pak

It might have also been

> is their any real analysis

driving some of the downvotes (I did not vote myself).

You wouldn't need to change the color palette. Freeze the gameplay and animate the world "flipping over" as a big 3D surface, like in Paper Mario. Continuity is established, problem solved.

> Provide both then, because typing a URL is a pain on most mobiles, especially urls more than 5 characters.

If people are already running up against 5000 SMS/month limits on their phones, I'd wager that they can spend 20 keypresses typing in a URL for something interesting. Frankly, it's probably the same cognitive load as finding the QR decoder app, lining up a good shot, waiting for the white balance and focus to settle, hitting the shutter button, ...

> What sense would it make? You have an existing brand and an existing site, why have to create a new brand and a new site just to link to that?

Don't understand your point. I'm talking about making a short URL redirecting to whatever the QR code points at, maybe even off the company's existing domain. E.g., http://acme.com/foo

I can't speak to QR code usage in Japan or Korea, I just remember seeing them used in CS course material at my college and none of my fellow students ever used them. Those were real geeks, and they weren't even interested enough to try.

> Wait that has no relevance whatsoever to the previous half of the paragraph, you said "normal people" would always ignore QRCodes, but anyone can recognize a barcode and knows (at least roughly) what it's about.

There's a distinction between recognition, and being able to use them. As far as recognition goes very few non-techy people know what a QR code even does, because there is no obvious indication that it contains a URL. My point was that barcodes usually have the numbers below them and that (IMO) facilitated both consumers and cashiers understanding that it just contains a product number, as opposed to being the Evil Mark of Satan that the red beam reads. It is a more transparent system. I know it seems weird, but for non-geeks, being able to see what is going on is an important step in not demonizing new technology.

> Google Goggles can attempt to do that with all the power of Google behind it.

I could be wrong, but I'm pretty sure all the OCR is done client-side. And what's your point? A QR code containing a URL requires a device with a network connection to provide anything of immediate value.

Here I'll give you a "real world hyperlink": print the damn URL in English so that I can remember it if I don't have my phone handy. If you're running a print ad you can afford a short snappy URL. Bonus, I can even use it on my laptop that doesn't have a QR code app! Or any other browser for that matter!

The main problem I have with QR codes is that humans can't read them. Therefore, they will be relegated forever to the sphere of computerish heiroglyphics that normal people will always ignore, because they have no obvious real world relevance. By way of contrast, most of the old EAN barcodes have the number printed under them, so a human can see what they encode when the barcode scanner breaks. QR codes, not so easy.

As suggested by another commenter, decide on some way of making the URL easy to OCR, perhaps by surrounding it with targeting sentinels and using a standard font. I mean, if Google Goggles can OCR text on a menu or poster in real time... a QR code is just making things complicated.

Also, let me tell you this really great joke: CueCat.

This whole article is predicated on "social search" being better than pagerank-scored search, which is only sometimes the case.

Sometimes I care what my friends think about a search query, e.g., "sushi restaurants near foo." It might help me to know that my friend Bob, who I trust, has been somewhere and thought it was good.

Sometimes I do not care at all what my friends think about a search query, e.g., "medulla oblongata." In fact, showing results that my friends might have liked in connection with that phrase may distract from the scientific information I am looking for, especially if they are pages of quotes from Fight Club.

Being able to distinguish this intent, and putting a good UX on top of it, is what Facebook needs to do to make a significant entry into the search market. (I'm not holding my breath.)

tldr version, minus the usual stackoverflow backandforth:

Named elements are made into properties of the "window" object in Chrome, IE, and Opera. That would include any element with an "id" attribute and most elements with a "name" attribute. Thankfully, any vars you declare in this scope override the DOM ones.

With some quick testing in Chrome, IE, and Opera, it appears that these elements are enumerable off of "window" in Opera but not Chrome and IE (perhaps those two browsers create them only when you access them).

Yep, here was the previous article.

http://news.ycombinator.com/item?id=2847675

"I'll show a second example which does pass data along with the URL" ... never happens.

Also, when replying to people directly, try to thread your reply onto theirs or that person might never see it.

Anyone upvoting this, can you please explain why? This article has numerous issues that would make me hesitate to recommend it to anybody else.

I'm not going to list them all, but here's a start.

