Oh, wait.

Do the copyright mafias really have no shame?

Diigo looks like the best replacement I've found. The interface is slightly cluttered but otherwise it seems to support the features I liked about del.icio.us. And the ability to auto-add new bookmarks to del.icio.us simultaneously is really handy.

Good find.

... Delicious and Flickr. They just killed Delicious, and I'm hoping Flickr isn't so far behind.

I used to use Yahoo Mail, which was a great webmail service for its time... in 2000. I also used Yahoo Auctions until that folded. Before Google, I relied on the human-assisted Yahoo Directory for my web searches. I liked Yahoo Games, when they didn't have much besides pool and scrabble and word games.

But all of Yahoo's services have turned into ad-laden, bloated interfaces with out-of-date technology. It seems that the company has been unable/unwilling to innovate and has just been milking their previously respected brand for ad revenue. Flickr and Delicious were the only two services that seemed to resist this trend :-/.

I guess it's time to export my Delicious bookmarks and find an alternative host for them :(. SimPy and Del.irio.us used to be a couple of pretty nice open-source clones, but seem to have disappeared. Anybody else have a recommendation for a site with similar functionality, clean interface, and good browser addon support?

Opponents of software patents should root for Microsoft here, regardless of how you feel about the company (I loathe their philosophy but like a few of their products).

Believing in justice means believing it applies even to your enemies and opponents. Besides, we don't want the Supreme Court setting some awful pro-software-patent precedent that will haunt less-deep-pocketed open-source developers down the road.

I recently came across Rick Cattell's site which addresses just the questions you're asking.

Rick Cattell has written an excellent comparison guide of horizontally scalable datastores of different types (RDBMS as well as a variety of NoSQL systems).

Cattell has also written an academic paper with database expert Mike Stonebraker, which weighs the system design factors required to make a datastore scalable.

Executive summary of Cattell's work: although NoSQL may be a huge fad, the things that make a datastore scalable can be implemented in SQL RDBMS systems as well. Also, implementing do-it-yourself ACID in NoSQL systems is extremely difficult and error-prone, and is a significant advantage of most RDBMS systems. Stonebraker is the author of VoltDB, which is an open-source RDBMS designed for horizontal scalability, but they give a very fair and thorough look at competing datastores as well.

Wellll... sorta. Significant digits are not magic. They express the relative uncertainty in a quantity.

The relative uncertainty does not propagate equally under all types of arithmetic operations (see http://en.wikipedia.org/wiki/Significance_arithmetic for addition and multiplication).

This is especially true for exponentiation. If we start with the premise that AT&T's data usage grew 30x over 3 years, to one significant digit, that means that the real value is somewhere in the 25-35x range. That gives us a range of growth rates of 192% to 227%, assuming annual compounding... considerably less than one significant digit of uncertainty in the growth rate.

For linear (non-compound) growth, 25-35x growth over 3 years corresponds to 800%-1133%. That range is not easily exactly describable by significant digits either.

Significant digits suck. They are an imprecise and opaque way to express relative uncertainty.

You're right, continuous growth is a much more useful model for this sort of thing, but calling is "115% growth per year" is nonsensical in that case.

I still maintain that this whole mess is due to the fact that people assume percentage growth to be nearly linear (because it's so often used to describe low growth rates like 1-10% where there's no confusion), and often forget to add the initial 100% contribution (pretty much for the same reason).

Yet another math fail. It's actually 966% per year, never compounded, because (1 + 3×9.66) ~= 30. It's incredible how many people make this simple off-by-one mistake.

Furthermore, non-compound interest has no bearing on any real-world financial or scientific calculation. It's a simplification that's just used to introduce students to the power of exponential growth and compounding, by way of comparison with linear growth.

Can you imagine a bank that offered 10% interest per year on savings accounts, but non-compounded? The clients would withdraw all their money every few days, and then re-deposit it and... boom! It's compounding all over again.

Funny indeed, but you've made the same off-by-one error as nearly every other posters here. 1000% growth every 2.031 years... what does that give us in 3 years?

(1 + 10) ^ (3/2.031) = 11 ^ (3/2.031) = 34.5

30x in three years is actually 1000% every 2.115 years, because:

(1 + 10) ^ (3/2.115) = 30

As far as I can tell, when I entered this thread, the number of people who actually understand that percentage growth represents the difference between the initial and final values rather than a multiplicative factor... that increased by 50%. (Previously, it was bunratty and some AC.)

WTF are you talking about? This is a trivial exponentiation problem and you're coming up with some crazy, imprecise and dead wrong solution.

Growth rate of 115.5% per year gives growth over 3 years of: (1 + 1.1155)^3 = 9.46

We want 30-fold growth in 3 years. That's 211% per year, since (1 + 2.11)^3 = 30.

Growth rate percentages over 100% are often confusing, because it's ambiguous whether they refer to the change from the original value or the total final value.

In the usual convention, 310% growth per year would mean a quantity that increases by a factor of (1+3.10)=4.10 per year. So 30x growth in 3 years would actually be an annual growth rate of 210% per year (because (1+2.10)^3 is about 30).

For this and other reasons, percentages are a horribly confusing and unintuitive way to describe high growth rates. It seems much clearer to me to say that AT&T's data traffic is approximately tripling every year.

Indeed. If something grows 30-fold in 3 years, it's growing at an average rate of about 211% per year, because (1+2.11)^3 = 30.

Apparently the author doesn't understand the difference between linear and exponential growth or the terminology used to describe them. I think that percent-per-year growth is almost always a confusing way to describe growth when the percentage is more than 10-20%. That's because people assume x% growth in one year corresponds to 2x% growth in two years. Which is pretty close to the mark for a small growth rates (say 5%), but way off for larger growth rates. It's easily seen in the Taylor expansion of the exponential function, e^x = 1 + x + (x^2)/2 + (x^3)/6 + ..., which is nearly linear for values of x1 but highly nonlinear for larger values.

Too hot? Have you been to Western North Carolina?

It's a mountainous, heavily-forested region. Snow in the winter, sometimes heavy snow. The Carolina coast can really roast in the summer, but it stays pretty cool in the mountains. I was born in North Carolina but mostly grew up in Michigan, and prefer cold to warm areas. Western Carolina is certainly warmer than mid-Michigan, but it has a rather pleasant climate in my opinion.

... is not data. The singular of anecdote is also not data.

Basically, one guy used one computer with flash on a few times, and with flash off a few times. What web sites was he looking at? How many did he have open at once? Why couldn't he just use an ad blocker rather than kill Flash altogether?

I think Adobe Flash sucks as much as the next FLOSS fanboy, but this is just an insubstantial anecdote. Couldn't the author at least run PowerTOP or some Mac equivalent, and try to figure out how much the processor is waking up with/without flash, how much disk is being used, etc?

Microsoft encouraged companies to build in-house web apps on top of IE6, using its many poorly-documented proprietary features. Many of those features were so poorly documented and maintained by MS that they won't even work with newer versions of IE!

Obviously, this was a poor decision on the part of a lot of IT departments and corporate web app developers, but I do think Microsoft deserves a good part of the blame for encouraging such departures from web standards.

(Writing this from Chrome, while I wrangle a recalcitrant IE6 web app in another window... )