Proteins, Code & Fashion

AngularJS makes it easy to override their default exception handler. This means we can plug in Raven-js to send all client exceptions to Sentry :)

Just create your own module and service with the same name as the default AngularJS provider ($exceptionHandler):

angular.module('myExceptions', []).factory('$exceptionHandler',
          ['$window', '$log',
  function ($window,   $log) {
    if ($window.Raven) {
      console.log('Using the RavenJS exception handler.');
      Raven.config('YOUR_PUBLIC_DSN').install();
      return function (exception, cause) {
        $log.error.apply($log, arguments);
        Raven.captureException(exception);
      };
    } else {
      console.log('Using the default logging exception handler.');
      return function (exception, cause) {
        $log.error.apply($log, arguments);
      };
    }
  }
]);

Just get your proper Sentry public DSN and add your allowed domains in the Sentry security settings for your app.

If you want to increase the speed that the dock hides and shows, and remove the delay so that you don’t waste those precious milliseconds, copy/paste these commands into your terminal:

defaults write com.apple.dock autohide-time-modifier -float 0.05
defaults write com.apple.dock autohide-delay -float 0
killall Dock

You now have a fast hiding dock!

In December of 2010, NASA published a paper in Science (Wolfe-Simon et al. 2010, A Bacterium That Can Grow by Using Arsenic Instead of Phosphorus.) in which they claimed to have isolated bacteria that substitutes arsenic for phosphorus on their macromolecules and metabolites. Of course, the implications of this finding would have been immense… had it been true. Soon after publishing, Dr. Rosie Redfield presented meticulous criticism of the original paper on her Wordpress blog (http://rrresearch.fieldofscience.com/2010/12/arsenic-associated-bacteria-nasas.html), and was essentially shunned by the NASA group, citing the non-official nature of her blog. It took one whole year for Dr. Redfield to submit her rebuttal paper to Science (http://rrresearch.fieldofscience.com/2012/01/arseniclife-manuscript-has-been.html), and five months more for her manuscript to be accepted for publication. Overall, it took one year and a half for traditional publishing to catch up to a Wordpress blog.

If there’s one thing in this world that needs to die and be reborn it’s the scientific publishing industry. One hundred years ago, the centralization of scientific publishing was necessary because content distribution was a hard problem. Now, as we all know, the problem of freely distributing content has largely been solved by the internet. Scientific publishing, as it stands today, is actually a hindrance to scientific progress. I’m not talking about the scientific method here, but purely the mechanisms by which we organize, compile, and distribute the vast amounts of scientific information.

Science is an old world institution where credit and reputation come mostly from journal article authorships. You can tell how well a scientist is performing by looking at their list of publications and by calculating something called an “impact factor”. An impact factor gauges the importance of a paper by calculating the number of times it is cited in the two year period after it is published. This is kind of like looking at the financial statements of a business. Both scientists and businesses require money to fuel their operations, and both must generate value from the fuel. In business, this value is typically represented by money. In science, value is knowledge in the form of publications. As a result, science has developed a strong dependence on the academic publishing industry. Over the past half-century, publishers have leveraged this dependence to build up the $20 billion scientific publishing industry that exists today.

Unfortunately for science, the traditional publishing model has become costly, painfully slow, and inefficient. It can cost anywhere between $1000-$5000 to have a paper published and even more (upwards of $8000) to publish an Open Access paper (free to read without a subscription). It can take up to a year for a paper to become accessible to subscribers (and/or the public) as it goes through the drawn-out publishing process. I’d venture to say that, because of this process, the vast majority of papers that come out today report on data that’s over a year old. Scientists are busy people, but they do most of the heavy lifting throughout the process. In academic publishing, unlike many other industries, the two most important inputs are provided virtually free of charge by researchers. Publishers argue that they add value by supporting peer review, typesetting, printing, and web publishing. The reality is that their main role is in representing their brand (the journal’s brand). All other processes can be fully automated over the internet (and they basically are already!).

As a way of mitigating the costs of publishing, some fields such as physics and math have collectively decided to bypass traditional journals altogether and use Open Access “e-print archives” such as arXiv.org. But, biologically-oriented fields are way more stubborn, most likely because of how afraid many researchers are of being scooped and the potential for valuable intellectual property. They’ll be sticking with traditional publishers until a reliable and trustworthy alternative comes along.

Between the rise of Open Access publishing houses and preprint servers such as PLoS and arXiv, and the more recent global boycotts of specific publishers (http://thecostofknowledge.com/), a lot of change is starting to happen. Many people have written about this issue (http://michaelnielsen.org/blog/is-scientific-publishing-about-to-be-disrupted/, http://www.nytimes.com/2012/02/14/science/researchers-boycott-elsevier-journal-publisher.html, http://scholarlykitchen.sspnet.org/2010/01/04/why-hasnt-scientific-publishing-been-disrupted-already/) and awareness is growing, but the future of scientific communication has yet to be discovered.

You know what’s fun? Google Insights for Search. That’s what’s fun. You know what else is fun? Asking google for fashion trend advice. It’s fun because it’s kind of like asking Ronald McDonald for cooking tips.

Here’s what’s happening with “hawaiian print”:

In the #menswear world, you might think that everyone’s wearing Hawaiian prints. Maybe that’s true in that world, but if we venture out into the mass market it looks like Hawaiian print was hot in 2004… and kind of died off since then. As we might expect, it seems to peak mid-summer, every summer.

How about “camo pants”:

Camo pants are strange because there are all kinds of people that wear them, so it’s not as easy to know if they are actually “in style” or not… BUT here’s the kicker. Camo pants usually peak in the fall/winter season. This year the peak has turned into a plateau! Camo pants (or camo everything!) might actually be in style this summer!

