You are correct that there is no good way to store grid-level amounts of power. The best we have is pumped hydro, but we will never build any more of that. (Environmentalists are working hard to try to tear down existing dams, so good luck building new ones to make pumped hydro storage. And the best sites have already been built anyway.)

I do have hopes that the Ambri liquid-metal batteries will work as promised. I'm not an expert on this stuff, but the technology does seem to make sense as far as I can tell.

http://www.greentechmedia.com/articles/read/Slideshow-Update-on-Ambris-Liquid-Metal-Grid-Scale-Battery

Practical grid-scale energy storage would really help solar and other renewable energy sources become practical and dependable.

The point is valid, but not that helpful. Yes, our current system for finding useful apps is imperfect.

Sometimes when you invent a better mousetrap, the world doesn't figure it out and beat a path to your door. It would be great if the best ideas always win in the marketplace of ideas, but sometimes they don't.

And, if you can solve this general problem, you will be very popular.

I think social media can help a bit, but it's no panacea. (TFA noted that the voting for apps doesn't favor the best apps, and the voting system is arguably a form of social media.) Sometimes I hear about cool stuff on Facebook or whatever, but marketers try to spam us even on Facebook and its signal-to-noise ratio is degrading.

P.S. I live near Seattle. I went to Emerald City ComiCon again this year, and I arrived too late to get one of the spaces in the convention center parking. I wound up getting a good space using the BestParking app. So, it worked well for me the one time I tried it.

I found this app by doing Google searches for parking the night before, and finding the bestparking.com web site, which advertised that they had a mobile app.

Consider the history of Microsoft. In the past, Microsoft didn't expend any significant money or effort on lobbying in Washington, D.C. Then during President Clinton's time in office, Microsoft faced serious threats from the Federal government... the worst being that a Federal judge actually ordered that Microsoft be split up. This order was voided by a higher court, so it didn't happen... but you had better believe that Microsoft took it as a hard lesson.

Microsoft now spends a great deal of money and effort on lobbying in D.C. I don't blame them for self-defense via lobbying. (I do blame them for attacking other companies via lobbying, if they do. See below for allegations that they do.)

Google isn't waiting for D.C. to turn on them; they are lobbying to "manage their relationship" with the Federal government. So is Facebook.

http://www.politico.com/news/stories/0411/52483.html

Here's an article from 2008 about Google learning the importance of lobbying. It includes allegations that Microsoft was using its lobbying infrastructure to try to prevent a deal Google was trying to make with Yahoo.

http://dealbook.nytimes.com/2008/10/20/google-learns-lessons-in-the-ways-of-washington/?_php=true&_type=blogs&_r=0

Now I'm picturing Google using its leverage to attack Microsoft, and Eric Schmidt saying "The circle is now complete. In 2008, Google was just a student... now I am the master."

Other important questions: how many of those 7.1 million have actually paid for the policies, and how many just went through the web site? Also, how many of these policies are insuring the previously uninsured, and how many are insuring people who lost their previous insurance due to the ACA?

I don't have those numbers. Nobody seems to have those numbers... Kathleen Sebelius has said "we don't know that" (see YouTube link below).

I have a suspicion that if the numbers were good, somehow they would have the numbers.

The DailyMail article says that a RAND Corporation study estimates that the number of previously uninsured people who have actually paid for their policies is: 858,000 (well under a million!). I haven't found a source for this. I believe they computed this number themselves, by reading the RAND report and by using the percentages in that report.

Avik Roy read the same report, and reports the number as 1.4 million +/- 0.7 million, i.e. 700,000 people to 2.1 million people, 95% confidence.

I believe this is the RAND Corporation study being discussed: http://www.rand.org/content/dam/rand/pubs/research_reports/RR600/RR656/RAND_RR656.pdf

References:

http://www.dailymail.co.uk/news/article-2594309/President-plans-victory-lap-strong-Obamacare-enrollment-Sebelius-faces-unpopular-law-blank-stare-tough-questions-remain-whos-signing-up.html

http://money.cnn.com/2014/01/30/news/economy/obamacare-premiums/

https://www.youtube.com/watch?v=dXDdmRaJy2c

http://www.forbes.com/sites/theapothecary/2014/04/09/rand-comes-clean-obamacares-exchanges-enrolled-only-1-4-million-previously-uninsured-individuals/

I'm wondering if the $25,000 round is inert, or if it includes guidance.

