Posts by Henry Wertz 1

Surprised

I'm surprised there's this much backlash, honestly.

I was surprised when LogMeIn dropped the free service, instead of just having a trial or putting service limits on it or something (after all, how are they going to grow their user base? I wouldn't have recommended buying LogMeIn service if I hadn't been able to try it first...) But I didn't end up with any problems with the service.

Is it fears they'll force-bundle LastPass with some other junk? Or that they'll turn it into some mutant hybrid of LastPass and LogMeIn's password service (that I'm not going to bother scrolling up to see what it's called?) Other problems with LogMeIn? General buyout fears?

Does seem odd...

Does seem odd. To be honest, I doubt ICANN are being bigoted, they probably are rigidly following some procedure. But I do think they need to reconsider this procedure if this gets a 0 while .spa or .radio get enough points to pass.

Good luck with that

First off: Hey Feds, good luck with that. It's business suicide for a business to slip intentional backdoors into a service in this day and age, and it WILL be found out. And probably exploited by hackers. And spread all over the media, so you'll lose all your business.

"Why doesn't Obama just issue an Executive Order telling them to stop demanding back doors?"

Because he doesn't want to. Like most members from the US's main two political parties, he will say what he thinks people want to hear at election time, so I'm sure at some point he vaguely intimated he'd do something about this little surveillance problem. But (based on his actions in office) he is actually a staunch supporter of ubiquitous surveillance. And based on both words and actions in office, he's a supporter of the NSA's programs in particular.

"Do they really need such a device when they have a cell with Bubba waiting for Contempt of Court?"

They don't need it either way, they don't have to be omniscient. Unlike UK, however, as much of a zeal as the US's main 2 political parties have shown in ignoring the constitution, if you're asked to give up crypto keys in the US you can still take the 5th (invoke the 5th ammendment right against self-incrimination.)

Good on them

"That would not really explain any shortcomings in London, though. On the face of it FTTC should be easy as it is densely populated and would not require fibre to be pulled in over great distances just to get anywhere near the customers. "

Actually, old and densely populated cities are the best to run fiber through for this reason, but the worst to run fiber through because there'll be layers of pipes, sewers, subways, possibly older abandoned subways, cobblestones and things instead of regular pavement, and relatively densely packed buildings. Also, there tends to be a relatively high rate of "NIMBYism" ("Not in My Back Yard") where the same people whining they can't get faster speeds will then show up at planning meetings to complain they want to dig up their street (or in the case of phone service... they'll whine they get bad cell phone service but then show up at planning meetings to rally against having any cell sites put up.)

Anyway, here's hoping they roll some G.Fast here. And that CL starts following normal industry standards for line length, instead of offering about half the speed they could be. I've got the worst of both worlds here - CenturyLink ran fiber to the cabinet to the front of the mobile home park I live in, and this runs VDSL2 rather than ADSL2+... so I cannot use a regular DSL/ADSL modem. Unlike DSL, ADSL, and ADSL2+, I have not found a single non-carrier-branded VDSL2 modem on the market in the US. Despite using VDSL, they only offer 7mbps at my distance (while line stats suggest the line should be able to do at least 15mbps on ADSL2+ and faster on VDSL.) They'd charge me 12mbps price for that 7mbps too, since they don't have a 7mbps plan any more.

Several points...

I must make several points...

1) I don't know that (as a practical matter) these will actually improve security. Presently, it's a mess, if I go to one store (with my mag stripe card) the checout never asks for my PIN; if I go to another, they ask for my PIN if I run the card as debit but not credit. I've seen a demo (linked off the Register!) probably 5 years ago where someone (with card holder's permission) cloned a chip card and ran a transaction on it (and the UK banks insisted it is secure even after being shown the video.) Apparently, these cards being rolled out in the US are not chip'n'pin, but chip'n'signature. I think require PIN is the best way to ensure security.

2) Theory aside, as a practical matter, I've had *one* ATM that did not want to read my card, and zero checkouts act up reading it. The door cards, I doubt anyone was responsible for cleaning the reader; on ATMs presumably the ATM owner cleans the reader when they add cash, and at checkouts I'm sure if nothing else the cashier cleans the thing when it starts acting up for them. Or they're self-cleaning, or immune to dirt... I don't know, I'm just saying I have not had mag stripe problems even to the degree I would expect, let alone what you'd expect based on those mag-stripe door locks.

3) I was going to comment how I don't want a card that can be copied or have a transaction run against it while it's still in my pants. But it turns out, these cards are not NFC, they rely on contact with the card reader! 8-)

Everything's not a database.

