msg

Read Dan Ariely's chapter 1 in Predictably Irrational. It is about relativity and the decoy effect. If you have a choice between A and B, then adding a third choice that looks like a defective version of A (the decoy) tends to nudge people to pick A over B.

As he explains it, choosing between two options in a vacuum is hard. You have to list pros and cons and weigh them carefully. We prefer to create yardsticks and say "better" and "worse". When you add the defective option you create a no-brainer yardstick in the mind of the chooser.

Doesn't this sound suspiciously like New Coke? I don't believe that the marketers didn't know this stuff, or that they were caught by surprise by what played out next.

I deduce from this that the fault was not in the language, but in the company, specifically in the resume filter that asks for 3+ years of production Clojure. A programmer who learns languages can learn Clojure for production work. That's better than trying to bring a soaring program back down to earth (or for C++, chaining it to a cart and dragging it into the salt mines).

But don't listen to me, here's Paul Graham from Beating the Averages (Footnote Edition). And ponder this question: why didn't Yahoo just rewrite the templates in C++?

[1] Viaweb at first had two parts: the editor, written in Lisp, which people used to build their sites, and the ordering system, written in C, which handled orders. The first version was mostly Lisp, because the ordering system was small. Later we added two more modules, an image generator written in C, and a back-office manager written mostly in Perl.

In January 2003, Yahoo released a new version of the editor written in C++ and Perl. It's hard to say whether the program is no longer written in Lisp, though, because to translate this program into C++ they literally had to write a Lisp interpreter: the source files of all the page-generating templates are still, as far as I know, Lisp code. (See Greenspun's Tenth Rule.)

It is about whether minority viewpoints get a fair shake on Wikipedia.

As it turns out, no.

And it turns out minority means "outside the male, machismo-dominated world of the alpha techno-geek."

It also turns out that even esoteric technogeekery is outside the tiny circle.

And it only takes one alphageek to make it so.

This is an organizational problem with Wikipedia, not just an isolated incident.

Sony's private key is, essentially, a random number.

As for copyright protection, you should reflect on this page:

http://www.cs.cmu.edu/~dst/DeCSS/Gallery/index.html

And then this one:

http://www.cs.cmu.edu/~dst/GeoHot/

To the point with Harry Potter: you can think of a continuum of infringement. If you tear a page out and copy it, you will violate the copyright. If you reuse characters or scenes and use them to create your own work, you are making derived works (JK put her stamp of approval on non-commercial fan fiction by the way). But the farther you get away from the original characters the less you can call it infringement. As the link between the two works becomes more and more abstract and less concrete, the new work is more protected. At some nebulous point it is no longer infringing.

For instance, Harry's story and appearance is superficially similar to an earlier work called The Books of Magic, a comic series by Neil Gaiman. You can read his take on it here:

http://journal.neilgaiman.com/2008/04/fair-use-and-other-thi...

http://en.wikipedia.org/wiki/The_Books_of_Magic

Essentially he throws up his hands and says "borrowing is a fact of life in genre fiction. Deal with it, I did."

How does this apply to computer code? Well, the farther away you get from the original copyrighted code, the less infringing it is. There are extremely limited protections for derived works that amount to Person B having a copy of Person A's code in front of them and editing it.

What those Touretzky pages from CMU show is that there is no such thing as a continuum from computer code to an uncopyrightable abstraction. It is abstraction, mathematics, numbers all the way down. In fact, it is speech that is protected by the First Amendment and suppressing it is a constitutional issue with narrow exceptions.

So no, you can't use a book model for code. It is not the same thing at all. Executable ideas, math, algorithms, just don't exist in the fiction publishing industry.

That is what copyright is for, not patenting. If someone doesn't have a license to distribute your code, you can collect damages for copyright infringement.

But patenting covers independent inventions of the same algorithms. And it's frequently used for ideas that are profoundly obvious, like a straightforward application of machine learning techniques to a particular data set (Bilski). So if you think of an idea by yourself, you can be sued for implementing it.

As a developer, it is recommended that I not read patents, as if it is determined that I read a patent that I then violated, it will be triple damages in a lawsuit. So I have no way to avoid implementing somebody else's patent.

Now Sony's lawsuit isn't about software, or patents. It is about certain special secret integers. It so happens that used with readily available software they blow a wide hole in Sony's horrible security system for the PS3.

But Sony doesn't own those integers. It doesn't own the hardware it sold to people under pretenses that it would always run Linux, and it doesn't have the right to control how people would choose to use their hardware and GPU. Hotz never used PSN software or any of Sony's other products.

When I used it Qt was dual licensed, one way for commercial closed source, and GPL for open source. It can just be forked right?

Outlook for Mac 2011 came out. It has some major missing features (won't run rules on the Exchange server, no auto-archive), but the calendar works and it's a serviceable mail reader. Better than Entourage by far, I would say.

