Alan Sondheim Phenomenology of Linux Beyond the traditional division of graphic user interface (GUI) and textbased interface, the unix and linux system/s create a unique environment problematizing machine, boundary, surface, and structure. The environment has implications far beyond a kinesic study of a particular technology; these tend toward an (un)accountancy of splintering or sputtering, stuttered linkages, microsutures, scanning intention across or among traditionally “isolated” platforms. Begin with the apparent file structure: 1. Working within the files, there are several domains: the formal tree- organization of the operating system (beginning with the root and ascending/ descending); the accumulation or heap of files within the local directory (these files may or may not be related beyond their common path); and the imminent domain, the file or files currently open or in the process of being modified (these are nonexclusionary). 2. The graphic interface opens to shells as well, and since the interface devolves from a blank screen, there is simultaneously potential (click anywhere on it) and absence (nothing visible), reflecting upon the human operator / monitor interface as well. 3. Errors may or may not be characterized by error messages, which are inscribed by a process evolving from the root cause; there is then both the symptom (program x misbehaving) and the message (Error: ) that intersect: the message may be the (only visible) symptom, and the symptom itself may carry the message. 4. It is easy to assume that source code is equivalent to bones and operable binaries to flesh; or the kernel as fundament, and file structure as slough. I would rather argue for a system of cubist plateaus of intersecting information regimes, with vectors/commands operating among them. In this sense it is information that is immanent within the operating system, not any particular plateau-architecture. 5. Language moves among performative, declarative, and neutral /dev/nul regimes; again, the boundaries are blurred, even on a technical level. Programs, more properly scripts (an apt word, since code is inscribed) call up different languages, shells, other programs, internal or external conduits (see below); internal and external interpenetrate here. 6. The division between GUI and text-based net access is blurred; shell accounts use IP and can open X Window and browsers, just as browser GUIs can share window space with shells. 7. The space of the operating system is problematized since machines carve out what I call fractal channeling, ports and commands rapidly shuttling back and forth between traditionally external netspace and internal vehicle space. Channels may open to other shells which may open to other channels; loopback channels operate within the local vehicle (internally), for example, and may be used to communicate with incoming on a local talk application. In shell-to- shell, both are equivalent on the screen: think of this as screen-resonance or system of strange attractors. 8. Furthermore, within the screen-resonance there are the spaces of the user/s on the system, partly application-dependent, shuttling among persons, tenses, and semantico-grammatical categories (Whorfian, in other words). Two linked talkers may be opened in relation to a net browser on an X Window while top (a program monitoring machine processes) is also running, and files are being transferred from a cdrom to hard drive. Attention moves among these spaces/applications, blurring distinctions; the talkers, for example, may demand considerable psychological investment, while anomalies in one or more of the other applications also call for immediate examination and response. If errors etc. appear, the anomalies (in relation to the normative ongoing chat) may best be described phenomenologically by Schutz’s relevance theory, consider lifeworld strata, projects, and presentifications —in spite of the fact that all of this is primarily read and written to, inscribed and counterinscribed. 9. One might argue that the fractured domain in its entirety is never grasped—nor is there a “domain” and “entirety” at all. If we extend inscription and counterinscription, taking into account fuzzy and fractal channeling (deconstruction of category object/arrow theory), we can work toward a loosely defined sememe undergoing continuous and fairly rapid transformations, which are not necessarily charted from either interior or exterior (meta-) positions. The traditional metapsychology of the user splits, just as it splits beneath the sign of morphing gender in MOOs and IRC; it is always already possible for theory to take morphing into account (as if morphing is being-accounted-for and therefore accountable), but this is a posteriori; in fact the splitting problematizes any metapsychology insofar as the mind is considered a somewhat closed (hydraulic model) frame, as opposed to a fuzzy communicative systemics paralleling the description herein of the operating system itself. 10. It is not difficult to see, not the operating system as mind, but both mind and operating system as challenging dyadic conventions of interior/exterior, grammatical tense and person, and so on. As I have mentioned before, Merlin Donald takes steps in this direction; one can also consider an accumulation or sememe of flows moving among bodies, organs, and so on, along the lines of Deleuze and Guattari. 11. Within and without all of this, the cyborg model, based on the suturing of disparate epistemes, becomes oddly antiquated; it accounts well for prosthetics, robotics, and machine/organism navigation, but remains based on traditionally separate ontological domains. Instead, think of spread epistemes and ontologies—for example, the distinction between declarative and performative becomes oddly confused in the case of basic HTML coding (that is. without “refresh” or JavaScript), which flows texts around screens. 12. Finally one might bring up postmodernisms, with their flows, partobjects, relativities, multiculturalisms, incommensurability of commensurable languages (and commensurability of incommensurable language)—as well the postmodern architectures, with their deconstructions, skewlines, and exposures/doublings, baring the systems, decomposing them. And it is true that such architectures have their equivalent among the operating system architectures; the operating system kernel for example may be equivalent to the control center of a building, and the communicative flow through a building has its equivalence with the fractal channeling described above. Nevertheless, I would not want to push this analogy, to the extent that the postmodern is representative of a stage (that is, post-Fordism among other things), and not necessarily the (de)construct of a broken episteme more or less permanently on the (broken) horizon and always- having-been-present. For the operating systems under consideration may be likened to the production of a scanning electron microscope, a case in which scanning is related to phenomenological intentionality instead of the discrete world of envisioned objects and flows described in, say, Gibson’s work. The difference, yet to be accounted for, never to be accounted for, lies between the optical circularity of the phenomenology of the image produced by the light microscope, and the exaggerated dimensionality and exploratory scanning of any electron microscope, such as the tunneling or even the recent development of the scanning probe, which promises to “image single electrons,” one might almost say, bits and their own architectures down to that very level (Scientific American July 1997.)