Disqus - Latest Comments for aaronsw

Re: Raw Meat: A Libertarian Puzzle

Aaron Swartz — Thu, 20 Dec 2012 07:11:46 -0000

Nowhere, wrote it and decided to just post it here.

Re: Raw Meat: A Possible Explanation for the State-National Poll Divergence

Aaron Swartz — Wed, 31 Oct 2012 12:04:34 -0000

Cell phone polls sometimes also use RDD, though.

Re: Raw Meat: The simple function call

Aaron Swartz — Sat, 01 Sep 2012 07:38:40 -0000

In Python, 1b and 1 not only share the same mindset -- they have exactly the same implementation. I guess in Ruby you'd do it with a hash:

# version 1c
choose('color', from: ['red', 'black', 'green'], ignore: 'ie7')

but the difference doesn't strike me as significant.

I didn't say fluent interfaces _were_ "random poetry" -- I said they had the _same problem_ as the poetry: they used the language for its English properties rather than its Python (or Ruby or whatever) properties. The Perl snippet is there so you can understand the difference. That they are extensively used and old is irrelevant. I suggest you try rereading this section of my piece.

> For instance, did you go to the RSpec source code to see what the "should" method implementation looks like? Isn't it very clear by itself?

No, it's not clear at all -- it has the same problem as Rails, the problem I discuss at the end. It's so unclear that it's pretty much unusable by a mental-model programmer.

To say that RSpec is OK, because it's just another language, like SQL, is absurd, because it's not another language. It is, quite obviously, Ruby, and since you're testing Ruby code it couldn't really be anything else. To write the Ruby code you need to read it as Ruby, but to work with RSpec you need to read it as RSpec (or worse, English). Doing both is what drives you mad, just like trying to read the poetry as both English and Perl.

In his book, _The Unschooled Mind_, Howard Gardner collects a series of studies that find the key difference between kids who learn to program and kids who fail at it is that the failing kids treat their programming language as English. They create a variable called "biggest_number" and are upset that the computer isn't smart enough to remember to keep the biggest number in it. The successful kids give up on it as a language and see it instead as a series of logical rules to be carried out precisely; they're mental-model programmers.

Things like RSpec push in exactly the opposite direction, literally taking us back to the state of kids who are unable to program. That's a dangerous habit.

Re: Take a step back (Aaron Swartz's Raw Thought)

Aaron Swartz — Mon, 20 Aug 2012 10:57:55 -0000

I haven't deleted any comments (except a test one I made). One got caught in the spam filter and I just freed it.

Re: Take a step back (Aaron Swartz's Raw Thought)

Aaron Swartz — Thu, 16 Aug 2012 16:30:27 -0000

In the middle of it, but really enjoying it.

Re: Take a step back (Aaron Swartz's Raw Thought)

Aaron Swartz — Mon, 13 Aug 2012 16:24:51 -0000

Related: http://sivers.org/comfort

Re: Open Data Creates Accountability - Sunlight Foundation Blog

Aaron Swartz — Sat, 07 Jul 2012 13:11:07 -0000

Yes, the line between documents and databases is a somewhat fuzzy one, but that doesn't mean it doesn't exist. I think it's pretty clear that grepping bill text is on one side and summing campaign contributions is on the other. And if your theory is that, well, we'll learn out how to do useful things with data someday, why not save the open data advocacy until then? There's plenty of other data sets to practice on in the meantime.

As for the argument that things are better now because the ads are public, I don't think that's quite right. In the old days, people handed suitcases full of cash to LBJ's deputy, which he spent on ads. Now people wire money to Barack Obama's former deputy, which he spends on ads. In either case, we can see the ads and make judgments about whether to believe it. What's the difference?

I think I kind of made this worse by conflating personal corruption with political funding, but the personal corruption story also hasn't changed much. Before, lobbyists would buy beefsteak, bottles, and blondes for legislators. Now they add the money to a charge account which can be cashed in when the legislator retires and becomes a lobbyist. Again, what's the difference?

Re: Open Data Creates Accountability - Sunlight Foundation Blog

Aaron Swartz — Sat, 07 Jul 2012 12:56:32 -0000

Yep, there's no question FOIA can disguise wrongdoing. And I'm not sure FOIA has a great track record of exposing wrongdoing as such. The reason I like FOIA is because it lets you get more of the gestalt of what things are like inside the government: the way they think about issues, the words they use to talk about them, etc. It's far from perfect transparency, but it is interesting in a way that database releases usually aren't.

Re: Open Data Creates Accountability - Sunlight Foundation Blog

Aaron Swartz — Fri, 06 Jul 2012 14:29:40 -0000

John Wonderlich is one of my heroes, so I'm very grateful that he gave my piece such careful consideration. But I still feel like we're talking past each other a bit.

My concern about open data is not with the "open" part -- I agree that people should know more about their government -- it's with the "data". My claim is that these goals (transparency and accountability) are best achieved through documents and investigations, not databases. And my question is: when has a database really helped increase accountability?

