I teach a postgraduate class at Birkbeck, University of London which explores issues around journalism in an age of the Live Web. In the first session, as a way of the students getting to know each other, and for me to know them, I get the class to interview each other and publish a story. The idea is that they produce something and post it online, raising issues of what is ‘journalism’?, speed, mobility, informality, open and closed source, content relationships etc. As proof of concept, while they are working I take a photograph and post to Flickr and Twitter. I deliberately make the picture as bland and ‘unprofessional’ as possible so that the emphasis is on the issues not the image. This exercise is also an experiment with jpeg.
My iPhone takes the image and the camera software compresses it using the jpeg protocol, adds the suffix .jpg and saves the image file to the phone memory. Already protocol is ‘withdrawing from view’. Having done its work between the capturing and the saving, it retreats, hides, leaving only its trace behind – encapsulated in the suffix. But the jpeg protocol actant, in its network alliances has had real effects. Jpeg is the chosen compression protocol in the iPhone software. Apple’s designers could have chosen GIF, or PNG or even RAW as a way of encoding the digital information its CCD provides to the software. They didn’t. One reason is because of the high compression offered by jpeg. This allows more images to be taken more quickly (writing a large file to memory can be a slow process, forcing the user to wait). This can be a selling point for consumers. It also allows iPhone images to be flexible and interoperable, allowing iPhone app developers to create apps that manipulate or share the images using standards they can work with, building the sort of App ecosystem that Apple is now seeing as central to its business and locking-in the user to that ecosystem and arguably proprietary space. In effect Apple uses an open standard as part of a closed business strategy.
I used one of those closed Apps to upload my image to two social media spaces – Flickr and (via Mobypicture, the default image service with my Twitter app) Twitter. Again the jpeg protocol’s alliances not only enabled these practices but constructed them in particular ways. As a compressed image, my jpeg-encoded photograph, could easily pass through the narrow bandwidth available on mobile networks (itself the site of other protocols and actant-alliances, e.g. TCP/IP, DNS etc). As a jpeg-encoded image, Flickr could recognise and post it to my page immediately. If my image had been encoded using a different standard e.g. Tiff, Flickr would have had to decode and re-encode it using the jpeg protocol to make it work and visible in its databases. Like Apple, Yahoo’s Flickr engineers chose to use the jpeg-protocol to encode the images it archives and displays. To a certain extent they had no choice. As imaging devices standardised around the jpeg protocol and other sites and services defaulted to “upload your jpeg” (sic), Flickr needed to ensure its(!) images were interoperable and the barriers to entry for consumer/imagers was as low as possible, the site was as fast as possible and the tools available were as attractive as possible. This is where again the jpeg protocol established new and powerful alliances around its metadata. The jpeg protocol allows EXIF metadata to be added to the file during compression. Camera and image management software can write information about the moment of taking, the choices the imager made and the location. All of this can be read by Flickr’s software and made added value for the Flickr imager by being added to the database, the page and Flickr as well as the imager’s archive. It is important to separate the metadata that jpeg enables from that Flickr allows a user to add. As I upload, or later I can add other information to my image: title, description, tags etc. This information is added to the file’s entry within the Flickr database. But other information in that database has been provided through the jpeg/EXIF protocol, which during encoding adds metadata to the image file – not the database entry. This information is then read into the Flickr database entry.
Again, in the Flickr practices of uploading, publishing, searching, downloading and printing through a jpeg-aware desktop programme, we are seeing the traces of the jpeg protocol but not the protocol itself. The ‘Flickr image’ both on my page and within Flickr’s databases, searches and streams of imag(in)ings, the sharing, embedding and imagining practices that happen around those streams within the distributed scopic regime, have been enabled and set in motion by a protocol. Flickr’s business (and the ecosystem of apps, programmes, widgets, search engines, artworks and PhD theses that surround it) are in/enfolded with jpeg.
I also uploaded the jpeg-encoded file to Mobypicture. I didn’t choose to. Within my Twitter app I chose to add a picture to my 140 character ‘publication’. Because Twitter itself is only based around textual information, working only with text (and of course network) protocols, another ecosystem has developed around Twitter which allows/encourages users to share non-text information such as images and videos. My app, defaults to Mobypicture. One the reasons it does that (leaving aside any possible commercial tie-ups) is that the iPhone camera software, the Twitterlator app software and the Mobypicture software (as well as the browser software on phones and desktops) are jpeg-aware. They are built with that protocol in mind. They expect to receive jpeg-encoded information from other actants in the network.
The extra-textual information I added to the Tweet – my #birkbeckmedia tag, link to the class website, geolocation, comment etc wraps around the link to Mobypicture. The image (on another site/server), the website, the ecosystem around my tag are brought together within the 140 characters while at the same time remaining distributed objects. This complex in/enfolding/outfolding, distribution/concentration, maybe deterritorialisation/reterritorialisation or exteriority/interiority are the traces of protocol and standards: http, TCP/IP, DNS, XMPP, OAuth and of course jpeg. These protocols withdraw from view. In Heidegger’s terms we don’t see them until they break. In ANT terms, they’re black boxes, so everyday they become transparent like Serafina Pekkala, the witch in Philip Pullman’s His Dark Materials, not invisible, just unnoticed.
The practices of distributed imag(in)ing that make up the new scopic regime create distributed objects. The jpeg-encoded image that was on my phone is now on my phone, my computer, in my computer’s cache, on a Flickr server, a mobypicture server, in viewer’s browser caches and possibly their iPhoto library folders etc. It’s jpeg-encoded metadata is now in databases and search caches. Maybe to even talk of ‘in’ and ‘on’ is problematic. There is no single, simple image-object. Protocols have constructed that object as a distributed multi-object, constantly in motion at different scales, within different alliances. This, of course has profound impact in terms of ideas of intellectual property, privacy and ethics. Jpeg as protocol-actant-object is in/enfolded with those discourses, practices and institutions. But again the only way of approaching that in/enfolding is through the traces. Protocol itself, like Keyser Söze, just slips out of sight/site.
At one level the jpeg protocol’s work is done when it finishes encoding and compressing digital information from a CCD chip into an image. At another level, its work is just beginning. Given the right equipment and skills we could possible see the protocol working to encode, compress and create a file. It is possible to envisage a laboratory device (or software) that made that process visible, like a microscope showing the wonders of nature in slow motion. What that equipment or software could not show is that other work: the alliances jpeg sets in motion or supports, the business, cultural and political practices it is in/enfolded within, the regimes of truth it is implicated in.