"At their core, if a program makes use of storage, it's an effing database! "

No it's not.

Javascript

I do disagree with "no javascript". But how about "think about what javascript you use?" Those pages that seem slow? I have written up a Greasemonkey script to log settimeout and setinterval calls.. they aren't even animated, but some of these banner ads javascript is so inefficiently coded that they'll call some function 100 times a second just to check if it's time to rotate the ad yet.

I would say "keep the javascript, but lay off the timers." Again moderation is the key - I'd say feel free to use timers when you need too, but it's typical for a site with no apparent activity (page and ads have loaded, and nothing's animating) to have 5-10 timers running every 100ms (10 times a second), a few running every 80ms (~12 times a second), so you'll have like 100 calls a second to various javascript, and that's when there isn't a banner firing off yet more every 10ms (100 a second).

This seems to be the problem. I've been to sites that make heavy use of javascript (but no random banners) and it runs great, it's all those timers that hurt things.

Yup

"So we're in a race to the cellar for cost/compute and watts/compute. Effectively TCO/compute in my book. I just can't see Intel being any more worried in that market, here or in China."

Yup, I didn't think Intel had anything competitive, but Intel's got some low power Xeons that use much more power than contemporary ARM, but also run quite a bit faster, so the performance per watt is pretty good. Atom also, (pretty slow but low power), but apparently Intel's planning to keep it for "consumer" usage and not for servers. Whoever gets best instructions per clock will be used. I guess Google, Facebook, etc. are looking at Xeons, Power8, and ARM primarily, but whatever gets bets performance per watt and density will be what they use. They don't have a bunch of purchased non-portable software, so they have no actual reason to run Intel if there's something better.

I can tell ya, back in the day I ran Linux on an ARM, MIPS, PowerPC, Alpha, PA-RISC, as well as x86. Debian do a good job of this, you had a complete distro on all of these, not some noticeably smaller subset of packages.. To bootstrap linux on a new CPU you basically get gcc, the kernel, and glibc to build on it (there's a few lighter libc for embedded use too), and (in Debian's case) a build system builds all other packages from source, logging which packages failed to build. Once your GCC and glibc are up to snuff, everything should build (it won't try to build something platform specific like virtualbox, it'd be flagged x86-only). ARM, and POWER/PowerPC have been supported for years, modern ARM and Power8 are already supported so I don't think a company with a home-built Linux-based software stack should have huge problems porting over if they prove more power-efficient (or to a future power-efficient chip, once it's got gcc, Linux, and glibc on it everything else should follow and you end up with a full distro on there too. Hopefully the C/C++ portions of your custom software stack build and whatever else (python or shell script or whatever) should just copy straight over.)

So name and shame?

So name and shame? Anybody? I want to know which vendor or vendors to avoid. I'm disinterested in using vendors who use legal threats to bully security researchers instead of taking their lumps and fixing the products.

Not interested

Don't get me wrong... I'm not going to go rallying against 4K sets. Since LCDs use conventional techniques, so making the same-sized 4K LCD panel should cost about the same as a 1080 panel. The decoder chips should cost about the same (decoding H.264 or -- hopefully they come to their senses and use H.265 -- shouldn't cost much more than an MPEG2/MPEG4 decoder already does.) If the 4K and 1080 sets cost the same (within a few dollars), well, OK.

That said -- I think 4K is useless. Personally, I'm not buying new TVs, period, as opposed to computer monitors. HD over SD? Yeah, it looks better even at a realistic viewing distance (unless you have a very small TV). 4K over 1080? I don't know anyone that sits close enough to their TV to possibly notice the difference*. Furthermore, I have zero interest in buying it when there is no 4K cable and no OTA 4K content (I think DirecTV and Dish Network may have a channel or two), and (since there is no standardization), if and when 4K OTA came out I'd be stuck buying ANOTHER new TV anyway.

*I'm curious if people swearing up and down that 4K looks better are looking at the same content in both 4K and 1080 -- when HD came out, the tendency on in-store demos was to use over-sharpened video with the contrast set excessively high to make it look "vibrant", and some blurred out dim crap on the SD sets. I heard several people (watching an ad ON THEIR EXISTING SD TV showing an HD set with excessively sharpened forest scenery on screen) exclaim how much sharper that HD picture was, until I burst their bubble and pointed out they were watching the ad on their existing TV, so it couldn't be sharper.

What I'm seeing...