John responds by pointing to FOIA, FACA, and FECA. FOIA requires the government to release documents; exactly the kind of non-database-focused measure I support. As for FACA and FECA, John points to general points (most politicians take money from corporations, most FDA advisory panels are funded by the drug companies) that could easily be made through documents and investigations, without even looking at the data that's been released (e.g. John doesn't use a single name or number). So my question stands: when has a _database_ really helped increase accountability?

In response to my point that transparency just moves corruption one step further into the shadows, John replies "Well, at least we've gotten it out of the street!" But why is that a good thing? "Huge personal bribes and free mansions are political currency in other countries," he brags, "and in the US they’re the rare exception." Should we be happier that companies buy votes through multi-million dollar SuperPAC ads rather than multi-million dollar homes? If so, why? Because it looks cleaner on TV?

These are the questions I'd like to see answered. I hope John will indulge me by attempting to do so.

Re: Oh, The Places Asana Will Go – A Dr. Seussian Poem by Our Kenny Van Zant

Aaron Swartz — Mon, 02 Jul 2012 21:24:08 -0000

Thanks for sharing this. It's terrific and it's really nice to see inside the slick shiny box that is Asana.

Re: Shots For Signatures

Aaron Swartz — Thu, 21 Apr 2011 14:44:20 -0000

The even more interesting question is how to do you design an appropriate price discrimination strategy for buying votes -- obviously there are a whole lot of people who will give you a vote for free and a whole bunch more who will do it pretty much at-cost. But are there enough of them to win the election? And if not, how do you get their votes at that price while paying more for the increasingly-expensive marginal votes you need to get a majority?

Re: Galston & Kamarck on Moderates

Aaron Swartz — Thu, 24 Feb 2011 09:38:25 -0000

They're not shadow caucus leaders, they're called the Majority and Minority Leaders. It would work just like in the Senate (nobody looks to the President pro tempore to set the agenda).

Re: The Confusion

Aaron Swartz — Mon, 24 Jan 2011 17:49:06 -0000

What causes these weird update numbers?

Re: Vertical Integration

Aaron Swartz — Mon, 17 Jan 2011 17:12:12 -0000

In what way does not including a Flash plugin protect their stranglehold? It has two effects: it makes Apple's products less attractive (the opposite of what you suggest), and it discourages the use of Flash on the Web. But those discouraged don't leap to make iOS Apps instead; they use HTML5 which works on Android just as well.

Re: Vertical Integration

Aaron Swartz — Mon, 17 Jan 2011 17:09:51 -0000

Lucky? It seems to me Apple can extract the rents because the monopoly rights they have to their software and hardware designs are even more exclusive than the carriers control of the spectrum. Apple was able to play AT&T off Verizon in gaining access to spectrum, but there's no one else those companies can order an iPhone from, since the government has outlawed their production by anyone other than Apple.

Re: The Case For Federal Higher Education Policy

Aaron Swartz — Sun, 16 Jan 2011 20:11:47 -0000

Doesn't this argument go one step further? Why do we only fund American resesarch? Since academic work is published to the whole globe, wouldn't American citizens benefit from funding good research no matterd what country its conducted in?

Re: an answer to John Gruber: Google dropping H.264 is good for everyone

Aaron Swartz — Wed, 12 Jan 2011 20:03:09 -0000

What about H.264 is closed? I think you mean "patent-encumbered".

Re: The Public Domain That Might Have Been

Aaron Swartz — Mon, 03 Jan 2011 11:01:57 -0000

In particular, the issue is that they decided already-dead people lacked incentive to innovate. There's absolutely no reason to make a law that retroactively gives incentives to already-dead people.

Re: Inclusionary Zoning

Aaron Swartz — Mon, 03 Jan 2011 10:58:53 -0000

Do you support Section 8?

Re: Building a censor-resistant web with distributed hashes

Aaron Swartz — Thu, 23 Dec 2010 04:32:49 -0000

What does that mean?

Re: Building a censor-resistant web with distributed hashes

Aaron Swartz — Thu, 23 Dec 2010 04:32:49 -0000

What does that mean?

Re: Analogy of the Weekend

Aaron Swartz — Sun, 19 Dec 2010 11:14:56 -0000

A shocking number of Bell Labs employees work at Google Labs now.

Re: Scripting News: A Web Trust to publish and store our creative work

Aaron Swartz — Mon, 13 Dec 2010 17:34:27 -0000

I wrote some ideas about how to technically implement such a system on my blog:

http://www.aaronsw.com/webl...

Re: Open Data needs Open Source tools - A Healthy Information Diet - InfoVegan.com

Aaron Swartz — Tue, 28 Sep 2010 12:00:00 -0000

What about thedata.org?

Re: Joaquin Phoenix and the 'I'm Still Here' hoax: So maybe he really is as great an actor as Brando

Aaron Swartz — Fri, 17 Sep 2010 00:18:30 -0000

Definitely increased.