The Navy has been talking about railgun projectiles with GPS guidance. All it would take is movable steering fins and a computer to drive them.

http://www.wired.com/2012/08/guided-supersonic-bullets/

No matter how good your targeting computers are, I think you need active guidance any time you are talking about a 100-mile range.

If I have done my math correctly, it will take about 67 seconds for a Mach 7 projectile to travel 100 miles (and that's assuming constant speed, not accounting for drag). That's a long time for a free-flying projectile to be subject to random winds.

Of course I'm not a physicist or ballistics expert. If I have made a mistake here, please let me know.

that "warmer" sound is called "low pass filtered".

Well, partially I guess. But there is also distortion, and some people like the distortion.

I used to work for JJ Johnston, and one time I helped him create an "ideal LP simulator". This was a digital filter that simulated the distortion of an ideal LP being played by an ideal needle. The filter converted the stereo into M+S (mono plus difference, the way an LP encodes it), upsampled by 4, and then applied different distortion functions to M and S (to simulate the different motions of the needle) before downsampling by 4. The M signal distortion was symmetric and the S distortion was asymmetric (as gravity affects the up/down motion of the needle in an asymmetric way). JJ got the distortion functions from a paper published in the AES journal. The upsampling/downsampling was to wipe out unpleasant harmonics that could be introduced by the distortion.

The effect was very subtle, so it was easy to miss, but overall I liked it. Someday I want to re-create that effect and make it work as a plugin on a Linux media player, and I'll use it when listening to music (at least sometimes).

We also made a couple of ideal tube amplifier simulators. Good times.

these implementations pay a penalty for relying on the VM's tracing garbage collection, and this penalty is loss of finalizer functionality.

And this is why it is best practice in Python to use the with statement, to make sure that things get cleaned up when you are done with them.

In CPython you can get away with just dropping your objects on the floor, and the reference counting will clean them up for you. In other implementations, not so much.

This works okay in CPython:

def read_data(fname):
    f = open(fname)
    return f.read()

In CPython, with no references to f it gets cleaned up, and the open file gets closed.

In any version of Python, this works great:

def read_data(fname):
    with open(fname) as f:
        return f.read()

The with statement ensures that the open file will be closed, no matter how the function is exited (including by exception).

And I actually prefer the with statement these days... I like how it makes the lifetime explicit for the stuff it wraps.

Raymond Hettinger has expressed the opinion that the with statement is one of the really best ideas in Python. I think that was during his PyCon session where he listed his favorite things in Python.

You actually have two complaints, not really related.

One complaint is the GIL. As I understand it, the GIL is mainly needed because the CPython reference counting model needs to touch a whole bunch of reference counts when churning objects, and this needs to not screw up. Thus Jython and IronPython, leveraging the garbage collection of their respective underlying VM platforms, didn't need a GIL.

The other complaint is that the language is too dynamic. That's just part of the design of the language. I can somewhat sympathize; I don't overuse the dynamic nature of the language (I know I can rebind "list" to do my own thing, but I never ever do that), and since I don't do that I would rather not pay the runtime penalty for it.

PyPy's JIT works by watching what the code is actually doing, and when it sees that (for example) your code never rebinds "list" it can strip out the dynamic lookup of "list" and just hard-code in a reference to the builtin "list" class. It can gain an extreme speedup when it detects a simple for loop, because it can replace all the Python VM machinery (creating new int objects, then freeing them) with a simple loop.

And, on the other hand, I do use the dynamic "duck typing" features to write less code, and I don't want to lose that. I love writing functional Python code where you can pass a function object in to another function and get a lot done with a relatively small amount of code.

It's possible that someday, some other language with a different static/dynamic tradeoff might eclipse Python; but I'm still using it because I get work done in it, and it is fast enough for the things I am doing.