What I've been seeing is this tendency for some ISPs to have dropped peering agreements in favor of a smaller number of higher bandwidth links. Some traceroutes between ISPs here in eastern Iowa that used to route locally (so 5 or 10 mile round trip) and more recently would route through Des Moines (~200 mile round trip) now route through Chicago (~440 mile round trip).

I agree with Big Ed, in principal, the cable plans all have GB limits, and overage, the customer's paid for those GBs. and the ISP should deliver. The provider of whatever service buys enough bandwidth to provide their service.

So, the squabble you have now is some of these providers (TWC in this case) failing to maintain adequate bandwidth to these exchange points that Level 3 and Cogent (to name two) use. In Verizon's case of slowness with Level 3 (and so Netflix) (even after Netflix paid some fee to Verizon), Verizon's own diagram showed they have plenty of backhaul (I assume fiber) running from the exchange point to their backbone network, and plenty of backbone capacity, but a link at the Los Angeles exchange that runs at 100% utilization. They have an 8-port 10gbps switch with only 4 ports hooked up, they could double their capacity at this exchange point for the cost of a few patch cables and solve the problem, they just won't.

It's tricky, because TWC (and Verizon etc.) really aren't throttling anyone, so it's probably not subject to the open internet rules. But, I do think they are being a bad actor by collecting plenty of money from their paying customers to maintain adequate connectivity, then expecting others to pay for it. If the US internet market were in better shape, it wouldn't be a problem, if your ISP failed to maintain good enough connectivity you'd move to one that does, but there's many markets here with few choices.

"As others have said Herron is a director of the company that makes this GPS device. If you look back a few years his co-director Dr Philip Tann was caught speeding and he too had his case withdrawn by the CPS. Strangely enough both got caught on the same road in Sunderland, how strange is that?"

Not strange at all. Given the comments on this model of laser speed gun being known to give improper readings, they probably were both improperly photographed, and were just two of the few with evidence to refute these false speeding claims.

Dealership

It's certainly possible the Taliban just ripped these trucks off, as they too over an area they sprung all the trucks from the dealerships.

But, it sounds like they are not hurting for money -- besides whatever backers they have, they also have income from selling off those artifacts that they aren't simply destroying, and by taxing exports of drugs (i.e. opium).

My guess.. I don't think the Taliban marched into a Toyota dealership in full dress uniform and said "Hey you, get me 100 trucks". The dealer would wonder "WTF", and even if they didn't have a problem selling to the Taliban, they'd have to order more trucks through Toyota and the question of who they are going to would come up then. Rather, I assume people may have come in in plain clothes and bought a few at a time. I mean, if someone popped into a dealership here, said "Hey I want to buy a truck" and slapped down a briefcase of cash, I don't think there'd be a lot of questions asked, they want to complete the sale!

cirby, pretty sure these are two different incidents.. the US supported Taliban opponents in the 1990s. They supported the Taliban themselves in the 1980s. I've never taken this as a "hate the US' meme, but rather a "be careful before you poke around overseas" meme.

I'd see gains for EU

"We expect that a suspension of Safe Harbor will negatively impact Europe’s economy, hurt small and medium-sized enterprises, and the consumers who use their services, the most"

And I think it could positively impact Europe's economy, help small and medium-sized enterprises; and "consumer" is a macroeconomic term, businesses have customers.

I would expect multinational companies to place more data centers within the EU (helping Europe's economy.)

I would expect small and medium-sized (as well as large) IT businesses to see an increase in business as (if they provide hosting or "cloud service") people move their online services into the EU; and if they don't provide hosting or cloud, some short-term business as others consult with them about what they should do.

I could see a further gain in these businesses as others OUTSIDE the EU move their data to exclusively EU-based data centers (as opposed to one that has data centers in both US and EU, since the US may then pressure them to keep non-EU traffic in the US so it's slurpable.) Either for privacy, or just to flip the bird to 3 letter agencies.

I'd expect a minimal one-time impact as non-IT-related businesses may hire an IT provider to see if they must move any services. But in most cases I'd guess they won't have to do anything (if they are using a provider with multiple data centers.. i.e. GMail or AWS or whatever... they should at most be able to tell them "I'm in Europe, move my data if it's not already here.")

Are these even contradictory?

For sake of argument, assuming both reports are true... The Greenpeace report comments on high defect rates among animals in the exclusion zone. The later report comments on quantity of animals in the exclusion zone. I'm not sure these are even contradictory statements -- if the defects do not make the animals sterile, then having animals with mutations and defects does not preclude the population increasing over time.

Furthermore, if there are any mutations that would increase hardiness in the presence of radiation, they may pop up in this area due to the survival advantage those animals would get compared to the rest, evolution in action.

