There were a few questions after my presentation on the details concerting serial and parallel processing in computation. Here is a pretty good link to a op-ed article by the chief scientist at NVIDIA: Life After Moore's Law By Bill Dally
Affective Relations with Computation: the problem of HCI:
There is currently a problem of transparency in the affective process of creation involving computed media. The process of computed media is generally obscured by perfect emulation of other forms of media. Lev Manovich’s critique of HCI (Human Computer Interaction) is merely a limited form of this issue.[i] The issue of HCI is not as Manovich states that it is bound to a cinematic mode but that HCI is a layer of mediation that obscures the procedural essence of computation. This layer of mediation draws on many different forms of media and epistemological organizations work to mask the procedural relation between artists and computation. In order for computed media and concepts to express and implicate truly new affects and concepts it will be necessary for those engaged to have a intimate and honest relation to the processes in which they explore.
An artist or thinker who wants to approach computed media on its own plane of immanence is to engage with the procedures that computation makes possible. This is not to say that the ability for an HCI to adequately emulate processes of previous media is bad. It is the mutability of computation that is its tremendously useful power. Yet for computed media to implicate something new, practitioners must engage with it on a much more intimate level.
This of course occurs in most experimental media in which the line between artist and engineer is largely extinct. There is however often a barrier to communicability of the experience of these practitioners and those who would be affected by their creations. It is fortunate that to experience a sculpture, painting, or even film can connect a viewer to the process by which the artist and medium are engaged. One can see and feel the artists brush strokes, chisel marks and imagine the directors splice at each cut. Computed media does not provide us with this kind of connection as long as it is used to emulate other media. For example, if digital-cinema is to critically approach computation it should impart to a high level of procedural transparency in which the viewer can be actively aware of the computed sequence that produces the images before their eyes.
[i] Manovich, Lev, "The Language of New Media," MIT Press, Cambridge, Massachusetts: 2001. pp. 88-93.
Here is a link to the full essay for those interested.
Showing posts with label Computation. Show all posts
Showing posts with label Computation. Show all posts
Friday, May 21, 2010
Thursday, May 13, 2010
From Algorithm to Process: Striation and Smoothness in Computation
Last night I talked about the main thrust of the essay I'm working on for the end of our class. It will primarily be a starting point in developing two terms: the algorithm-image and the process-image. I opened my talk last night with a quote from Cinema 2 in which Deleuze talks about how the expression of movement-images and time-images will not be developed any further with electronic media than by pre-existing cinematic forms. Instead electronic media will have to forge new affects and concepts in its own right.
"The new automatism is worthless in itself if it is not put to the service of a powerful, obscure, condensed will to art, aspiring to deploy itself through involuntary movements which none the less do not restrict it. An original will to art has already been defined by us in the change affecting the intelligible content of cinema itself: the substitution of the time-image for the movement-image. So that electronic images will have to be based on still another will to art, or on as yet unknown aspects of the time-image."
"What is important is that the cinematographic image was already achieving effects which were not like those of electronics, but which had autonomous anticipatory functions in the time-image as will to art. Thus Bresson's cinema has no need of computing or cybernetic machines; yet the 'model' is a modern psychological automaton, because it is defined in relation to the speech-act, and no longer, as before, by motor action (Bresson was constantly thinking about automatism)."
(Cinema 2, 266)
It is in this spirit that I wish to make an attempt at describing the qualities that are significant to artists and thinkers in their current and future engagement with computation in general and computed media in particular. Instead of focusing on the issues of subjective bodily experience as Hansen has argued, I would like to take the position that Delueze has proven to be most interested in, the artist-artwork relation. The terms algorithm-image and process-image will be developed from the examination of the affective and conceptual relations that artists and thinkers enter into in their engagement with computation. The rough distinctions between these two term are as follows:
Algorithm-Image: finite, closed and striated process.
A prime example which we talked about in class would be most of video game narrative. There are choices to be made yet these choices are limited and the final outcomes are predetermined and of a fairly small set.
Process-Image: open, able to implicate a continuous becoming.
