The Virtual Lightbox: The Potential of Peer-to-Peer Humanities Computing

This paper will demonstrate and discuss the Virtual Lightbox, a newly developed image-based software tool (programmed primarily in Java) in the context of the network paradigm known as peer-to-peer computing. We believe that the Lightbox, which may be imagined as an image-based whiteboard for the Web, will be of broad interest to members of the international humanities computing and digital library community.

We are also particularly interested to discuss the Lightbox in terms of peer-to-peer computing, the best-known exemplar of which is probably Napster (the controversial MP3 trading application). Online projects in humanities computing have heretofore depended primarily on the client-server architecture underlying the World-Wide Web. As a number of commentators have recently pointed out, the popular emergence of peer-to-peer applications may represent the first major paradigm shift in network computing since the 1991 launch of the NCSA’s Mosaic browser (Tennant 2000); we will therefore present the Virtual Lightbox in the context of these developments and their implications for humanities computing.

The Virtual Lightbox is a newly developed open source software application, supported by the Center for Computational Sciences at the University of Kentucky and jointly designed and implemented by this paper’s authors. The Lightbox was conceived to function in a networked environment, as a browser helper application. It presents users with a display surface analogous to a real-world lightbox or tabletop; images may be imported into the display area through manual selection (from both local storage and remote http addresses), or (with suitable server side scripting) as the result of a search. Once placed on the surface of the Lightbox, images may be dragged, arranged, and layered at will to create juxtapositions for comparison. Images may also be enlarged and reduced in their proportions, from thumbnails to full size. At present, there is no comparable network-based tool that facilitates image selection and comparison. Comparison itself is an example of what John Unsworth has recently termed a “scholarly primitive.” Thus, the Lightbox is founded on the assumption that computational tools should reflect intellectual practice.

As participants in a recent discussion on the LOOKSEE (image-based humanities computing) listserv pointed out, image-based tools for the desktop are plentiful (Photoshop, for example, or the University of Maryland’s PhotoFinder); however, image-based tools for networked environments are comparatively scarce (but see the Institute for Advanced Technology in the Humanities’ Inote, or the new tools that Luna Imaging is producing). In designing the Lightbox, we therefore sought to combine the need for an adequate image-based comparison tool with the collaborative potential embodied by networked environments. Participants in a Lightbox session may share images (including their actual on-screen configuration) with one another. Each user’s actions with the Lightbox are emulated in real-time for the other participants in the session. Thus, if one user slides an image across the surface of the Lightbox, all other participants in the session observe this behavior. Likewise, a suite of basic image processing features (sharpen, brighten, opacity, etc.) is available for collaborative use. (These features reinforce our analogy to the numerous text-based whiteboards that are now available.) Authority control is obviously important here, and users of the Lightbox can opt for different layers of authority to avoid initiating contradictory or exclusionary actions.

A typical usage scenario might resemble the following: Four scholars (each at a different
institution) who share an interest in Blake’s use of color schedule a day and time for a Lightbox session. At the appointed hour, each loads 6-8 images from their hard drive into the Lightbox for inspection by their distant colleagues. Each takes a turn at “directing” the virtual conversation, bringing his or her images to the fore, zooming them to highlight a detail, and offering commentary via a realtime chat facility that is part of the Lightbox. (Or perhaps they might agree to place a conference call instead.) The observing scholars may each individually annotate the images belonging to the current discussion leader, keying their notes to specific regions of interest on the image. At one point, one of the observing scholars wishes to bring an image into the viewing area and juxtapose it with an image belonging to the conversation’s current director (but without assuming authority over the conversation). When finished, the four scholars each save a copy of the chat session’s transcript for future reference.

Question: Could the tasks just described be performed asynchronously? Yes, probably. But most scholars will attest to the importance of dialogue and interaction with their colleagues, and robust tools for realtime distributed collaboration will enable those intellectual exchanges to flourish in fruitful and unexpected ways.

We also believe that the Lightbox has obvious pedagogical implications, particularly in the arena of distance learning. We think the Lightbox could potentially find an audience in a field as far removed as medical informatics.

In order to accomplish the above behavior, we have had to draw on some fairly advanced tools and protocols for building network architectures. These are bound to become more prevalent in humanities computing, and the Lightbox provides a useful example of how these network protocols have been implemented in a humanities context. In particular, we make use of the Java Shared Data Toolkit (JSDT) and Remote Method Invocation (RMI).

The JSDT is an API provided by Sun Microsystems. It provides a platform to support highly interactive, collaborative applications. The toolkit provides full duplex multipoint communication between connected applications. It also provides a common programming interface for multiparty communications. The wide variety of technologies that have been implemented below the interface -- such as T.12x (standards based multiparty communications protocol) and TCP/IP -- are not visible to the programmer. The API has taken ideas extensively from ITU T:122 recommendation for Multipoint Communication Service for Audiographics and Audiovisual Conferencing Service definition. RMI, meanwhile, allows parameters, return values, and exceptions to be serialized and sent from one Java Virtual machine to another across the network. For example, the movements of an image across the surface of the Lightbox are serialized as a set of x/y coordinates which are then distributed to participants in a session (thus eliminating the need to transmit the image across the network in its entirety). The fusion of RMI and JSDT in the Virtual Lightbox, makes it possible to efficiently send and receive both the image and control data over the network using a peer-to-peer model.

In this paper we will address the Virtual Lightbox not only from a technical standpoint, but also reflect on its potential significance in relation to the kinds of electronic scholarly resources we currently create in the humanities. We’re thinking in particular of the rise of the “archive” as the predominant trope in scholarly electronic editing. Many of the major scholarly electronic projects in the humanities -- the Blake Archive, the Rossetti Archive, the Dickinson Electronic Archives -- use that word conspicuously in their titles. One of the reasons why “archive” has been so naturalized in discussions of electronic editing is because it is clearly compatible with the basic assumptions of client-server computing: that there is a centralized repository (the archive) which houses material for users to retrieve on demand. The interactive nature of peer-to-peer computing, however, promises to inspire new kinds of electronic resources, whose locus is outside the archive. In particular, networked spaces may come to resemble laboratories; that is, venues for synchronous collaboration rather than remote collections of static resources. The Virtual Lightbox, for example, in conjunction with streaming video and (especially) emerging Internet 2 technologies (such as the NCSA’s Access Grid), would allow the Blake Archive to function as a site for sophisticated scholarly conferencing.

References

Inote and the ImageSizer
http://www.iath.virginia.edu/inote/

LOOKSEE: Resources for Image-Based Humanities Computing
http://www.rch.uky.edu/~mgk/looksee/

Luna
http://www.luna-imaging.com/

PhotoFinder
http://www.cs.umd.edu/hcil/photolib/

Tennant, Roy. “Peer-to-Peer Networking.” Current Cites 11.10 (October 2000). http://sunsite.berkeley.edu/CurrentCites/2000/cc00.11.10.html

Unsworth, John. “Scholarly Primitives: what methods do humanities researchers have in common, and how might our tools reflect this?”
http://jefferson.village.virginia.edu/~jmu2m/Kings.5-00/primitives.html

The Virtual Lightbox: An Image-Based Whiteboard for the Web
http://www.rch.uky.edu/~mgk/lightbox/