Time for the mutant superdeer!

"One Windows"

I'm neutral on this right now -- I don't use Windows on desktop, phone, or tablet. But...

First off, the Win8-era plan was pure madness, letting desktop, phone, and tablet groups work totally independently and having 3 mutually-incompatible sets of APIs? Ugh (especially daft that Win8 would schizophrenically bolt on an unrelated desktop interface that is not compatible with either of the other 2.)

That said.. I'm reasonably impressed that (in the length of time they've had) they've managed to get all 3 compatible enough that Visual Studio can emit something that runs on all 3. (I have the feeling they are probably just as incompatible as always and Visual Studio has some compatibility libs and contortions to make them compatible though.) This making them compatible really is the best way for Microsoft to have a reasonable chance of getting significant development for any of them though. I do still think calling this "One Windows" is sort of a fantasy, if the compatible level is only ~60%. Will it be enough? I have no idea.

It's fine...

I think everything worked out OK. She can ask others to tone it down, but can't expect that everyone else must accommodate her apparent squeamishness about swearing. They didn't. She went on to work on some project that matches her collaboration style better.

To those of you who seriously think everyone else should change -- that's a load of crap. Some people are uncomfortable with the swearing and blunt assessments. But, others are equally uncomfortable being expected to curb their tongue, be nice and understanding and respect others feelings at all times and... I can't even finish the sentence, it's so syrupy it's damn near making me sick.

Really, there's plenty of projects so both types can have something to work on without trying to force other's behavior.

Privacy issues

I'd say one other reason people are not trying out Win10 is privacy. My friends* who run Win7 are concerned about the amount of information being sent to Microsoft in Win10 (and, with a lack of information saying what exactly is phoned home, they tend to assume it's sending virtually everything.) The best thing Microsoft could do to assuage these fears is list exactly what types of info are sent, how to turn these off if they wish, and what is sent when "everything is off" (probably automatic update-related traffic.)

I think, realistically, Win10 probably doesn't collect an unusual amount of info... but the privacy policy saying what info is sent to Microsoft, well it looks pretty bad when it's like "searches, microphone input, your E-Mail, login, password, user directory contents, may all be sent to Microsoft." But I assume defaut web search is bing; Cortana uses the mic; Outlook Online (or whatever it's called) would have your E-Mail going through them; only active directory the login and password and possibly user directory contents; OneNote will store your files with them too, but only if you actually use them.

I've never had problems cancelling...

I've never had problems cancelling service. It's easy:

1) If you have any cable modems or boxes, you'll have to go in to return equipment anyway. So do it in person.

2) If it's on the phone, let them know you want to cancel. Don't let them keep making offers if you intend to cancel, point out you expect service to be cancelled now. If they persist on yammering on about whatever, point out you will be stopping payment on any attempts to withdraw money after that so they might as well cancel now. They should cancel. I've never had to go past this point.

3) Write down when you made this call. If they try to claim you owe more money later (beyond some fractional month's worth if you cancelled mid-month), point out the date you cancelled and stand firm, don't pay it. If they try to put a negative note on your credit, the credit agencies must correct inaccurate information if you request it, point out the date you cancelled and they are obligated to remove this. (If a company makes too many false filings to the credit agency, they get into some kind of trouble too.)

Collect != collect

“The Prism programme – which is another name for foreign intelligence collection subject to judicial supervision under section 702 of the Foreign Intelligence Surveillance Act – is NOT based on the indiscriminate collection of information in bulk, as a report from the US Privacy and Civil Liberties Oversight Board makes clear,” said Litt in a statement.

The important thing to realize here, the feds have already come up with doublespeak for situations like this.

a) "subject to judicial supervision" -- the FISA court initially just rubber stamped anything that came by their desk. As the NSA expanded what they slurped in, the FISA court did eventually state their reservations on the scope and scale of this program, and found some uses of it were flat-out illegal. They "supervise" insofar as they release legal judgements on the program, but there seems to be no penalty for FISA judging them to be illegal.

b) Collect doesn't mean collect. The NSA and federal gov't have intentionally redefined plain English to fit their purposes. They define information as "collected" or "intercepted" NOT when it's pulled off the wire and dumped into some database they can search at any time (i.e. when anyone who speaks English would say it's collected). They say information is not collected until someone at the agency has done a query that actually pulls up that information.

Wholesale rates?

I don't know for sure what the wholesale rates are... but, T-Mobile US has free, unlimited (but throttled to 128kbps) data roaming overseas (Canada and Mexico are not throttled). This would take 64 seconds to rack up 1MB of data, so I'm quite sure T-Mo is not paying what you are for roaming. They also have things like $50 for 500MB of full-speed data, which is $0.10 a MB -- so they're paying less than that.