There are not many examples of this. Yet, the fractal video I showed in class and Benjamin's example of data-moshing are images that implicate a specific process at work in computation. With and through these images a certain open-ended implication of a infinitely itterable procedure is affectively grasped. In fractal we are presented with a sequence of a single geometry repeated such that a infinite space can be easily imagined with each part necessarily engaged with and situated in a unique relation with the whole. In data-moshing we are shown a glimpse of the process by which movement images are in process. Striped of their unique frames, movement becomes a series of vectors. Movement as process and in process.
I will also examine what I find to be the true issues embedded in HCI. I believe that Lev Manovich's critique of HCI is merely a limited form of the essential issue. The issue of HCI is not that it is bound to a cinematic mode but that HCI is a layer of mediation that obscures the procedural essence of computation (Hansen, 34). The layer of metaphor that draws on many different forms of media and epistemological organizations work to mask the procedural relation between artists and computation. In order for computed media and concepts to express and implicate truly new affects and concepts it will be necessary for those engaged to have a intimate and honest relation to the processes in which they explore.
Tuesday, May 4, 2010
The Digital and Immanence Co-Implicated, or A Minor-Digital
In reading Hansen I've been returned back to the concept of the smooth and the striated being co-implicated. One of the most interesting aspects of this was illustrated by Deluze and Guattari in ATP when they said that all progress must take place within striated space while all becoming occurs in smooth space (ATP, 486). I talked about this a little in my presentation (part B). Immanence is implicated by striated technology that operates at such a scale and velocity that immanence is able to be implicated. The way in which striation is connected to technology is given an example in Mandelbrot's fractals (ATP, 486-488). These fractal images would not have been possible with out the increasingly iterable computational tools available to Mandelbrot. The systematic approach of fractal geometry is a good example according to D&G because it "provides a general determination for smooth space that takes into account its differences from and the relations to striated space" (ATP, 488).
The example is but one of many instances in which a striated system can function in a way that implicates an immanent concept. For me, I think that this is also the case in cinema. The technology of the movement-image is such that it accelerates the static technology of photography to such a velocity that not only is movement implicated, but time as well. Deleuze would argue that the implicated becoming in cinema is not movement but time. Hansen's search for what is "new" in new media holds a similar affinity to Deleuze's search for what is co-implicated in cinema (Hansen, 21). Although he takes his conclusions in a rather non-Deleuzian direction I believe his starting point is very important. What does the striation that produces the digital (image/media) co-implicate? What kind of immanent concepts can the digital imply? Hansen seems to focus on the way in which the digital striates the status of representation in media and therefore shifts the origin of meaning and interpretation to the body.
If we put Hansen's conclusions aside how would we provide a "general determination [of] smooth space that takes into account the differences from and the relations to striated space"? (ATP, 488) A single equation can be repeated enough to imply a conception of natural forms in the case of fractals (ie. the geometry of leaves and mountains). A single frozen slice of action/time can be taken again and again a such a rate such that it implies movement and time in the cinematic image. What does the digital striate? How are these qualities expressed in the 'major" media of digital? How are they different and related to the immanent qualities that it quantifies? If there are indeed immanent potentialities within digital technology would a "minor" digital be possible? What would minor digital be like?
For this reason I'm not sure that the reason that data-moshing as Benjamin posted for us implicates movement or time but something quite different (see Movement and the Digital Image). Just as the cinematic image implicates movement and time because of its unique relation to time so then would minor digital implicate immanence due to it's unique relation to smooth space. Data-moshing perhaps asks questions about the digital itself. It foregrounds what Lev Manovich describes as the digital image as only being the surface of pure data (Hansen, 32).
As you may of noticed, I'm not really sure what the digital truly striates. My initial instinct is that the digital striates process and procedure. This is what Manovich calls the processural image (Hansen, 9). Maybe what the digital striates is then a certain cognitive logic and the thing that a minor digital can implicate is a continuum of thought.