- gives the Google AJAX-ified query string without explaining the difference

- perpetuates the 256-char query string limit myth [1]

- totally muddles distinction between POST and GET, e.g., "Therefore, we POST information to a server when we want to give information to us."

- talks about examples that are never shown

- python examples have no error handling

[1] http://www.w3.org/2001/tag/doc/get7#myths

Luckily, throwaway phone numbers are as easy to get as throwaway email accounts these days: http://voice.google.com

Yep, and that's assuming 8 random bytes from extended ASCII. The other point of the article was that nobody actually makes a password from random characters because words are easier to remember. And I think it's disingenuous to suppose people will enter alt-codes and that nonprintable characters would be allowed, so assuming MENSA-quality users with internal random number generators, we get 95^8 ~= 6.6E15, a clear loss of entropy.

Aha, indexing the characters out of "false" and "true"... clever.

Why would I "kick send"? Wouldn't I press send, or touch send, or maybe kick to send, or send a kick (violent)... is kicksend supposed to be a new verb? I can't get past the name.

It sounds like a mispronounced "quicksand", a meaningless mash of two words. Neither the noun+verb nor the verb+verb interpretations are grammatically sensible, and the verb+noun meaning just doesn't feel right.

How do cars defrost the rear windshield? Embedded heating elements in the glass. That's pretty proven, and easy to manufacture. The exact same thing could be used here.

I'm not really



a fan of the padding



on these "new looks".

Also, I liked the little headers that said earlier this week, earlier this month, etc.; why did those have to go?

I don't believe that any DNA virus would replicate via a dsRNA intermediate, since that would be an unnecessary step and therefore anti-selective. You are correct that ssRNA viruses create a dsRNA template while replicating but I would think that this state is transient and therefore not a great target, unless this treatment can target dsRNA strands littered with replication forks.

I am not convinced, especially when the company is hyping it to kill "virtually all" viruses. Far from the majority of viruses contain dsRNA. For example, the Ebola virus and rhinoviruses, which cause the common cold, all contain an ssRNA genome. (For this reason I am confused why somebody quoted in the article specifically uses these examples.) HPV and many other categories of virus [1] contain only DNA. It would seem that only this category [2] of viruses would be affected.

[1] http://en.wikipedia.org/wiki/DsDNA_virus

[2] http://en.wikipedia.org/wiki/Double-stranded_RNA_viruses

Chrome runs a separate process for each tab, and that process includes the layout engine and the JS engine (V8) running in tandem. V8 has to be able to block the layout engine from rendering, otherwise it will draw invalid page states. The execution model (greatly oversimplifying) is that the browser runs a tight loop that monitors for changes to the DOM state and input events, and then synchronously executes JS code bound to these events (and the native handlers if they are not overridden), finally re-rendering parts of the viewport that become "dirty".

In a browser that isolates tabs into their own processes, like Chrome, try opening a tab for www.google.com and then enter this in your address bar:

    javascript:document.getElementById('lst-ib').onclick=function(){while(1);}

When you click on the search box the entire tab will freeze, because the tab process is stuck inside the JS engine running the infinite loop function we bound to the click handler. It's not a matter of bad browser design, that's how the JS execution model is intended to work.

"Web workers" is the HTML5 proposal for running JS code that does not block the event loop. They do not have direct access to the DOM and can only communicate with JS code in the main event loop via a message passing API.

It was broken from the start, since it was introduced in conjunction with Win95 and at that point it was designed to trigger the Start menu with a single press.

How would JavaScript be able to override default event handlers if it weren't tied to the UI thread?

Synchronous XHR will tie up every other browser too--because it's a blocking call, hence the "synchronous"...

A rare example of appropriate use is an onbeforeunload handler that saves your work on a webpage when you close the window, because in this case, asynchronous calls will never complete between the click and the window unloading so you have to synchronously block the unload event handler. Now, go post that on this SO thread [1] and make some points ;-)

[1] http://stackoverflow.com/questions/2088318/is-there-any-reas...

Try hooking your computer up through a hub or something and use WireShark on your traffic. You never know on a Windows box. I've had deeply rootkitted machines that didn't get picked up by any scans.

In any case, I'd recommend against changing your MAC address because when you do get a hold of the network admins they may be rather disgruntled that you're violating network policy (almost any university that registers MACs prohibits MAC spoofing, otherwise you could just use the MAC of somebody that you saw on a different subnet).