Rolling releases and review

"still, with the versions? you better believe it cuz you can't go to a rolling release with stable tracks or anything, that's craziness!)"

I like version numbers. If I want to follow current, I just dist-upgrade as a new release comes out. If I want to follow stable, I install an LTS; if I want more updates I enable backports. There is the downside of the "big update" when you run a dist-upgrade or full-upgrade to go to the next LTS. But the upsides are 1) If you don't like something in the new LTS, you can actually go back, and if it proves buggy you can go back and wait it out. 2) 3rd party software can say "this requires version x.y of this distro" which you can't do if there are no versions.

Don't get me wrong, I see the appeal of a fully rolling release (I've used Gentoo after all.) I just also see the appeal of doing it the other way too.

"going to self destruct like all boontoos from the last several years do?"

I have one system that removes the old kernels, so I have the latest kernel and one older version. I don't know why one system does this and not the rest. I'm not running seperate /boot so I don't rapidly run out of space.

"Not even a good try: this wasn't a review of an OS, just a superficial review of a GUI, supplemented with a list of version number bumps."

I found it perfectly fine. He points out he found surprisingly few changes between the previous version and this one, other than the version bumps. He reviews the changes he DID find (which were in fact superficial GUI changes) and comments on the stability of the system (which, as he comments, can vary a lot on these October releases since they usually have massive changes being worked in.). I found this review perfectly fine; it made it clear to me that if I wanted a full-on review I could simply proceed to read the review for the previous version.

Friend's 386

My friend had a 386 than ran pretty well, but every so often would spontaneously reboot. You know where this is going -- after this last time it rebooted, he opened the case up, to find a few mice with one pissing on the motherboard. The system was rebooting each time the mice pissed on it. The largest hole in this case was a an open spot where a different motherboard would have had a 9-pin serial port (about the size of a VGA port.)

Good move

I don't have the concern "I wouldn't trust Experian to do the monitoring", they are one of the three big credit agencies (Experian, Equifax, and TransUnion). But it'd gall me if Experian lost my data, then got away basically penalty-free (either actually got paid to provide protection, or provide it "on the house".. since they are a credit agency this'd cost them almost nothing.) I'd prefer T-Mo get protection from someone else then collect from Experian for it.

US patent trolls

""filed in 1992, granted in 2001"

nine years! is that typical in the USA?"

Yup. Non patent trolls, you can get a fast track patent in a matter of weeks, and conventional months. But patent trolls use tricks to intentionally delay this process. Read on...

There's a few particular attributes of the US patent system that makes them take this long. Intended use was, back when people were working on whatever invention and were worried about someone filing days or weeks ahead of them. So, you can file a patent describing the basic device, and have some time to refile to add additional claims to the patent as you refine or add to your design. When you do this, though, it resets the time limit so you'd have a little more time to flesh out your design if you wanted.

The problem is, a patent troll can subvert this system, they will make frivilous "continuation applications" on the patent (legalese for a requests for more time). The classical "submarine patent", until 2000 one could keep getting their patent extended out, and it'd be valid 17 years from the date it's actually granted rather than the original filing date. This was put to a stop for patents filed after 2000, they're 20 years effective filing date. The new-style submarine patent, there is a tradeoff where the longer they submarine it, the less time they have left on the patent to collect on it. But, a patent troll can file a patent, add claims onto it for years, then when they pop this patent up and start swinging it around, defendants are expected to show prior art effective the original filing date, not the years later date the claims were actually added.

Huh....

I have to be honest, I assumed Adblock and Adblock Plus already had the same owner.

Apple's "solution"

I don't think there's much market for either one of these. But having a specific protocol that others can follow is really the only way to have any chance of this kind of thing ever catching on (for anything more elaborate than "overpriced radio-control on/off or dimmer switch"). I've got little interest in that kind of thing. But that interest drops right to "less than zero"* if it requires buying components piece-by-piece from a single vendor just because vendors can't or won't standardize.

But, I'd like to make it clear, Apple's "solution" of requiring a specific chip and firmware from Apple, in no way solves any of the problems of security, power use, etc., It solves the "problem" of Apple wanting their products to only interoperate with other Apple-approved products, bringing my interest to well below zero. Network security in no way requires a special security chip (AES accelerator? Sure you can have one but not required). And "security through obscurity" doesn't work.

*What is less than zero interest? I don't know, I suppose not only having no interest in the technology for myself, but telling others how dumb it is and why they should not buy it?