Google's spam filter is training the internet to produce better and better content. Over time a rising tide lifts all boats.

In 2000 they were working on not returning results from various subdomains off the same base domain.

This year they're working on content farms.

Next year we won't be talking about plagiarized or duplicated content. We'll be talking about Google giving us content that is superficially different but not unique enough.

Five years from now we'll be talking about how Google is giving us unique, relevant content in the first ten results, but it's not personalized enough. Et cetera.

Just like in the AI world, every time a computer does something we thought only a human could ever do, instead of admitting defeat, you say, "This just proves that chess|poetry|music composition|playing Jeopardy doesn't take real intelligence." The truth is that you're not defining intelligence, you're moving goal posts. And if there's one thing Matt's post shows, it's that we moved the goal posts.

I think Matt likes the criticism. When Google's critics have to move to more and more sophisticated attacks, it shows that Google's facility is growing.

Google is not just raising the bar on the whole internet. You are raising the bar on Google.

When you're on your deathbed, you're not going to tell your grandchildren "I sure wish I had upheld the strict letter of copyright law on the internet more." But you might say "remember that time I saved the BBC from itself?" And like Big Fish, it will be almost true.

Unless Bing weighs clicks on Google SERPs more strongly than clicks on Bing SERPs! After all, Google is the market leader...

This is the problem, we really don't know how hollow Bing is any more. If it's a thin UI over results that grow closer to Google, it's intellectually shallow. I wouldn't trust Bing search to combat, e.g., click fraud and SEO better than Google. If anything I would expect the edge cases and tricky behavior to derange Bing further.

I work for one of those big companies. After going over trouble spots in the resume and asking why they want to work here, I tend to ask one or two very similar questions that 1) force candidates to do nontrivial coding, algorithms, and design while writing a complete program, and 2) elicit a wide range of responses. I don't require a certain language or ding on syntax. We usually spend the rest of the hour on the question.

In general, bad candidates get me hinky in the resume portion (or worse, don't really want to work here!) then crash and burn on the programming question. Borderline candidates do ok in resume but make me feel like they're not that comfortable programming. Great candidates usually do something clever during the problem.

I believe that doing homework isn't convincing. It's too easy to cheat. That's probably why the commenter elsewhere on the thread got bounced from Google. Doing week-long trial runs would be ideal, but impossible at my company. We sometimes go contractor-to-hire though.

Can you be a great programmer and fail my interview? I think it boils down to three things: skills, communication, and pressure. If you lose your skills and communication under pressure for an interview day, you probably won't survive in my environment over the long haul.

I thought he was saying something about tankers.

They're not just susceptible to Google bombing. They are at the mercy of some easy to automate SEO with the bing toolbar installed. If I were a black hat I'd install the bing toolbar, craft some perhaps human looking Selenium code, and start the clicking. Over a period of a few weeks I'd be owning the top results in Bing.

Look at it like a machine learning problem. If you slavishly match the test set you are doing something called overfitting. Your performance on the rest of the web will decrease, because you inherit bias from the test set.

I agree we should stamp out racial prejudice everywhere it appears, like that well-known racist Mel Brooks and his Canadian caricature, Dark Helmet. Neo-colonialist interpretation: inside every big black man is a whiny white man who can't breathe in this thing.

And let's not forget that closet bigot JK Rowling. Did you know the Dark Lord is really black? The movies got it all wrong.

Or the League of Shadows in Batman Begins... it was called League of Assassins in the comic books, so we can blame the interpolating hand of director Christopher Nolan...

When is knowledge more than mere facts? When it's taken for truth and lived.

The war between science and religion is only real for people who believe that science precludes a religious outlook, and vice versa. Science is about pattern recognition from observation. Religion is about living life from revealed or derived wisdom. Religion can be informed by science if it accepts science as part of wisdom. Science has obvious limitations to only what can be observed and explained or duplicated, and it can't get outside that box.

But if you take a fact and intend to spend your limited life pursuing it, it should be real. It should be the truth. You shouldn't be wasting your life working from faulty assumptions, because as far as we can observe (pace Hinduism), this one life is our only chance to do some damage.

Are there major differences in the life you live if you have a different set of facts? Do you have to decide? I would say yes for myself. But apathy like that of the editorialist will kill any amount of wisdom.

The first one I bought was the hardware accelerated version of my guitar tuner, gstrings. I also bought a cool sequencer/composer called Loops. Over time the barrier has gone down significantly and I've bought more.

My price level for no-brainers feels like $2. More than that and I search for reviews and such.

I tend not to buy unless I can try first.

It is ambiguous: "a lost (life's work)" is the intended interpretation.

"rediscovered" is more like it.

Of course you could talk about "a (rediscovered life)'s work" in an interesting way too...

And where you are driving to!