You make it sound as if it were no big deal to remove the GIL. It has been tried, and Python got 2x slower, so that attempt was abandoned. Python 3.2 gained a different implementation of the GIL, and that fixed some problems, but other problems still occur.

The GIL is Python's hardest problem.

https://www.jeffknupp.com/blog/2012/03/31/pythons-hardest-problem/

https://www.jeffknupp.com/blog/2013/06/30/pythons-hardest-problem-revisited/

As noted in the above referenced blog, you can use Jython or IronPython to avoid the GIL; PyPy will be using Software Transactional Memory to avoid the GIL; and you can use the multiprocessing module to use multiple cores without GIL problems. You do have options other than just using CPython.

If removing the GIL was as easy as you seem to think, it would be gone now, at least in a fork of CPython. Yet still it remains.

no....
If OLPC wants a new mission, it should be to develop FREE educational software that runs on standard Android tablets.

Sure. They give away all the software they write already, and I assumed that they would give away any Android software they write. It seemed so obvious that I didn't feel the need to put in the word "free" but I guess I should have. Thanks for the comment.

one of the largest problems is that there are not a lot of books printed in the languages of 3rd world countries. and absolutely no advanced education texts.

I agree. IMHO that is the "killer app" for the OLPC devices, or for Android tablets. It's great that they can run software, but it's essential that they serve as textbook reading devices.

I think the ultimate educational tool for developing countries would be a ruggedized Android tablet with an e-ink screen. Color is a nice-to-have rather than essential, and the dramatic increase in battery life would be the win.

This suggests another idea for OLPC's future: they could make a deal with Amazon and ship a customized Kindle e-ink reader. Or maybe make a deal to ship a customized Nook. But either way, ride the coattails of a company that already spent the R&D and focus just on adding educational software.

I disagree with your analogy in which you compare stealing a car with patent grabbing.

Um, no. I said that stealing a car with the keys inside is still stealing; despite the fact that the car owner "enabled" the theft by being careless with the keys.

In the same way, being a jerk with patents is being a jerk, despite the fact that the USPTO "enables" patent jerks by granting patents that should not be granted.

Even if Apple were to have freed every one of their smartphone inventions, there would still be lawyers arguing that those inventions are not comprehensive, that their client has patents that fall between the gaps.

Your language loads the question here. You presuppose that Apple "invented" everything that they patented; I think some of it has prior art, some of it is obvious (like the patent on dialing a number by touching the number on the touchscreen), and some of it is probably genuinely new and interesting but by now I'm predisposed to assume Apple is a jerk.

And as for your main point, you are simply mistaken. If Apple had not patented anything, but just made a public demo of the iPhone, nobody would be able to file a patent after that and use the patent to hammer Apple. Apple's iPhone would now be prior art. If you disagree, please post some sort of reference to any example that would disprove this idea.

I work at an IP company, and it has been hammered into me not to disclose anything that might be patentable, because once the idea has been publicly disclosed, it becomes not patentable. I'm not a lawyer of course.

the analogy here is a defense system vs having weapons of your own.

I understand the possible use of patents as a defense. For example, Google: I am not aware of any case where Google tried to shut down a competitor using patents. This despite the fact that they own many patents.

However, Apple has been trying to use patents to prevent Samsung from selling phones. I have a problem with that.

you can't blame one company for doing it in a world in which everyone has the bomb.

Um, the whole point of my comment is that I do blame one company for doing this. I blame Apple, I blame Microsoft, I blame any company that is getting egregiously bad patents and enforcing them.

I agree.

If OLPC wants a new mission, it should be to develop educational software that runs on standard Android tablets.

You can buy "white box" Android tablets at amazingly cheap prices because they are mass-produced in China. While these tablets fall short of the ideal devices imagined by OLPC, there is absolutely no way for OLPC to get their costs down to match.

You can buy at least three Android tablets for the cost of one OLPC device. You could bundle tablets with a keyboard, a carrying case, and maybe a solar panel, and still massively undercut the OLPC's custom hardware.

Cheap Android tablet's don't have great battery life. But I bought one of the original XO-1 laptops and it only had a few hours of battery life, so clearly OLPC must consider even the limited battery life of a cheap tablet to be sufficient.

