Follow Slashdot stories on Twitter


Forgot your password?

Comment: Re:of course it wasn't NK (Score 1) 236

Everything about the attack has seemed to be inconsistent with North Korea's tendency towards propaganda.

It just seems... odd... that the attackers behaved consistently like disgruntled employees/ex-employees.

Then Sony started talking about North Korea for whatever reason, and I think the attackers saw that and ran with it, thinking it was a great way to send Sony on a wilde goose chase. Heck, they might have intentionally left evidence pointing towards North Korea from the beginning (I suspect the various tools that the FBI thinks imply NK have already been traded around via underground methods and are in the possession of people other than their original creators/users...). Once there was public talk of NK, I think the attackers just decided it would be effective to screw with Sony regarding "The Interview". It's probably nothing to do with any moral objections to the movie - but it's a great way to cause a major financial loss for Sony and make them think someone else is responsible.

Comment: Re:Weird article (Score 1) 176

by Andy Dodd (#48618925) Attached to: Army To Launch Spy Blimp Over Maryland

Also note, one of the reasons the project got canned was supposedly its inadequacy at friend vs. foe identification.

Privacy advocates are freaking out about a system that apparently can't even reliably tell the difference between "friendly" and "hostile" let alone "that vehicle belongs to John Doe! FOLLOW IT!!!!"

Also, tracking ground targets over terrain (land) is likely feasible at FAR shorter ranges than the 340 miles given in the article.

Comment: Interesting (Score 1) 140

by Andy Dodd (#48566685) Attached to: $35 Quad-core Hacker SBC Offers Raspberry Pi-like Size and I/O

Hardkernel used to be one of the #1 purveyors of Samsung Exynos development boards (The other being Insignal). Unfortunately, both Insignal and Hardkernel's BSPs for Exynos boards tended to be vastly outdated. (Hardkernel was even violating the GPL with some of their Android 4.2 releases for some of the Exynos 4412 boards for a while - putting up binary images with no source code in sight.)

Now even Hardkernel is putting effort into non-Haxxinos boards...

Comment: Re:Had access to Glass for an afternoon... (Score 1) 73

by Andy Dodd (#48502003) Attached to: Intel Processor Could Be In Next-Gen Google Glass

That doesn't help either, but even for non-native stuff, the Haswells are about on par with Cortex-A15s (if Chromebooks are any indication) - and the Cortex-A15s are pretty much non-starters in the "small device" category, which is why Qualcomm's Krait architecture has become so dominant even in devices that are wifi-only. (See Nexus 7 2013 version)

Dual Cortex-A7s seem to be the "go-to" for wearables nowadays, providing similar performance (at greatly reduced power consumption) compared to the dual Cortex-A9s of the OMAP4. There are very few dual-A7 solutions out there, so the go-to seems to be to disable two of the four cores of a Snapdragon 400 (quad-A7). Nearly every Android Wear device except the Moto 360 is using a "crippled" Snap400, and the Moto360 keeps getting slammed for battery life due to being an older OMAP3. (One of its updates greatly improved that, but the "crippled" S400s still win as even on the same manufacturing process, Cortex-A7 is much more efficient than A9, and the A9 is more efficient than the A8 in an OMAP3. Add to that the S400s being on a 28nm manufacturing process instead of, if I recall correctly, 45nm for OMAP3/OMAP4.)

Comment: Re:Had access to Glass for an afternoon... (Score 3, Informative) 73

by Andy Dodd (#48498761) Attached to: Intel Processor Could Be In Next-Gen Google Glass

The raw performance of the OMAP4 wasn't the issue, BUT the fact that it's an EOL architecture no longer supported by TI is showing in the current software quality of Glass. Ever since Google deployed KitKat to Glass (which has not been deployed in production to ANY other OMAP4 device), Glass has been unreliable and suffered from wildly inconsistent battery life. XE19.1 was a big improvement, but it was still a significant backwards step from pre-KitKat Glass. Then Google went and fucked it up again with XE21 - Twice in one week I had Glass run out of battery in only 8 hours with effectively zero usage other than sitting on my head idle. (1-2 notifications/hour, no Navigation, etc.)

Even before KitKat, the OMAP4 was a woefully inefficient CPU due to its age. A Snapdragon 400 with half the cores disabled would provide a MUCH better experience - more efficient/capable GPU, more efficient video encoding/decoding engine (no burning your head when recording), more efficient CPU.