I... don't know about you, but if I were a blackhat hanging on to a XSS or CSRF bug in Facebook, I'd expect it to be worth a lot more than $500 on the open market. If it could quietly expose arbitrary user data, it could be worth tens of thousands easy; if it could arbitrarily change it, I don't even want to think about how much that is worth...

Yeah, it can, but perhaps you develop a habit for it whenever creating variable names, so there is some visual harmony between klass, old_klass, klass_dismissed, etc. when reading the code.

Pretty slick--have to say I enjoyed the big mouse cursor coming up and editing the top paragraph, that was rather clever.

In trying the demos, it behaved rather unobtrusively and flexibly, and despite looking a little like Word's contextual ribbon it didn't make me want to set anything on fire after a few minutes. That's a first for an in-browser rich text editor.

I think it's interesting that "old" rich text editors like TinyMCE/FCKEditor tried to copy Word 2003 right down to the massive toolbars, and then barely anybody needed all that crud so it fell out of fashion in newer, slim designs (e.g., StackOverflow, Quora). Nowadays, it appears that Word 2007/2010 has finally gotten enough mindshare for the ribbon/contextual palette model that it's finally being comfortably re-implemented as a web UI widget.

The flash crash was not precipitated by a "fat finger" event but rather an intentional, albeit aggressive, move by one mutual fund, and according to regulators it was exacerbated by HFT's, not "corrected" [1][2][3][4]. Do your research--like so many you probably believed the BS that was handed out to pacify everybody when it was big news, and missed the SEC/CFTC report that came out later. More recent mini-crashes have occurred in commodities markets [5], again driven by HFT.

What is my disaster scenario? Do I really need to explain why extreme volatility can vaporize investor confidence in a market? If an event like this was caused by one market action, you can even imagine malevolent investors trying to intentionally trigger a crash or a mini-crash (holding a short position on the side so that they profit). If enough people are able to do this with regularity, the market turns into /b/ and becomes worthless for traditional users. I won't pretend to have enough imagination to see all the ways this could go wrong; we can leave it up to Wall Street to make it a reality. It would seem like common sense, however, that if neither investors nor regulators have a firm grip on the steering wheel (and maybe that's the best way for the economy to run, who am I to say?), we would be better off not putting the market in cruise control at 600mph without carefully considering all the ramifications.

[1] http://online.wsj.com/article/SB1000142405274870402930457552...

[2] http://www.bloomberg.com/news/2010-10-01/automatic-trade-of-...

[3] http://www.nytimes.com/2010/10/02/business/02flash.html?_r=1...

[4] http://www.reuters.com/article/2010/10/01/financial-regulati...

[5] http://online.wsj.com/article/SB1000142405274870432280457630...

Oh, just because it hasn't happened yet, we should pretend it isn't possible? You remember the flash crash. That took out the market for a day. People waved their hands and said it was an innocent mistake. Nothing changed.

When scads of people were writing CDO's and swaps and other clearly garbage papers in 2007, they probably used your logic to justify their recklessness, and here we are.

If those extra packets could suddenly and uncontrollably leak out on the Internet and destroy everything, requiring a trillion dollar government bailout to fix it and ten years to restore confidence in the infrastructure, yes I would be pissed and doubly so that this inadequate analogy was used to defend it.

Indeed, if Nigerian 419 scammers have already picked up photoshopping IDs [1], you can imagine what a low barrier this will be. The real question is if you can verify the ID electronically from the image, and the answer will generally be no.

[1] http://www.419eater.com/html/hall_of_shame4.htm

None of your points rule out my first implication, and I don't believe the logic for your second point was part of Paul's statement. Despite YC companies being a small part of the world economy (or whatever), they are disproportionately mentioned on HN for obvious reasons.

I appreciate your candor on this, but I'm curious if you are implying two things: 1) you think that people interested in reading about Airbnb are behaving irrationally, and 2) people rarely get angry at YC companies?

I find the interaction between HN and YCombinator rather fascinating, because you obviously have a financial interest in many things posted on the site, but you also clearly feel some responsibility to play the neutral host. I wonder this responsibility comes from HN's value to YC as a barometer of trending ideas (requiring you to retain the faith of your userbase), or from some higher sense of "journalistic" obligations.