Log in

No account? Create an account
Previous Entry Share Next Entry
This Month in Evil
I'm gradually catching up on my tech blogs from Pennsic. Remarkable the things one misses while away:

On the one hand, there's the Net Neutrality mess, as Google caves on a point they've been arguing as a matter of principle for ages, and the wireless providers using that as evidence that NN is a bad idea. They're saying it's a necessary compromise, not a business decision -- but really, this smells very much like inter-company horse-trading. It certainly fails the "appearance of impropriety" test to me.

Then, Google gets to be on the receiving side of (for my money) an even bigger evil: Oracle suing them for using Java in Android. The one silver lining here is that it might renew the well-deserved decline of Java. The language is old and creaky, and has long since been passed by better options, and now we're getting a great reminder that it is owned and patent-protected by a company that is happy to sue people using it. Time to move on to better things, and tell Oracle to f*** off.

On the good side, RIM (a company I usually pay little attention to) is coming down on the side of individual privacy, at least for their customers: they aren't giving backdoor access to countries that want to spy on their citizens. Surprisingly gutsy move, I have to say, and it's causing some middle-eastern countries to shut down Blackberry. But it makes a certain amount of business logic: corporate customers want that assurance of privacy, and it may be worth RIM losing some customers in more authoritarian states in order to reassure the ones elsewhere.

Still a week or two behind in my reading; it'll be interesting to see if there have been meaningful changes in any of these...

  • 1
(Deleted comment)
The silver lining on Net Neutrality: It seems to me that a disincentive to use wireless internet devices is called for. Our growing obsession with being online 24/7 is not healthy, in my humble opinion.

A comment and a question

I thought RIM already gave most NATO governments access to Blackberry communications for the purpose of the war on terror? I'm curious why they won't give India the same...

On net neutrality: It makes some sort of sense to me to prioritize small volume communications over large (so if the network has bandwidth = 100 units, and there are 50 people trying to send/receive 1 unit and 10 people trying to send/receive 100 units (each), then I would prioritize the 50 small and have the large packets take the remainder. But I'm not sure how feasible this is or what other implications this would have, do you? Also, are there any thoughts of doing something like the "smart" appliances (which will do a load of laundry (for example) when power is more available/cheaper) to allow users/machines to voluntarily deprioritize certain things (such as downloading software updates or TV guides/programs) on the networks?

Re: A comment and a question

I thought RIM already gave most NATO governments access to Blackberry communications for the purpose of the war on terror?

Do they? That would be disappointing, although I suppose not astonishing. The hard but inconsistent line does seem odd if that's true, although I suppose it could simply be based on how much their customers trust various governments.

It makes some sort of sense to me to prioritize small volume communications over large

I actually have no problem with prioritizing purely on the basis of size, or indeed almost any content-neutral mechanism. The real issue of NN is prioritization on the basis of *origin* (and more generally content, although discriminating based on origin is the likeliest abuse). Google's taken a appropriately solid stance on that historically (and they have good commercial reasons to do so), but it sounds like they're knuckling under now.

Also, are there any thoughts of doing something like the "smart" appliances

There are certainly some tools like that in place already. For example, it's not unusual to schedule big downloads like OS updates overnight. At the moment, though, I suspect end users don't do much of that because there is little economic incentive to do so, which can lead to a certain amount of tragedy of the commons. Such incentives could be created, but would be controversial.

I should note that I'm a relatively conservative economist when it comes to this sort of thing: in principle, I actually like the idea of metered pricing, where you are charged at least roughly based on how much traffic you generate. Properly implemented, that could provide this sort of incentive for responsible usage, as well as giving the ISPs better incentive to play fair with NN principles. But end users tend to get cranky about metered pricing, and it probably only works well within a strong regulatory framework (which the ISPs fight tooth and nail against, since it would reduce their ability to charge monopoly rents). So the odds of it being done right aren't good, and without those sort of price signals, end users don't have a lot of incentive to play nice.

Re: A comment and a question

Congestion pricing is what would (IMO) really do the job, rather than metered pricing per se. The catch there is that I can't see how to do the pricing signals....

(As for RIM and NATO, if they are doing it I suspect it's because they're based in a NATO country and the Canadian Forces have more guns than RIM does. :-)

Re: A comment and a question

Yes, true -- I'd sort of elided that step. I tend to think of congestion pricing as a specialization of metered pricing, but you're correct that it's the part that's relevant to the question at hand.

(In principle, the best way to do the pricing signals would seem to be auction-based, where the price-per-bit goes up as congestion does. In practice, though, that's a hard sell to end users...)

Re: A comment and a question

Another point to this may be that individual consumers as a group don't use enough bandwidth during the most congested hours to make it valuable to do that (as opposed to doing this type of pricing for companies/institutions). That's often the reason given for electricity smart metering not existing.

Re: A comment and a question

Actually, my impression is that individual consumers *do* use a lot of bandwidth, and that moreover it's wildly uneven. Basically, video and stuff like that consume the vast bulk of cable bandwidth, so a modest number of consumers are eating most of the bandwidth.

The analogy with electricity is imprecise in some important ways. In particular, at the local end you simply have corporate and consumer use on different systems, with totally different pricing structures. Indeed, as far as I can tell, certain consumers use a great deal *more* bandwidth than many companies, since one video stream is a lot more bits than many email accounts. So cable congestion is mainly about consumers vying for the bandwidth with each other. Whether this is actually a problem is an open question: the ISPs claim it is, but I gather that other studies imply that we're nowhere near filling the overbuilt pipes.

Where there clearly *is* a problem is wireless: with the rise of smartphones, we're overloading the poorly-designed infrastructure rather quickly. Hence, the wireless ISPs seem to be moving towards at least partially metered/tiered service despite consumer objection. I don't have much problem with that *provided* it's content-neutral. My gripe with the recent weakening in Google's stance is that they appear to be caving on precisely this point, which is a recipe for creating precisely the sort of walled gardens I don't want to see...

When you are caught up you will find that RIM caved to UAE

Ah. Pity, although not astonishing...

The Java situation is way less clear than that. Google apparently created their own VM and tools that will translate Java bytecode to their VM's bytecode, and carefully avoided calling it Java to avoid being bound by Sun's (this predates Oracle) licensing terms. I don't inherently have a problem with that, but they did build a system that was compatible enough for developers to use standard Java development tools and develop Java apps that would run on their for-legal-reasons-not-Java VM. I didn't care for that when Microsoft did it, and I'm not convinced of the purity of Google's motives in doing it.

Also, if it was worth it to Google to go to all that trouble to have their own version, I somehow doubt this is a sign that Java is on its way out.

HAH! I knew it. I knew it. I told my colleagues that Oracle was gonna start being obnoxious about Java, and they didn't believe me.

Hahahahahahahah. (I don't like Java very much, but I like Oracle even less. They've started charging for *security patches* for Solaris. All versions.)

Edited at 2010-08-26 12:36 am (UTC)

  • 1