One of the nifty things about the OLPC custom design is that it's easy to repair. But with the massive cost advantage of a generic Android tablet, whole spare tablets could be shipped.

The promise of Sugar never was realized. For example, one of the reasons I bought an XO-1 laptop was that I was excited by the thought of the "show source" key, where you were supposed to be able to go anywhere in the system, hit the "show source" key, and find some kind of editable Python source code you could tweak. I never did find any source to tweak before I gave away my laptop. (It's in India now!)

Another part of the OLPC custom hardware was the "mesh" networking, which aimed to make it possible for multiple students to cooperatively share limited networking resources. Did that ever actually get used? All the photos I have seen show students in classrooms, and if the classroom has WiFi then an Android tablet would work fine. If the "mesh" networking is valuable, then maybe OLPC should invest in a one-off gadget that just does that, and plugs into the USB port on an Android tablet.

Apple is not the problem. The patent system is.

Can't we agree that both are?

If you leave your car with the keys in the ignition, then it is partially your fault when someone climbs in and drives away. However, the person who stole the car is also to blame. It's not a valid defense to say "He left his keys in the car so it wasn't stealing."

If all Apple wanted was to make sure nobody else got patents on all this UI stuff, they simply could have fully published the details of how their phone worked, and nobody filing after that would be able to claim to have invented it. And I'm not a lawyer but I think Steve Jobs's public "one more thing" demos would have sufficed to make all those UI features unpatentable by anyone else.

But that wasn't enough for Apple. "Patented!" crowed Steve Jobs. Apple patented everything they thought they could get away with, including totally obvious stuff like squishing your fingers together to make things get smaller on the screen, and spreading your fingers wide to make things get bigger on the screen. Come on, that is totally obvious and there even was prior art on it. So we return to where we started: the USPTO is a problem because it let Apple patent obvious stuff, but Apple is part of the problem for trying to patent obvious stuff. (Fortunately the "pinch-to-zoom" patent was in fact invalidated, due to Samsung winning in court against Apple!)

Samsung is going to go scorched earth on this new lawsuit. Millions for defense and not one cent for tribute. And Samsung has the millions. I hope Samsung wins big and invalidates all of Apple's patents.

(And then, as long as I'm dreaming, Samsung can go invalidate Microsoft's mobile patents next.)

They have a website devoted to this:

http://www.foldscope.com/

And the news on the web site is that they will give away 10,000 of these to people who volunteer to test them. If you think you could do a good job of testing, maybe you should sign up.

http://www.foldscope.com/#/10ksignup/

To me, the most impressive part is that he claims they have very accurate focusing. I believe he said "micron" focusing. I'm not sure how that works, but the paper is cut to a very accurate shape (the video showed some sort of computer-controlled cutter, it might even have been a laser cutter). By moving a tab I guess the paper can be made to flex predictably to focus the lens?

The biggest part of this story is that it is now easier to make a trojanized version of a legit app. But it has been possible from day one.

Android apps are written in Java, and Java bytecodes can be decompiled into something remarkably similar to the original source code. Then the source code can be edited and complied back to an app. Hey presto, you have a hacked up version of the app.

http://stackoverflow.com/questions/12370326/decompile-an-apk-modify-it-and-then-recompile-it

But -- and this is important -- the person using this attack has no way to sign the malware with the same signing key as the upstream source of the original, legit app. This means that it is much harder to trick someone into running the malware.

So, if you get an app from the Google Play store, and later someone tries to overwrite your app with a new build that is malware-infected, Android will refuse to install the new app, because the signing key isn't identical.

http://developer.android.com/tools/publishing/app-signing.html

So, if a user gets an email with an attached "free" version of an app that normally costs money, and that user has not previously installed the legit version of the app, and that user sideloads the malware version, then that user will have malware on his/her Android device.

So, as usual, it's easy to protect yourself: get apps from the Google Play store, and don't sideload apps unless you are certain they are clean.

For that matter, if you are browsing the Google Play store and you see an app that has only been up for a day, and claims to be a miraculously free version of a payware app... just say no.