Posts by Kanhef 615 publicly visible posts • joined 3 Nov 2007
There are 8,760 hours in a typical year, give or take a few. By some fairly basic calculations, the Azure uptimes should be 99.5098% and 99.8757%. I don't know how CloudHarmony came up with their numbers, but I wonder if Microsoft 'encouraged' them to use some alternative voodoo calculations so they can claim "99.9% uptime", when any service that is down for more than 8.76 hours clearly fails to meet that standard.
This will result in more neurotic parents freaking out over every slight variation in the kid's vital signs, because they don't have the medical knowledge necessary to understand what it means and what is or is not cause for concern. I wonder if some hospital executive is behind this, trying to see how many more urgent care visits they can get and bill for.
Sure; the Shellshock bug was introduced back in 1989. Once it was reported, though, there was a patch available in 12 days, from someone who maintains Bash for free as a side project. Microsoft pays thousands of people to work on Windows as their full-time job, so not being able to respond to vulnerability reports in a timely manner is embarrassing, to say the least.
The only way security-minded people would accept a version of Skype as 'compromise-free' is if it's completely open-source, and can be reliably compiled to be byte-for-byte identical to any distributed binaries. If we can't inspect the code and prove that there are no backdoors or weak, home-rolled crypto systems, it will still be considered compromised, no matter what anyone at Microsoft says.
If every node delays every packet by a random amount in the same range, all this will do is slow down the network. With enough packets to analyze, the randomness averages out and isn't a significant obstacle. A better approach might be to add delays depending on the speed of the individual connections between nodes; the idea is that all traffic takes the same amount of time to transit through a node, no matter where it came from or where it's going.
As Steve Davies mentioned, there will be security vulnerabilities found in the .NET libraries; it's a question of when, not if. The real problem is what to do about them, now that the libraries are bundled with apps.
On the one hand, they could let developers release new versions of their apps every time the libraries are updated. Realistically, most of them won't bother, which creates a large attack vector. I'm sure VXers will find a way to take advantage of it, such as convincing users to install vulnerable apps which can be exploited. (E.g., "you need XX video player to watch this clip of [celebrity]".) Will antivirus programs have to start flagging anything with outdated libraries as potentially harmful? This way lies madness.
The alternative is to push security patches through Windows Update. Except this is supposed to be cross-platform, so you don't necessarily have Windows Update. Maybe solvable with an updater service, but now that also has to be bundled with apps as well, and could lead to issues with multiple instances and version incompatibility if you install several .NET apps. Even with that solved, pushing updates could break signed apps.
Anyone have better ideas on how to not have this turn into a nightmare?
Why would any app need a list of all other (running) apps? Gathering that information and sending it off for 'analysis' definitely counts as spyware. If app functionality depends on the presence of certain other apps, the OS should provide a means to query whether those specific apps are installed, rather than revealing all of them.
I can understand the need for some of the permissions, such as access to the camera and storage so you can post photos. But I'd love to see their explanation for why it needs to be able to modify contact information, read text messages, change network connections, or modify battery information.
Unless Windows 9 (or whatever it's called) goes back to a familiar user interface, I think this will be an increasingly tempting option for enterprises. If they're going to have to retrain users anyway, why pay an arm and a leg for Windows + Office licenses, plus the inevitably required hardware upgrades? In most cases it would be cheaper to hire an on-site migration assistant from a distro provider than to stay with Microsoft.
Trust management problems in 73 percent of the top 1000 apps, but only 36% of the next 9,000 most popular apps. Webkit issues in 77% of the top 1000, but just 6% of the next 9,000. Why are the most-downloaded apps so much more prone to security problems than ones that aren't quite as popular?
Amazon and Ebay aren't actually that unreasonable – they're probably trying to look up your postal/zip code so they can automatically calculate shipping costs. Still, it would be nice to have the option to turn that off, in case you're shopping while not at the location you want things delivered to.