I clicked through to the Click Forensics site to try to understand where they got the 20% fraud number from. It is not there. There is a press release but no methodology or anything else but the bare number.

Understand I do think they are doing sound work that could identify fraud. They use a learning engine on data like IP, user agent, referrer, and post-click behavior, and extract features.

What's really missing is distinguishing fraud from what they call low value, unlikely to convert clicks. You might be able to uncover fraudulent behavior patterns as opposed to crappy inbound links. But what is the ratio? What is the method? I looked everywhere for this without reading and watching their marketing sessions...

An important point appearing elsewhere in the comments: who is a programmer anyway? If you can write Hello World in C, are you a programmer? If you can't write your own compiler are you a non-programmer? This is an unanswerable question.

Maybe we could focus on kinds of thinking that help you program well. There is nothing you can't learn eventually. However, we might find some common ground by looking at how good programmers have trained their minds.

We can list some desirable skills for a programmer that reach across languages and tools. There is an old saw by Joel on Software that some people just can't understand pointers and recursion. Or on the next level, that some people can't understand multithreaded/parallel programming. I think I'd say that an instinct for simplicity or mathematical beauty and symmetry also plays a role in making good designs. In other words, making not just consistent mental models that you can simulate in your head (as another comment says), but making beautiful ones.

I was sitting with my manager yesterday. He was giving me requirements for a system I'll have to build over the next two weeks. Along the way we came across a problem where we need to take a bunch of multidimensional values and combine them together.

Instead of leaving it to me, he wanted to design it on the spot. He thought of this like voting where each value makes some contribution, but then the requirements dictated that we needed to add things like "some values have more votes than others", and "some votes have veto power".

As he was describing this pseudo-consensus algorithm with weights and vetos, I was surprised he didn't know how ugly this design sounded. It is hard to set up the contribution of each value, hard to predict the behavior, and hard to make intuitive.

It's not that this is a bad, uncreative idea. Voting is an interesting analogy for the problem, and intuition is a good way to take a stab at a problem. It's that when he started working it out he didn't see the multiplying problems, or the counterexamples that would cause trouble for his system. And I do think this is something that good programmers and system designers do well, even without much conscious thinking. They find the attack surface, figure out where their abstraction leaks, jump immediately to edge cases.

I'm going to have to pull him back early next week. :oD

You don't have to say that the programmer has a magic switch in their brain that enabled them to be a genius while everyone else is left out. Anyone Can Cook!

edit: wwhere => where

I don't find interleaving natural. It would be natural if we were using old school hypertext and reading from anchor in the original text to response in the new text. But as it is, I have to skim a whole email again to find the small portion that changed. If the reply is terse enough, I have to reread the large original to get enough context to understand what they are saying.

If the large portion is snipped, you can't read the whole thread in the final message(s, for forked threads), but have to go spelunking through multiple messages.

Also, on some mail readers it is hard to tell the difference between responses and text indented for visibility, like quoting or code or equations. Replies to replies get really messy.

I also don't like scrolling.

I got used to reverse chronological reading with blogs (the train has almost sailed on them, huh).

What top posting forces you to do is to form one coherent complete thought (if you plan to be understood).

About those big-box bookstores... Barnes and Noble is doing just fine.

http://mediadecoder.blogs.nytimes.com/2011/01/06/barnes-nobl...

It's been flat for a decade, so why now?

The Seahawks are in trouble, so he's turning to other interests?

15. My son is autistic and would score quite a bit higher.

I'm a programmer but I love talking, jokes, and making friends.

Like other social sciences, software engineering is an under-specifiable domain for double-blind studies. TopCoder might be a good example of how far you can take it. It provides a wealth of interesting comparisons that extend to the group of engineers that want to go do TopCoder, across the platforms that TopCoder supports. But it's a little artificial environment and doesn't tell you much about collaboration.

The other problem is that there's more variety and less consensus in judging the work of creative human thought than there is in something like focus group response to commercial messages, or even response to sociolinguistic cues. So all the results of your studies on software have high entropy.

Don't harsh on this. Analogical reasoning is where scientific theories come from. Well, perhaps from there and from elegant math.

I thought it was a very interesting metaphor, and you can start thinking of diseases of the national infrastructure, or the infrastructure of the body.

Science shouldn't stop at metaphor, sure. But it often starts there! Think about concepts from fluid dynamics that leaked into electromagnetism.

I usually have two passwords: A stupid one that I only use on stupid sites, and a good one that I use on crucial ones (Gmail, Amazon, bank). There's no relationship between the good and bad passwords, so I feel more confident that I'll be unaffected by data breaches.

I checked the torrent and I guess I never made a Gawker account.

Or a 13/80 difference, which is in the 15%+ range. It's a rounding error on a total time of many seconds, but not in the realm of the very fast.

My sister and parents recommended this to me:

http://www.mistakesweremadebutnotbyme.com/

Time to go get it I guess.