pants, while my friends, associates, and employers respect my ability to solve arbitrary new problems accurately and quickly.

In practice, RDP can be a security risk. It can be mitigated, but hasn't been in practice.

Is this security risk any different than any other method which allows relatively-insecure* remote logins to a general-purpose computer?

For a long time, I had the default RDP port forwarded to my Windows 7 desktop computer at home -- open for all manner of fuckery -- out of sheer laziness on my part. Nothing bad happened, even though it was simply a username/password combo (and both the username and the password were relatively common English words) to gain access.

*: No two-factor, no cryptographic authentication, no....

A crude proportional (self learning) controller is desired.

Is there anything in your post that is not solved by a Nest thermostat, other than monitoring boiler temperature and comparing that to room temperature?

And as the house temp rises close to the thermostat setting, the circulating water temperature could become less hot.

What merit is there to doing that, anyway? It seems to me that adjusting boiler temperature based on outside temperature is the best method, and (based on your description) you're already accomplishing that automatically: When it's colder outside, you want hotter water. Right?

I mean, the efficiency does improve with cooler water, but the efficiency of the system is already determined based on the amount of work that it has to do (which is dictated by outside temperature). What's wrong with that?

Lowering the water temperature based on inside temperature seems like a fool's errand: Once the system finally gets caught up to the thermostat setpoint, it would lower the water temperature. As they night (or day) gets colder, the water would continue to be cooler. Until it's so cool that it's no longer usefully warm, and then it has to play catch-up.

This will, at worst, cause temperature oscillations (even with the simplest mechanical thermostat running the circulating pump), and at best increase your time constant dramatically on days when you need your heat to be working at its best.

And in both of these cases, your family will hate you for it.

For instance, the weather is cooling off 20 degrees F over a period of about 12 hours here where I am, today -- mostly during the day. The thermostat is keeping things warm just fine, which in your hypothetical case would mean a decrease in water temperature -- when you really, really would be needing the thermal mass of all that water to be hot, tonight.

I'm going against my normal judgement here, but...

Consider that most people think of their phone as a device that already does things, and not as the general-purpose pocket computer that a lot of folks on /. might.

They are probably happy that their new phone handles their email, browses the web, can open a PDF or Word document, and back their photos up to a cloud service. They may not notice (or care if they do notice) that it behaves differently than any other phone.

And, generally, the OEM pre-loaded productivity apps tend to do these things pretty well.

So is it a problem?

I mean, sure. We geeks will always root and use Ti Backup to freeze all unwanted apps, and pick our own apps as we deem fit for our own purposes. (We'll also install a magic ad-blocking hosts.txt and conduct other tomfoolery, because we can.)

The unwashed masses get a device that tends to do mostly what they want out-of-the-box with regards to OEM preloads.

This seems to be very different than the stuff that is genuinely crap (such as nearly every carrier add-on, particularly including NFL Mobile which Verizon has been pre-installing on Droids since they started selling Droids).

So the winner on the price-performance curve is still...X10?

The 100 Watt-ish LED.

The high-CRI LED.

You don't get to have both, just yet.

My Samsung S5 groks SD cards just fine, and I've ejected Knox.

It roots with towelroot, just like many other devices.

*shug*

Indeed, that does seem to be a problem. (I'm currently happy with my old and dumb 52" Samsung LCD, and will not be replacing it until 4k content becomes a non-streaming thing, so I didn't realize that "good" new TVs are all smart.)

So buy a smart TV, and then never, ever connect it to a network. Just add a Chromecast or a Pi 2 with XBMC or a FireTV stick or a Roku or an HTPC with Plex or XBMC, for family-friendly options.

They're all less than $100 (aside from a proper HTPC). Using multiples of the same device for different sets around the house, is also a usability boon: Everything works the same, no matter what room of the house you're in.

I'm in.

But big, dumb appliances are going to be hard: I've been through Whirlpool's washing machine and dishwasher plants in Ohio, and they are vast things indeed.

Big dumb pipe: Google is trying this, AFAICT.

Big dumb panel: Element is already doing this, assembling foreign subassemblies into flat-panel televisions in the US. They don't, however, have a high-end line with excellent panels every video connector imaginable. (Also, I want it to be 600Hz, so that it can do 24fps, 25fps, 30fps, 50fps and 60fps without aberration.)

The easy answer: Don't buy a smart TV. There are other, often much cheaper, options to network-enable a dumb TV.

Buy a TV because it has the glorious pictures that you find preferable at a price that you can justify, not because it's "smart." Buy a networked HDMI dongle/STB because you like its features and interface.

Keep the functions independent, and you'll be in far better shape -- both financially, and functionally.

And you'll never have to ask this question again.