I'm not trying to defend the NSA here, but as far as targets go, this one isn't unreasonable. People who are looking into ways to hide their online communication are more likely than the average netizen to be doing something of interest. They might be terrorists or other spies, they might be Chinese pro-democracy activists or Iranian counter-revolutionaries; in any case, the NSA wants to know what they're up to. Of course, there are also plenty of people doing nothing of interest who happen to be conspiracy theorists or just don't like being spied upon, but I don't know of an easy way to tell the difference short of spying on them more.
I don't like the overbroad dragnet espionage, but at least there were some attempts to focus on valid targets. If it had emerged that they were scrutinizing visitors to dailykitten.com, that would raise serious questions about their competence.
That would only happen if it were claimed that Paypal is inherently a criminal enterprise, not just that the CEO has committed a crime. Pretty much everything being sold on Silk Road was illegal, which is what justified shutting it down and seizing its assets. Even if the CEO was using Paypal for money laundering or such, most traffic on that site is legitimate, so they'd only seize the related accounts, not everything.
It could have been done reasonably well. Throw up a notification asking if people are willing to be part of an experiment on social behavior, which may alter their experience of Facebook for the next week. Explain that providing any more details about the experiment would alter people's behavior and invalidate the results, but provide more information about the study and which group people were in after it's over. Not complete information, but enough for reasonably informed consent, and far better than how they provided no information and obtained no consent.
Claiming that there's no battery, but it "stores the energy in a uniquely designed power bank." Sounds a lot like a battery to me. Somehow it's able to contain all the circuitry, the power-harvesting antennas, the Bluetooth antenna, the not-a-battery™, the accelerometer, and possibly other components. It can somehow tell the difference between being shaken intentionally by a person and being shaken incidentally when attached to a dog's collar. They're able to connect a wired diagnostic interface to it, but it's sealed and waterproof.
Supposedly they have working prototypes, but there's no clear demonstration or explanation of how they are used. My guess is that it would use some sort of roundtrip timing signal, which would only give a distance to the tag, and no information on direction. If the phone can track its own location accurately enough, I suppose it could do some sort of automatic triangulation. However, AFAIK, GPS isn't accurate to less than about ten feet at best, which is too fuzzy to use for finding something within a house only a few tens of feet wide.
It doesn't seem like this would be particularly hard to do; you could probably borrow a lot of the code from web browsers, which already do a fairly good job of handling malformatted HTML. There are three main types of XML error that I can see:
Orphaned tags with no matching closing or opening tag, which is what Trevor's problem seems to have been. Easy enough to delete or escape as text.
Transposed tags, such as < a ...>< p >< /a >...< /p >. This would take a bit more work to detect, but the fix is obvious.
Broken tags, particularly missing right angle brackets. Escape the left bracket and recheck the document, as this will probably create an orphaned closing tag.
and not one prone to collisions. If they're using something weak like md5, there's a potential denial-of-service attack here: identify a (legitimate) file you want removed, upload a copyrighted image or video carefully padded to have the same hash, issue DMCA notice, and they'll block access to both files.
Maybe not quite as well, but you can hide *nix malware fairly well using similar techniques. Keep most of the payload, scratch files, etc. in an encrypted virtual file system; to anyone else it looks like a regular binary file. The only exposed part would just open the VFS and load the rest of the code; give this a name confusingly similar to a known daemon, and it could easily be overlooked. It may not be as easy as it is on Windows, but don't say it can't be done.
Good analogy; much like the California Gold Rush, the people who really make a profit aren't the miners themselves, but the people selling equipment to them. In this case, it's the companies making Bitcoin mining machines. Low-end models cost $2000 or so, top of the line ones are $10,000 to $20,000. Due to the rapid pace of development, they quickly become outdated and too slow to be competitive, so people need to keep buying new ones.