This really depends on what's happening

This really depends on what's happening.

If, as Samsung says, this dimming thing really does kick in under real-world usage and save power, then I think it's legitimate. Perhaps the testing should be done with options like this on then once with them off (making sure brightness and contrast are adjusted, so the vendor doesn't just use inappropriately dim defaults), so you get kind of a typical and (somewhat) worst case figure.

If there are sets that actually detect the IEC clip and start power saving right away, that's cheating and I assume they'll get fines and possible lawsuits. I do wonder how many would sue though, I couldn't get that worked up over a dollar or two a year on my power bill. But (in the US) if the extra power use pushes some monitor that claimed to be energy star into not meeting energy star standards then that company could have problems with the EPA.

Well, whatever...

I would think Amazon would want to encourage Prime subscriptions, by making them as easy to use as possible. Amazon should sell Fire sticks etc. on their own merits, I wouldn't buy into a service that actively cuts out 3rd party clients.

NasaTV's odd history

NasaTV has been around for quite a while. A kind of odd thing, it used to be that NASA was requiring NasaTV to be carried in the clear, apparently. They eventually dropped this requirement (although I think it's included in any satellite TV package.)

So, if you had a Dish Network receiver with no service, you'd have 1 channel with either Charlie Eigen (head of Dish Network) imploring you to sign up, or (if you were using a hacked card that Dish Network disabled) a channel where Charlie Eigen would chew you out for being such a cheapskate and gloat a bit about burning out your card. And NasaTV.

Monopolies and duopollies

Yup, too many places have uncompetitive pricing.

Locally, CenturyLink (DSL) and Mediacom (cable) are SO overpriced that satellite -- yes, satellite -- internet service is not only price-competitive but actually slight less expensive than either duopoly landline option (while providing higher upstream speeds than either one provides for any price; and higher downstream than either's lower-cost options.)

Note wireless broadband is not viable in most of the US. Some lucky areas have a (usually local) provider that just uses wireless to provide broadband, priced accordingly to compete with whatever broadband options are available. But the "big 4" charge insane prices, they are looking to sell service for phones and not that interested in actual broadband. You can't get much better than $50 for 2GB.... they'll charge you like $10/GB (minimum 2GB), charge $30+/month if it's a phone for voice & texting even if the line's for tethering and you don't want the voice and text. If you use a non-phone (mobile broadband card or wifi-sharing box) they want to double-dip by still charging $30+ a month to share the data you've ALREADY paid for. Oh and overage is cash overage, not throttled. (T-Mobile charges a bit better price and has throttle caps, but still not particularly competitive with a landline. Plus they don't have service in my market.)

Vista and SQL Server 2005? Whoop-dee-doo

Vista and SQL Server 2005? Whoop-dee-doo. Honestly, I wouldn't use either one. But (as bad as Vista is) once it's already on there and already paid for, if it runs the applications why would someone want to (at that point) waste money on something newer? This can be replaced at hardware refresh time. And for the huge cost of SQL Server, and the well-known issues with newer versions dropping support for some old stuff, why wouldn't someone just keep using 2005 if it works? They might want to look into if it's possible to upgrade their software in case they have to get a new server & can't get 2005 for it. But why upgrade just to upgrade?

I thought you were going to rag on people STILL using DOS-based software, old Win-3.1-era software, and so on. Yes, I've seen it -- ridiculously, at the insurance office I saw, each insurance co. (that they sold policies for) seemed to use it's own software, which was often some 20-30 year old thing they wrote once then apparently never updated. This of course means they are running all these 16-bit apps, so no 64-bit Windows. And now that Microsoft has gone to the ridiculous "Who needs version numbers, we'll just keep calling it Windows 10" plan (I hope to hell they change their mind on this!!), they may end up with a situation in the future where "Windows 10" runs 16-bit apps still but "Windows 10" doesn't.

Or the bowling alley that needed some spare computers that a) would fit into a limited space, so not a full tower b) had to have PCI *AND* ISA slots to run some unholy combination of cards to run their bollowing alley scoreboards and stuff. Yep. It was very hard to find any spares for them.

High school?

High school? Isn't that too late? I mean, it's called "high" school for a reason. I would think this would be plenty late enough, too, for people to be wondering "should I really be working for the feds?"* and wondering about rates of pay.