-Ariel
The example is but one of many instances in which a striated system can function in a way that implicates an immanent concept. For me, I think that this is also the case in cinema. The technology of the movement-image is such that it accelerates the static technology of photography to such a velocity that not only is movement implicated, but time as well. Deleuze would argue that the implicated becoming in cinema is not movement but time. Hansen's search for what is "new" in new media holds a similar affinity to Deleuze's search for what is co-implicated in cinema (Hansen, 21). Although he takes his conclusions in a rather non-Deleuzian direction I believe his starting point is very important. What does the striation that produces the digital (image/media) co-implicate? What kind of immanent concepts can the digital imply? Hansen seems to focus on the way in which the digital striates the status of representation in media and therefore shifts the origin of meaning and interpretation to the body.
If we put Hansen's conclusions aside how would we provide a "general determination [of] smooth space that takes into account the differences from and the relations to striated space"? (ATP, 488) A single equation can be repeated enough to imply a conception of natural forms in the case of fractals (ie. the geometry of leaves and mountains). A single frozen slice of action/time can be taken again and again a such a rate such that it implies movement and time in the cinematic image. What does the digital striate? How are these qualities expressed in the 'major" media of digital? How are they different and related to the immanent qualities that it quantifies? If there are indeed immanent potentialities within digital technology would a "minor" digital be possible? What would minor digital be like?
For this reason I'm not sure that the reason that data-moshing as Benjamin posted for us implicates movement or time but something quite different (see Movement and the Digital Image). Just as the cinematic image implicates movement and time because of its unique relation to time so then would minor digital implicate immanence due to it's unique relation to smooth space. Data-moshing perhaps asks questions about the digital itself. It foregrounds what Lev Manovich describes as the digital image as only being the surface of pure data (Hansen, 32).
As you may of noticed, I'm not really sure what the digital truly striates. My initial instinct is that the digital striates process and procedure. This is what Manovich calls the processural image (Hansen, 9). Maybe what the digital striates is then a certain cognitive logic and the thing that a minor digital can implicate is a continuum of thought.
-Ariel
Friday, April 30, 2010
Smooth Hypertext
I just wanted to share a few blurbs with you which I came across recently and I thought they might be related to the concept of "smooth space" as Delezue and Guattari described it.
The first blurb is a work by Christian Hubert, which some of you might know already. This theory is in the form a hypertext and "It contains extensive paraphrases of readings, generally, but not always, clearly credited [...] Many come from readings in contemporary developments in science like chaos and complex systems studies, some theoretical biology, and some related philosophical writings, particularly the work of Deleuze and Guattari."
http://christianhubert.com/writings/index.htm
The second blurb is a background sketch on the history of the hypertext. I wanted to mention a text by Critical Art Ensemble to open up the field. It is the chapter 5: "Utopian Plagiarism, Hypertextuality, and Electronic Cultural Production" from the book The Electronic Disturbance. Here is an extract:
"Thinking about a new means for recombining information has always been on 20th-century minds, although this search has been left to a few until recently. In 1945 Vannevar Bush, a former science advisor to Franklin D. Roosevelt, proposed a new way of organizing information in an Atlantic Monthly article. At that time, computer technology was in its earliest stages of development and its full potential was not really understood. Bush, however, had the foresight to imagine a device he called the Memex. In his view it would be based around storage of information on microfilm, integrated with some means to allow the user to select and display any section at will, thus enabling one to move freely among previously unrelated increments of information.
At the time, Bush’s Memex could not be built, but as computer technology evolved, his idea eventually gained practicality. Around 1960 Theodor Nelson made this realization when he began studying computer programming in college:
'Over a period of months, I came to realize that, although programmers structured their data hierarchically, they didn’t have to. I began to see the computer as the ideal place for making interconnections among things accessible to people. I realized that writing did not have to be sequential and that not only would tomorrow’s books and magazines be on [cathode ray terminal] screens, they could all tie to one another in every direction. At once I began working on a program (written in 7090 assembler language) to carry out these ideas.'
Nelson’s idea, which he called hypertext, failed to attract any supporters at first, although by 1968 its usefulness became obvious to some in the government and in defense industries. A prototype of hypertext was developed by another computer innovator, Douglas Englebart, who is often credited with many breakthroughs in the use of computers (such as the development of the Macintosh interface, Windows). Englebart’s system, called Augment, was applied to organizing the government’s research network, ARPAnet, and was also used by McDonnell Douglas, the defense contractor, to aid technical work groups in coordinating projects such as aircraft design:
'All communications are automatically added to the Augment information base and linked, when appropriate, to other documents. An engineer could, for example, use Augment to write and deliver electronically a work plan to others in the work group. The other members could then review the document and have their comments linked to the original, eventually creating a “group memory” of the decisions made. Augment’s powerful linking features allow users to find even old information quickly, without getting lost or being overwhelmed by detail.'