I haven't worn Glass in nearly two months now. It's in desperate need of a hardware refresh to improve power management and stability, but Intel is the LAST thing Glass needs. Intel's mobile SoCs are worse than even Cortex-A15 in terms of power efficiency, which is why you see a number of Intel-based tablets and settop boxes, but next to no Intel-based phones (there are about as many Intel-based phones as Exynos5-based phones, another SoC that's woefully unsuitable to phones due to power consumption.)

Comment: Re:LMFTFY (Score 1) 652

by Andy Dodd (#48459497) Attached to: Two Google Engineers Say Renewables Can't Cure Climate Change

This has always been my opinion. We NEED another generation of modernized nuke plants to bridge us until renewables are more mature.

Trying to mass-deploy renewables now WILL fail. We simply don't have the energy storage technology to do it.

One more generation of nuclear will bridge the gap. And ideally, during that time, in addition to renewables, work will be done on next-generation nuclear plants that can use the current generation's waste as fuel.

If I recall correctly, the IFR reactor design in the 1990s had the potential, if it had continued development, to be able to supply 100% of the US energy demand for 100 years, using only the existing reactor waste at that time as fuel. The resultant waste from an IFR fuel cycle would only remain hazardous for 500 years (as the longest-lived waste products from LWRs can actually be used as IFR fuel)

Comment: Re:It boils down to energy storage costs (Score 1) 652

by Andy Dodd (#48459333) Attached to: Two Google Engineers Say Renewables Can't Cure Climate Change

Not entirely true, if anything, it's MORE dangerous if you're changing power levels to match load.

There's a reason France (along with nuclear-powered ships) are the only ones that do such a thing. (In both cases because they have to - those communities have gotten VERY good at doing so, but it's still NOT an optimal way to run a nuke plant and does introduce new ways for the plant to have an accident.)

Nuclear reactors have properties that cause delayed reactions to control inputs, if you don't handle these properly, Bad Things happen. (And in fact, such Bad Things DID happen in an extreme case at Chernobyl. They tried to restart a xenon-poisoned reactor too quickly, and when the xenon finally burned off, there was a massive power transient.)

Comment: Re:So, why the continued G-love? (Score 2) 105

by Andy Dodd (#48388065) Attached to: Google Wallet API For Digital Goods Will Be Retired On March 2, 2015

Pretty much everything on your list of stuff that got dropped was in the category of:
1) Very few people used it in the first place (Wave, Buzz, Orkut, Reader)
2) Was not really dropped but replaced with a similar service under a different name with a method of migration (Picasa got integrated into Drive for all practical purposes, and actually technically G+ replaced Buzz)

The only thing I'm not sure about is Google Health - although remnants of that have been getting integrated into Android lately.

Comment: Re:About time for a Free baseband processor (Score 1) 202

by Andy Dodd (#48385481) Attached to: Department of Justice Harvests Cell Phone Data Using Planes

"But at least the phone companies can know about it and mount a legal fight, if they so choose'

Really hard with current legislation.

Remember Lavabit? It's already been proven that the government has been using legal means to acquire the private keys of service providers for the purposes of MITM attacks just like this one.

Comment: This doesn't work (Score 1) 488

by Andy Dodd (#48368185) Attached to: Denmark Faces a Tricky Transition To 100 Percent Renewable Energy

Last I checked, Denmark is already filling in their holes by using Norwegian hydro.

Effectively, when Denmark has a surplus, they sell it to Norway and Norway throttles back generation at their dams (allowing water to back up in the reservoirs a bit).

Of course, because this is a surplus and Norway doesn't really need it (other than effectively saving it in their reservoirs), it's sold to Norway cheap.

When Denmark has a shortfall, they buy back energy from Norway, but because demand is high, the price is high.

At least as of 2-3 years ago, no country had achieved more than 20-25% grid penetration of wind/solar. Denmark was the highest percentage, and to achieve their high percentage, they were critically dependent on Norwegian hydro resources as a form of "battery". There's only so far you can take hydro-based storage though.

That's the big problem with renewables - we just don't have the energy storage technology yet to make them feasible.

Some people claim that the UNIX learning curve is steep, but at least you only have to climb it once.