Anyway, what Ben Burch said, what they really need to do is quit worrying about recreational drug use. They should be concerned if someone's addicted to anything (including alcohol) (since it could end up affecting their on-the-job performance if they become more and more addicted), but the odd puff or drink off-the-clock should be none of their concern. I think a problem they run into here... their pay isn't just below market rate... the market that rate is compared to expects a 40-hour work week, while the FBI expects to control your behavior 24/7, that's 168 hours a week. To be told what to do "off the clock", people expect on-call pay or some kind of compensation.. so the FBI is actually paying severely below rate.

*To be honest, I think this is one problem they're probably having.. I haven't heard about the FBI running illegal and unconstitutional programs like some 3-letter agencies I could name. But I get the impression some of the general public just view the whole thing as "the feds", and assume they act much more as a monolithic entity than they really do (and so would assume if one agency misbehaves they all do.)

Good reason not to deal with Apple

This illustrates a good reason not to deal with Apple. I won't buy an Apple product (why would pay more for a product that does a subset of what I can get from another vendor for less money?) and why I don't plan to develop for them either.

Why Microsoft why?

Why oh why did Microsoft not make the default "always show file extensions" like 15 years ago -- to me, for security purposes, showing the FULL file name by default on a system like Windows (where system behavior varies based on file extension) is exactly as obvious as the decision to turn off autorun. But here we are, with systems still supressing important file information by default.

Was it really a test?

Was it really a test? I'm rather suspicious of the garbage URLs.

If it was:

NOTE TO MICROSOFT: Per RFC2606, .test and .example top-level domains are set aside for tests and examples. example.org, example.net, and example.com domains are also reserved for tests and examples. All these domains are held by IANA (Internet Assigned Numbers Authority) so they will never be assigned. You are not to use (random crap).org, .edu, or .gov, because they (theoretically) could be assigned at any time (and, in fact, if these links had been kept up, scammers could have registered at least the .org domain, while they could never have gotten an (whatever).example.org domain.)

Second note: Perhaps you should put "test update" somewhere in the update description, so if it's leaked it's not so mysterious.

Why do pepole reply with posts like this? Because.

"Why do people reply with posts like this? It's like reading an article about a cat problem, and posting just to tell people that you own a dog, and dogs don't get feline infections."

I don't reply with posts like that (usually). But I can see why people do -- too many people comment as though it's a natural state for computers to have to be on this vigilant lookout for viruses, and spyware, and updates from the vendor that do bad things, and buggy updates, and weird software conflicts, and on and on and on. These people like to point out that this is just Windows, not the natural state of al computers.

board behavior and emissions defeats

"But the point is that these cars, when tested under the European testing, will not trip the US test defeat conditions - I understand they're quite specific, because they don't need to. So when they are tested, they are almost certainly not in the reduced emission mode."

Some of these cars when tripping no test condition should (just) meet European emissions -- European diesel NOx limits are about 5x US limits. Some of these diesels that were running at 40x the limit were violating BOTH US and European limits. Pretty sure these detect BOTH European and US test procedures.

"If they then pass the EU tests, then you could not sue the manufacturer for non-compliance or being 'too dirty'."

I wouldn't jump to any conclusions. In the US, EPA rules prohibit defeating (i.e. disabling) an emissions control device. This is not just if the emissions limits are violated. A car that runs squeaky clean during EPA testing, and dirtier (but still emissions legal) during real-world conditions, is still breaking EPA rules, since (per EPA rules) emissions controls that could have the car running cleaner at all times are being defeated.

Verizon Wireless' got a redundant network

Verizon Wireless' got a redundant network. As they were installing 4G LTE and the required additional backhaul to each site, they also rolled out quite a bit of site-to-site microwave backhaul. In some cases, it's used to get backhaul to sites where they simply couldn't get decent backhaul to it otherwise. I remember reading a few years ago they got it set up to work as a fallback as well, with automated failover. Now when a fiber cut knocks out 1/4 of a state or whatever, the affected area can usually be fed via the microwave links from areas outside the affected area, keeping voice and (most likely slower than usual) data going.

Of course that's wireless, not DSL. Reading about Sky Broadband, it's not really clear if (to get out of the area) it'd run over some Sky-owned fiber or BT or what.

Fuel range

"Far fewer fill-ups (for those in locations where petrol stations are busy, inconvenient or just particularly noxious)."

Don't list this as a reason to get an electric vehicle, it's disingenuous at best. These electric autos that have like a 75-125 mile range, that's obviously not comparable to any gasoline (or diesel) powered vehicle. Tesla (with bigger battery pack)'s ~265 mile range? Pretty good but still lots and lots of conventional cars get better range than this.

The other points are true (depending on which electric vehicle you get.)

Why are these on the open internet?!?

