Pi

If you think of pi as a simple ascii-replacement cipher, then, at some point, your last name will be encoded within it.

Pi is an irrational number which means that (among other things) it has no repeating number patterns - which suggests that the number sequences in pi occur randomly.

Not only is your last name encoded within pi - your first and last name are as well (at some point).

Your entire families names as well - in order - are also encoded in pi.

Your favorite book - word for word - is encoded in pi.

The Bible, Quran, Bhagavad Gita - all encoded in pi.

Every word you’ve ever said, in the order you’ve said them, is encoded in pi.

Every word anyone has ever said in the history of human kind is encoded in the infinite depths of pi’s numeric sequence.

Every word anyone will ever say until the sun blows up - in pi.

Trippy.

However, when there were similar numbers getting in touch to ask if Rebekah Brooks was really in trouble for raping a monkey, and why the BBC was claiming that, as a special summer treat, the Prime Minister had told the nation’s teenagers they didn’t have to pay for anything any more, we realised something was wrong.

I don’t think you can make this stuff up! /HT Paul Stack

BBC sign language interpreter sacked for ‘changing the news’ | The Poke:

Source thepoke.co.uk

Healthcare.gov Meltdown: An Alternate Explanation

I’ve been fascinated by the meltdown of healthcare.gov. Such an important site, so politically charged - and it’s not working at all.

I’ve read a number of articles on this - starting with the criticisms:

And then these posts, which were written just before (or at) launch:

If you read any of these - make it the last one. It’s a fascinating look at the way a CTO thinks - about Web Servers and the current web development culture.

Anyway - back to my alternate theory. Everyone keeps saying the site is bad:

Part of the problem, according to a number of designers, is that the site is badly coded, which makes the traffic problems more acute

Here’s the thing: aside from the large amount of javascript - there is no code behind this site.

It’s completely static - written with Jekyll (a static site generator written in Ruby).

So how, then, is the site failing so badly?

Here are the problems that I’ve read about:

  • No one can sign up reliably
  • Confusing, poorly-written directions
  • Poor user experience
  • Intermittent “Site is down” messages

In everything I’ve read there’s never been a 502/503/504 error on the site which would indicate server load. There *has* been a report (I can’t find the link) of a 500 occurring on a JSON look up.

This got me thinking.

What if the site is *too fast*.

Think on this a moment: it’s being hammered right now due to the outage etc - but it’s still up. It seems to have always been up - right through the traffic problems.

The API calls, however, are a big problem. And I think I know one reason why… have a look at this:

This snippet of code is from “all.js” - a massive concatenated dump of … some really bad Javascript.

What you’re seeing is a jQuery function that looks for all links with the title “glossary” on the page, and then, for each one, it’s executing a remote call via $.getJSON.

That is a really, really, really bad idea if you want to keep your API servers up and happy. 

And that’s what I think is happening:

Poorly-coded Javascript is being served rapidly to a whole lot of people, which is just slamming the API servers… wherever they are.

I’ll guess the API integration was a black box to the Devseed team - “just send the data over here”. 

I’ll guess further that the Devseed team had no way of knowing what kind of load the API endpoints could handle.

And I’ll guess finally they didn’t understand that it would all look like their fault when these servers didn’t respond properly.

OK, back to work now…

They were planning 32 servers, between staging, production and disaster recovery, with application servers for different environments,” said Cole. “You’re just talking about content. There just needs to be one server. We’re going to have 2, with one for backup. That’s a deduction of 30 servers.

"Everyone has a plan - until they get hit" - that’s Mike Tyson. 

I can’t imagine going live with the Obamacare website… running on 2 servers.

Open by design: Why the way the new Healthcare.gov was built matters | E Pluribus Unum

Source e-pluribusunum.com