In 2012, the photograph of Barack and Michelle Obama embracing after his re-election was 'liked' over 4 million times. That photo, like the 250 million other images uploaded to Facebook every day is standardized; it is a JPEG-encoded image. When Obama's staff uploaded that image to Facebook, the company's software recognized it as JPEG-encoded data and created 4 different sized images using the JPEG compression standard built into its photo-management system. The image became visible to Obama's “friends” (and, apparently, to the NSA’s). It became ‘Like-able’ and ‘Share-able’ generating new data points in Facebook. If Obama's staff had tried to upload a different type of image — a RAW-encoded file or a Photoshop file, for example — Facebook’s software would not have recognized it. It would have refused to upload and the data and image would have remained invisible. The digital image, the political message would not have done its work.
We commonly think of JPEGs as a type of image. We talk of taking JPEGs, sending JPEGs, looking at JPEGs, posting JPEGs. But technically, JPEG is a compression standard, not an image type (the formal name for the latter is JFIF or EXIF). That compression routine is built into a digital camera’s software, taking the data stream from the camera's sensor and compressing or encoding it into a form that can then be read by software in Web browsers and operating systems, photo management and editing applications, and even surveillance systems.
JPEG was originally developed by the Joint Photographic Experts Group (hence its name). Despite many attempts by companies to claim ownership, it has effectively become an open standard, and therefore pervasive. Anyone looking to develop an imaging device or service uses JPEG. Social imaging applications like Instagram, lifelogging devices like Memoto and walled gardens like Facebook know that to use JPEG is to have a head-start: no need for plug-ins or learning curves. Like VHS and MP3, JPEG has lock-in as a standard. By adopting it, we can get on with the business of selling or using gadgets and services.
JPEG compresses image data... very efficiently. Let's take the example of Obama's photographer. When she pressed the software button, light hit the camera sensor, an array of silicon, solar or photovoltaic cells. Some of the light's energy was absorbed by the silicon, knocking electrons loose which were forced to flow in a particular direction creating a current: photons became electrons, light became electricity. In order for the software (including JPEG and Facebook) to be able to work with it, an analog-to-digital converter (ADC) in the camera converted this electricity to digital information a RAW data file.
JPEG’s job was to get rid of as much data as it could while still making the image recognizable. It sampled the frequencies of different colors in the data captured by the sensor, attempting to determine how much they contributed to the visible image. Then it applied a series of algorithms (Discrete Cosine Transform, quantization, and Huffman coding) that discarded “extra” data. JPEG depends on the human eye's forgiveness. It fills in the gaps of the data it had removed (this is why JPEGs sometimes look blocky or smeared). The camera software compressed light-as-electricity-as-data into a visible file that the human eye could see as the First Family. That visibly powerful result, the encoded photograph, was written to the memory card ready to be uploaded and then decoded by the JPEG software built into photo management software, web browsers and surveillance systems.
Even Google, with its hegemonic power and massive reach, has not been able to dislodge the simple JPEG algorithm. In 2010 the company devised a more efficient compression algorithm and file format, WebP (pronounced “weppy”). Images encoded through WebP are visible in Google's own browser (Chrome), email system (Gmail), photo-management software (Picasa Web Albums) and search engine, but not in other leading browsers or photo management systems such as Firefox, Safari, Internet Explorer, Aperture or Lightroom. If you try and upload a photograph encoded with WebP to Facebook, the social network's software will scratch its virtual head, grey out the file name and fail to acknowledge the data. Literally, WebP does not compute.
As the world's most popular photo site, social imaging is a core part of Facebook's business. The company makes money from the Open Graph that draws the connections between users, their “friends,” and their content - connections that are mined and sold to advertisers. Images are a crucial part of that. Every time we upload an image of us hugging our wife and someone likes it, another data point is added to Facebook's map of relationships: Paul likes Obama. A new connection or data relationship is established. We can think of Facebook as a “relationship engine” generating new connections and content as images are uploaded, liked, shared and tagged. Images are particularly powerful connectors because they can be tagged with additional information, such as the date and location an image was taken and the people captured in the image. The tags establish connections which create even more data.
The importance of images in general and JPEG in particular can be seen in Facebook’s patents. Tagging digital media (US patent 7,945,653) cites “the association between the digital image and the email address [which] may be stored in the media database.” Managing information about relationships in a social network via a social timeline (US patent 7,725,492) states: “the social timeline further comprises photos of the members connected in relationship.” Facebook's lawyers as well as business people know that the image and the user are connected in Facebook’s database, ready for further association by subsequent use or for mining by advertisers and spooks.