Title says it -- why are these on the open internet?!?

Quite simply, specialized equipment (medical instruments, scientific instruments, "car computer" some auto shops have, to name 3...) should never be placed on the public internet. The OS itself will become increasingly out-of-date, and unlikely to have vendor patches for known vulnerabilities. And the application code, if it's fully custom it may or may not be following secure programming practices. If the application relies on some standardized libraries or web platforms or whatever, there could be more and more known exploits for these over time, which (again) may not ever be patched. You also don't have to worry about someone figuring out your admin password in bigguy 8-). I thought everyone knew this, I'm surprised to read about significant amount of hospital gear online.

I've heard of newer equipment using Linux instead. I'd expect the Linux install itself to be plenty secure but if the device has any web access, or administrative port, or whatever open, you are then still at the mercy of whatever application code the device uses, and if this is secure or not. Of course out of the box security won't help if your password's bigguy 8-) . So needless to say I still would not put it directly online.

Not even legal under US law

"Yes, the Americans are trying (as usual) to apply US law to other countries. But this seems to be deeper than even that."

Not even that, the NSA's argument in insisting it's actions are legal under US law (despite being explicitly illegal under the law), is to say the NSA's head lawyer determined it was legal, therefore it's legal. So really they're not even trying to apply US law overseas, just trying to get what they want.

One big problem

I thought I'd go take a look a this. One big problem -- Unity Web Player is only for Windows and Mac! This is the point where someone would make a snarky comment about Linux having "no" market share or something. 1) Simply not true. 2) This GCHQ-backed site is meant to attract people into infosec, I would expect the demographic drawn to this site to be significantly off the norm in terms of both OS and browser usage.

edit: Per Google there's kludgey way to make Unity Web Player work as a firefox plugin, that may or may not actually work. But still, I would hope something brand new would just use HTML5 or something.

So...

So, not only incompetent but evil. Defending the use of torture is evil. Diverting paying customers already-paid-for hardware to a third party (no matter who it is) is evil to a lesser extent. The actual diverting systems to the NSA -- well, who knows, Fiorina could have plausibly denied knowing the extent of the NSA's spying programs, and this may not have reflected on her too badly. Bragging about supplying hardware to what are now known to be illegal and unconstitutional spying programs? That's evil.

Also, the last thing the US needs is increased military spending.

QR Code? Why not I guess

QR Code? Why not I guess. I have no desire to scan my food and see where it came from, but QR codes are free to use, just printed on so cost is approximately zero. Compared to some few cents for an RFID tag, and most phones can't do anything with it. A little easier than going to a web site printed on there and typing in a production code.

Good on them

Good on them. You'd probably have distinct cpe (customer premises equipment) for EPON as rolled out by Verizon for FIOS, and EPON as rolled out by a cable co. then (unless the proper firmware can support both modes of operation.)

But, cable? Some of the cable cos are running their billing systems on a mainframe, it handles billing, sending "hits" to set top boxes (or disabling it as the case may be), provisioning or unprovisioning cable modems, and so on. Using DPoE sounds far easier than trying to come up with a totally new setup for EPON.

It'll be expensive...

I think this could be expensive.

The cars that were running at 5x the US legal limit may be within European limits; emissions limits up to Euro Tier 5 are almost 6x the US limit, and Tier 6 (introduced 2014) is still about 2.5x the US limit. VW's going to have problems with these cars in the US, but in Europe I doubt they need to adjust anything, and if they do it should be pretty mild, shaving a few .01 g/km off the emissions.

The ones that were putting out 30-40x US limit? Those are trouble. I'll just note that by 2010, VW had the only diesel road vehicles available in the US without urea injection. Navistar planned to continue using an EGR-only solution, but switched to urea injection almost literally at the last minute of 2009, after they found they had not found a engine management software and tune that'd meet emissions and maintain driveability, power, and fuel mileage. VW has perhaps two unpalletable choices:

1) Strictly software/tune update. This would need no new hardware (the EGR hardware's already there after all). But this likely would hamper driveability, power, and economy, and looks like lawsuit city as well as further harming their image (I mean, who wants a mandatory software update that does that?). They'll have to be very careful that a much more agressive EGR usage doesn't lead to any stalling, this would be an even bigger problem.

2) Retrofit urea injection. Hopefully there's a little empty space somewhere in the back to add the tank! This could be quite costly, a tank would have to be fit and something'd have to be added to the exhaust system for the urea injection. At $1,000 a car time 11 million cars it'd be like $11 billion. Ouch. But it shouldn't affect engine performance since it's exhaust aftertreatment.