Computer technology continued to be refined, and eventually— as with so many other technological breakthroughs in this country—once it had been thoroughly exploited by military and intelligence agencies, the technology was released for commercial exploitation. Of course, the development of microcomputers and consumer-grade technology for personal computers led immediately to the need for software which would help one cope with the exponential increase in information, especially textual information. Probably the first humanistic application of hypertext was in the field of education. Currently, hypertext and hypermedia (which adds graphic images to the network of features which can be interconnected) continue to be fixtures in instructional design and educational technology. An interesting experiment in this regard was instigated in 1975 by Robert Scholes and Andries Van Dam at Brown University. Scholes, a professor of English, was contacted by Van Dam, a professor of computer science, who wanted to know if there were any courses in the humanities that might benefit from using what at the time was called a text-editing system (now known as a word processor) with hypertext capabilities built in. Scholes and two teaching assistants, who formed a research group, were particularly impressed by one aspect of hypertext. Using this program would make it possible to peruse in a nonlinear fashion all the interrelated materials in a text. A hypertext is thus best seen as a web of interconnected materials. This description suggested that there is a definite parallel between the conception of culture- text and that of hypertext:
'One of the most important facets of literature (and one which also leads to difficulties in interpretation) is its reflexive nature. Individual poems constantly develop their meanings—often through such means as direct allusion or the reworking of traditional motifs and conventions, at other times through subtler means, such as genre development and expansion or biographical reference—by referring to that total body of poetic material of which the particular poems comprise a small segment.'"
The first blurb is a work by Christian Hubert, which some of you might know already. This theory is in the form a hypertext and "It contains extensive paraphrases of readings, generally, but not always, clearly credited [...] Many come from readings in contemporary developments in science like chaos and complex systems studies, some theoretical biology, and some related philosophical writings, particularly the work of Deleuze and Guattari."
http://christianhubert.com/writings/index.htm
The second blurb is a background sketch on the history of the hypertext. I wanted to mention a text by Critical Art Ensemble to open up the field. It is the chapter 5: "Utopian Plagiarism, Hypertextuality, and Electronic Cultural Production" from the book The Electronic Disturbance. Here is an extract:
"Thinking about a new means for recombining information has always been on 20th-century minds, although this search has been left to a few until recently. In 1945 Vannevar Bush, a former science advisor to Franklin D. Roosevelt, proposed a new way of organizing information in an Atlantic Monthly article. At that time, computer technology was in its earliest stages of development and its full potential was not really understood. Bush, however, had the foresight to imagine a device he called the Memex. In his view it would be based around storage of information on microfilm, integrated with some means to allow the user to select and display any section at will, thus enabling one to move freely among previously unrelated increments of information.
At the time, Bush’s Memex could not be built, but as computer technology evolved, his idea eventually gained practicality. Around 1960 Theodor Nelson made this realization when he began studying computer programming in college:
'Over a period of months, I came to realize that, although programmers structured their data hierarchically, they didn’t have to. I began to see the computer as the ideal place for making interconnections among things accessible to people. I realized that writing did not have to be sequential and that not only would tomorrow’s books and magazines be on [cathode ray terminal] screens, they could all tie to one another in every direction. At once I began working on a program (written in 7090 assembler language) to carry out these ideas.'
Nelson’s idea, which he called hypertext, failed to attract any supporters at first, although by 1968 its usefulness became obvious to some in the government and in defense industries. A prototype of hypertext was developed by another computer innovator, Douglas Englebart, who is often credited with many breakthroughs in the use of computers (such as the development of the Macintosh interface, Windows). Englebart’s system, called Augment, was applied to organizing the government’s research network, ARPAnet, and was also used by McDonnell Douglas, the defense contractor, to aid technical work groups in coordinating projects such as aircraft design:
'All communications are automatically added to the Augment information base and linked, when appropriate, to other documents. An engineer could, for example, use Augment to write and deliver electronically a work plan to others in the work group. The other members could then review the document and have their comments linked to the original, eventually creating a “group memory” of the decisions made. Augment’s powerful linking features allow users to find even old information quickly, without getting lost or being overwhelmed by detail.'