Obviously written by a lawyer: by saying the press report as a whole is inaccurate, they have carefully avoided having to either confirm or deny any of the specifics. They may or may not be having discussions with Lenovo about something. They might be considering a business alliance with a different company. They might be planning to sell the Vaio unit entirely. I suspect they are up to something, but they don't want anyone to know yet.
Degradation doesn't even have to be that obvious. If an ISP delayed every packet to or from Netflix by 1 ms, no one would be able to tell the difference. Add another millisecond of delay every week, and see how it takes for anyone to notice. Netflix videos would still work without any obvious glitches, but they'd be a bit slower and take longer to buffer than the ISP's service. If only streaming data is throttled, and not ICMP traffic (such as pings and traceroutes), it would be almost impossible to prove that anything unusual is happening.
This is essentially an off-the-shelf MIG welder with the gun attached to a 3D-printer style mount. Not a complicated idea, though getting it to work reliably is the tricky part. It's fed with standard spools of welding wire and inert gas; there's no way to throw random scrap metal into it, and you wouldn't want to anyway – the resulting objects would be little better than scrap themselves.
Depending on the wire alloy and shielding gas used, the printed parts could be reasonably strong. It'll never be as good as forged steel, but still more than adequate for most low- and medium-strain applications.
Microsoft could have gone to their community and said "Here are the problems with the current Masters certifications. Help us create a new set of certifications that will be more up-to-date and more accessible to people around the world." Once the replacements are more-or-less ready, they could have announced that the Masters certs were being phased out and replaced with Xyzzy certs. No outrage, users feel Microsoft cares about and listens to them, everyone's happy about better and cheaper certs.
Instead, they started by canceling the Masters certs with little explanation and no replacement ready, and now are scrambling to do damage control and try to rebuild bridges. From the phrasing of the announcement, I doubt they had any plans for a replacement certification before seeing what a negative response it got.
Garcia's work is about gaining access to a vehicle you otherwise can't get into, which usually means breaking into someone else's car. Miller and Valasek's work requires that you are already able to get into and start the vehicle; their paper doesn't tell you how to steal a car by itself, but Garcia's potentially does. That is why they're being treated differently, not the U.S. vs. U.K. legal jurisdiction.
will use any excuse to bury their heads in the sand.
Nothing here is at all new, or surprising, or in any way invalidates climate science. It's well-known that weather modeling is chaotic; small changes to input data result in disproportionately large variations in output. In this case, the output isn't all that different; there's a discrepancy between the test machines, but the overall result of the simulations are similar.
It's also well-known that floating-point calculations can produce different results on different processors. Chips are often designed to perform these calculations with more bits of precision than the output register can hold in order to produce a more accurate result. This is normally a good thing, but can be a problem when exact reproduceablitity between platforms is needed. Programmers have been dealing with this for many years; for example, back in 2000, Java added the StrictMath functions, which have consistent (but slower) results across all platforms.
Once attackers have been identified, redirect all traffic from them to a second server, full of good-looking but fake data. The intention is to make them waste time attacking the fake server, and even let them think they've succeeded, while preventing them from accessing the real database.
Should be easy enough to test - put a piece of pure nickel in, let it run for a while, then take it to an independent lab and have them do an elemental analysis of it. No concerns about revealing "trade secrets" there.
Of course, they'll never agree to it, as it would immediately show their claim of achieving fusion to be false. There are too many hallmarks of bad science to consider this credible. The only question I have is whether they've managed to fool themselves and actually believe what they're saying, or if they're just trying to fool everyone else.
It appears they asked the question, "which of these eight issues are you most concerned about?", as if people are only capable of caring about one of them, and assumed they don't give a damn about the other seven. A properly done study would have allowed people to indicate how concerned they are (from 'not at all' to 'extremely') about each of those issues.
Also, "seventeen years of continuous surveys" is flat-out wrong. According to the linked report, the survey was conducted exactly three times, in 1993, 2000, and 2010.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
