The Underweb
My dissertation (2007-2011), titled “Re-Framing the World Wide Web”, studies and describes how technical standards, protocols, and API's shape the aesthetic, functional, and affective nature of our most dominant mode of communication, the WWW. Furthermore, it examines the politically charged and contentious battle over the most important window on designed and de-signified information. As a remedy and way to catalyze the already non-linear development dynamic of this amorphous electronic infrastructure my dissertation proposes an alternative browser prototype and framework. I provide both written and practical portions, arguing a need for more user-oriented technologies that equally emphasize the ability to read, write, and publish in the internet without third-party involvement.
screenshot of the Underweb showing two video surfaces playing while editing their shapes to be non-rectangular
Underweb browser showing video encoding and streaming to an icecast server. The left image shows the streaming video. The right image shows a video player that is re-playing the live video from the server in an edited shape.
The written portion of the dissertation provides a critical analysis of the technological space of the WWW. I analyze the history of the WorldWideWeb as it has evolved in piecemeal fashion from a single ascii-based file format made for a single browser into a dynamic and interlinked software environment. I analyze the big changes that are on the horizon with HTML5 and how the new so-called standards tend towards a more centralized web experience, pushing more and more personal user data into private data clouds. I show that previous additions to the web infrastructure were all implemented in retrospect, and how a recent push towards standardization at the application layer of online communication may not be necessary or even practical considering the possibilities afforded by free and libre software development methods. Moreover, I consider what is lost in the simple, uniform, and effective communication format of HTML version 3, as well as the social and artistic significance of software methodologies that exist in a public sphere such as the WWW.
For the practical portion of this dissertation, I introduce a proof-of-concept browser-like media and communication application environment called the Underweb. Unlike the contemporary WWW, who's aesthetic and functional ideology is geared towards rectangular newspaper-like layout, single page-based interaction, and consumption of data, the Underweb aims to provide the user with more general layout mechanisms, dynamic interaction, and tools for writing and publishing of data. The Underweb can decode as well as encode audio-video streams on the net. It contains potential support for multiple markup languages, and includes an API for developing simple non-rectangular container shapes. It uses and exposes to the developer the lower level free software technologies that are employed, but concealed, by other contemporary browsers such as firefox, safarai and chrome.
screenshot of the Underweb showing two video surfaces playing while editing their shapes to be non-rectangular
Underweb browser in edit mode where users can directly edit shapes, texts, colors, and layouts inline.
These API's come from the free software libraries of glib, gtk+, cairo and pango. It includes direct support for reading and writing of files, albeit with no current security model or sandbox strategy. Underweb “pages” can be written in C, vala, javascript and python, but could potentially support other languages automatically through GOBject introspection. It also includes support for decoding many multimedia formats that are currently unsupported by standard browsers, as well as a simple software hook to internally embed other browser engines such as Safari's webkit. Furthermore, the Underweb browser also includes full socket support so that the browser is not only a client on the WWW, but also potentially the server.
pdf icon
August Black, Marko Peljhan. 2011. Underweb. Submitted to ACM WWW ‘12, but rejected
underweb logo, two letter in black and white spelling Un
three different shapes, red, green, and blue
Shapes in the Underweb are a collection of points (bezier or linear), a surface, and text. Surfaces may be defined by color, image, or procedure. Texts may be defined inline or loaded from a foreign server, and may be wrapped in a rectangular bounding box or within the bounds of the shape itself.
two images with sine patterns from yellow to green
The “surface” of a shape is only a chunk of memory. It may be written to as an image file or alternatively procedurally generated.
example of gui layout
The Underweb provides easy to use GUI widgets from the GTK+ toolkit.
example of gui layout with transparent background revealing desktop
Users may mix widgets and graphical components together. The background of the Underweb browser may also be transparent.
sinusoidal layout of repeating blue shapes on red
Shapes in the Underweb may provide their own drawing callback functions for animated displays.
2 web camera images of a person demonstrating video capabilities
Full audio video capabilities are included in the Underweb, including the ability to not only read audio and video, but encode and stream it as well. Above is a screen grab of an Underweb application document streaming live video to a server on the left and displaying the receiving stream on the right.
audio waveforms
The Underweb has a very simple built-in audio api. More complex signal processing can be done by writing puredata patches and controlling them vie the Underweb. The Underweb uses libpd to enable this. The above shows the graphical display of an audio buffer while playing.
bezier editing of GUI shapes
All shapes and media content are editable inline in the Undereweb.
website of UCSB's MAT program loaded in the underweb
Using the webkit and mozdev APIs, the Underweb can also view standard WWW content.