Computer technology continued to be refined, and eventually— as with so many other technological breakthroughs in this country—once it had been thoroughly exploited by military and intelligence agencies, the technology was released for commercial exploitation. Of course, the development of microcomputers and consumer-grade technology for personal computers led immediately to the need for software which would help one cope with the exponential increase in information, especially textual information. Probably the first humanistic application of hypertext was in the field of education. Currently, hypertext and hypermedia (which adds graphic images to the network of features which can be interconnected) continue to be fixtures in instructional design and educational technology. An interesting experiment in this regard was instigated in 1975 by Robert Scholes and Andries Van Dam at Brown University. Scholes, a professor of English, was contacted by Van Dam, a professor of computer science, who wanted to know if there were any courses in the humanities that might benefit from using what at the time was called a text-editing system (now known as a word processor) with hypertext capabilities built in. Scholes and two teaching assistants, who formed a research group, were particularly impressed by one aspect of hypertext. Using this program would make it possible to peruse in a nonlinear fashion all the interrelated materials in a text. A hypertext is thus best seen as a web of interconnected materials. This description suggested that there is a definite parallel between the conception of culture- text and that of hypertext:
'One of the most important facets of literature (and one which also leads to difficulties in interpretation) is its reflexive nature. Individual poems constantly develop their meanings—often through such means as direct allusion or the reworking of traditional motifs and conventions, at other times through subtler means, such as genre development and expansion or biographical reference—by referring to that total body of poetic material of which the particular poems comprise a small segment.'"
I will send the PDF of Critical Art Ensemble around for those who are interested to read further. Related to the above there is also a graphic/map which I have found on the Documenta 12 website through which they tried to visualize the concept behind the 'Documenta 12 Magazines' program. The website and description: http://www.documenta12.de/magazine.html?&L=1
Sunday, April 25, 2010
Movement and the Digital Image
As an initial reaction to a technical detail of Mark Hansen’s argument, I’d like to talk about movement and pixilation in the digital image. On page 9 of the introduction to New Philosophy for New Media Hanson writes:
“If the digital image is an accumulation of such discontinuous fragments, [pixels] each of which can be addressed independently of the whole, there is no longer anything materially linking the content of the image with its frame, understood in its Bergsonist-Deleuzean function as a cut into the flux of the real. Rather the image becomes a merely contingent configuration of numerical values that can be subjected to ‘molecular’ modification, that lacks any motivated relation to any image-to-follow.”
According to my, admittedly limited, knowledge of video files this is, strictly speaking, untrue. A compressed video file is made up of both the change in color pixel-by-pixel, frame-by-frame (called the I-frame) and the movement within those frames through time (called the P-frame). This is what allows for the technique called datamoshing in which I-frames are deleted but P-frames are not, creating a decidedly unique effect. This technique has been used by artists such as Takeshi Murata:
And Paperrad:
How this retention of movement-in-time effects Hanson’s argument theoretically I’m unsure. In Deleuzean terms, the strictly striated space of pixel grid is smoothed over by the P-frame’s access to movement which identifies larger forms on the screen and transgresses the borders of individual pixels. For a how-to video on datamoshing and more information about the process you can watch this video:
-Benjamin Schultz-Figueroa
Thursday, April 8, 2010
Computation and Immanence Part 1: Rhizomatic Computation
I realized after my presentation yesterday that I could have done a much better job making some of my questions more clear. In fact I regret not having paused before opening up the discussion into the realm of cinema. The thoughts I am developing in that area are murky to say the least. I'll try and explain where I'm coming from a bit more in part 3 of this subject. That being said, I'd like to recap the initial part of my presentation and begin to allude to why the questions raised are of interest to me.
Here is the link to the presentation site I was using.