Here’s a cool one, “selvedge denim”:

Now there’s a positive trend. Pretty clear that lots of men are finding out about selvedge denim these days. That’s cool.

Ok, so here are the last two interesting ones. Bow ties. Are they back in style or not?

Yes. Searches for bow ties are the highest ever in history (well, since 2004 to be specific). Bow ties are clearly coming back, at least for the time being.

One extra little thing I thought was interesting was the data on “how to tie a bow tie”:

Once again, a slight increase in the past couple of years… but what the hell are those peaks in December? WHAT IS HAPPENING IN DECEMBER? Someone please tell me… Christmas? what is it?

That’s all for today folks!

In the next 10-20 years, every baby born in the developed world will be allocated their own personal IPv6 address.

Why? Well, wouldn’t it be cool if everyone just had the same phone number for life? Fine, so there are services like Google Voice and the Yellowpages but come on, how many people actually use these things in real life? Instead, what if everyone just had a single address that they could link to their phone, computer, computer-phone, brain-implant-phone, tooth-implant-phone, eyeball-phone, or whatever you might have in your pocket in 10 years, to your personal IPv6 number!? Think about the possibilities!

Also.. IPv6 also has cool integrated security features that should make the world a better place.

Today is dedicated to IPv6 Launch Day and Giant Steps by Coltrane.

Today at Dolbeau I wanted to make a bell ring when a new order comes in… just like Jeff Bezos did in the early days of Amazon. I wanted to make sure that our bell rings in any browser that has the bell page open. It turns out it’s super easy to do using NodeJS/Express and Socket.IO.

This is how it all works:

1. Load the URL in any number of browsers
2. Send a POST to the same url
3. Listen to the bell ring in all open browsers!

Also you can specify which sound you want to play by using the ‘type’ POST parameter. Just check out the JS and HTML files and you’ll be good to go!

Fork it on GitHub!

Warning: This post is for people who know what Python, Django, and Debian are.

It turns out that most people ever only need a single VPS node (or less) for their Django site. When it comes to small sites, using Chef and Puppet for deployment is kind of overkill. In our previous iteration at Dolbeau I implemented Chef with the Deploy Resource. It worked alright, except for the fact that my deploys took way too long, not to mention it was super overcomplicated and scary when it failed. I wanted to convert my huge ass Chef deployment into a tiny Fabric script (with the help of Cuisine), so I did!

You can find it in a Github repo called deploy-django.

The stack is composed of:

• Nginx
• Gunicorn
• Django
• Memcached
• Postgresql

Supervisord is basically optional, but you need to edit the fabfile. Right now I personally just restart guncorn using a kill -HUP during the deploy process. This seems to work fine for my purposes. I’ve included an example init script in the provision folder.

To use the fabfile, you first need to install the requirements:

pip install -r requirements.txt

Then, edit the deploy_settings.py file:

APP_NAME = the name of your python app
APP_REPO = the git repo
APP_REPO_BRANCH = the branch you want to deploy
APP_HOSTS = a list of hosts
REMOTE_USER, REMOTE_GROUP = the linux user that should own the deploy directory on the server

DJANGO_STATIC_ROOT = the directory name of your static_root in your app
DJANGO_MEDIA_ROOT = the directory name of your media_root in your app

DEBIAN_PACKAGES = a list of debian packages to install, the default list contains a bunch of things that you may not need

DEPLOY_TREE = defines the deploy directory tree. The default deploy tree creates a /srv/<APP_NAME> for your app.


DEPLOY_TREE = {
    APP_NAME: {
        'releases': {},
        'shared': {
            'logs': {},
            'media_root': {},
            DJANGO_STATIC_ROOT: {}
        }
    }
}

Then, you edit the server config files in the provision directory. The provision directory is structured as if it is the root directory of your server’s disk and it contains basic config files for your stack. You will need to edit these files (and rename some of them) before trying to deploy, they are NOT templates.

/etc/
    rc.local
    init.d/myapp
    gunicorn/myapp.conf.py
    nginx/
        nginx.conf
        sites-available/myapp
    postgresql/8.4/main/postgresql.conf
    supervisor/conf.d/myapp.conf
/srv/myapp/shared/settings_local.py

Then you can try to initialize, provision and deploy:

fab initialize

This will set up the required user and inject your public ssh-key for passwordless login.

fab provision

This will install the packages and set up the server config files

fab deploy

This will deploy your app!

It should all work… if it doesn’t then feel free to file an issue or fork the repo.

Brands are all about emotion (duh!). It’s funny because when you’re working on a startup that designs, manufactures, and sells mens accessories, it’s really easy to forget that simple fact. Especially when the company basically consists of a handful of people doing all kinds of random shit. Also, I have to admit that sometimes I’m just not that good at emotions.

Think about Tiffany & Co for a second. Tiffany’s makes you think of love. Not just any love mind you… true love. The only love that’s worth loving for… obviously. That’s what they’re about! They even have an entire website about true love (http://www.whatmakeslovetrue.com/). I bet you didn’t know that! So what that they’ve been around since 1837… I’m pretty sure love has been around for longer than that. You have to hand it to them though… they really kill it with their branding.

Now, back to reality. The challenge with a new company is that the possibilities are endless. We can do whatever the fuck we want! Sounds fun in theory but it really isn’t easy for a control freak like myself :) At some point a decision has to be made about the central “tenet”, or to be super nerdy, the prime directive of the business. If you can pick 1-3 words that describe the emotion of your business, what would they be?

At Dolbeau, we’ve made some mistakes and we’re still working it all out (as we should be), but we have a pretty good idea of where we want to go. We think we can be pretty funny dudes sometimes, but we’re also deadly serious. So there you have it: serious, with a side of funny. There has to be a better way to phrase it though… but we’re making baby steps here.