I think the statement that best sums up the issues for parts A through C is this:
The rhizomatic potential of computational space is suspect.
This suspicion is expressed by Deleuze and Guattari in the introductory chapter in ATP (a thousand plateaus). From my point of view, contemporary computation and the tools derived from it are structured such that a relational arrangement that is purely rhizomatic (as well as completely immanent) is not achievable. There are two areas in this claim that should be clarified; computational space and rhizomatic relation.
1)Computational Spaces are media and mediated environments which are predicated on modern computational tools such as the ubiquitous cpu and the directory(folder) based operating system. This necessarily includes spaces as large as the global internet to small local spaces of the desktop computer running locally executed software. The common features that I am interested in are the way in which relations are organized and the methods by which relations and information are transformed. Information and connectivity are organized into hierarchical folders or grid like databases. These pieces are transformed through Boolean logic statements sequenced together to form more complex procedures. It should be clear that computational space is a delimited space in which only certain kinds of information, relation and becoming are allowed.
2) Rhizomatic Relation in its most pure sense should not be achievable within a computational space. For such relation to occur information, relation and becoming must not privilege any hierarchical structure. In rhizomatic computation there would be no folder analogy, no grid of data, no predefined logic of transformation, and no predefined order of operations let alone a general dismissal of true simultaneity. This, in general, is not the nature of the computational space we utilize today.
Finally we come to the danger of "false multiplicities" as mentioned in ATP a couple paragraphs following D&G's initial suspicion of the information sciences. The internet as we experience it today sometimes seems like it holds the potential for a rhizomatic space. The scale at which web sites and individuals are connected is staggering. The velocity at which these connections are made, broken and remade is amazing. It is at the limit of individual perception that the potential for false multiplicities present themselves. I would not go so far as to say that the rhizomatic potential of computational space is not such that immanence cannot be inferred by our experience. However, it is difficult at the scale and velocity of our current global network to see a qualitative difference between space that is purely rhizomatic or merely a hierarchy so vast that it is indistinguishable from the rhizome.
-Ariel
Here is the link to the presentation site I was using.
I think the statement that best sums up the issues for parts A through C is this:
The rhizomatic potential of computational space is suspect.
This suspicion is expressed by Deleuze and Guattari in the introductory chapter in ATP (a thousand plateaus). From my point of view, contemporary computation and the tools derived from it are structured such that a relational arrangement that is purely rhizomatic (as well as completely immanent) is not achievable. There are two areas in this claim that should be clarified; computational space and rhizomatic relation.
1)Computational Spaces are media and mediated environments which are predicated on modern computational tools such as the ubiquitous cpu and the directory(folder) based operating system. This necessarily includes spaces as large as the global internet to small local spaces of the desktop computer running locally executed software. The common features that I am interested in are the way in which relations are organized and the methods by which relations and information are transformed. Information and connectivity are organized into hierarchical folders or grid like databases. These pieces are transformed through Boolean logic statements sequenced together to form more complex procedures. It should be clear that computational space is a delimited space in which only certain kinds of information, relation and becoming are allowed.
2) Rhizomatic Relation in its most pure sense should not be achievable within a computational space. For such relation to occur information, relation and becoming must not privilege any hierarchical structure. In rhizomatic computation there would be no folder analogy, no grid of data, no predefined logic of transformation, and no predefined order of operations let alone a general dismissal of true simultaneity. This, in general, is not the nature of the computational space we utilize today.
Finally we come to the danger of "false multiplicities" as mentioned in ATP a couple paragraphs following D&G's initial suspicion of the information sciences. The internet as we experience it today sometimes seems like it holds the potential for a rhizomatic space. The scale at which web sites and individuals are connected is staggering. The velocity at which these connections are made, broken and remade is amazing. It is at the limit of individual perception that the potential for false multiplicities present themselves. I would not go so far as to say that the rhizomatic potential of computational space is not such that immanence cannot be inferred by our experience. However, it is difficult at the scale and velocity of our current global network to see a qualitative difference between space that is purely rhizomatic or merely a hierarchy so vast that it is indistinguishable from the rhizome.
-Ariel
Subscribe to:
Posts (Atom)
