In the course of the dialog, the following exchange took place between ALAN SONDHEIM and JOHN CAYLEY, forwarded onto the nettime listserve by rr. I've reset the exchange, indented to show prior text and subsequent text referring to it, with any requoted text excised. Please see the nettime list archive for February 2004 for the context and further posts.

A few notes on codework for an upcoming talk -

These are great points, Alan, I am touched, and look forward to talking to you about this more. I get the clean/dirty distinction and accept that, despite my (un)conscious desire to de/inf(l)ect the categories and structures of so-called traditional/canonical literary practice, I do work with/against a notion of clean, functional literary and programmable-literary objects. (Not so sure about this, on reflection later, when you consider that I do a fair amount of *ucking around with generated transitional texts and make points about transcultural language issues. The work still 'looks' pretty clean tho.)

A difficulty I have with codework discussions - codework is usually taken as pseudo-code or broken or dysfunctional code somehow pasted onto or into a work - as if that would give the work credence, legitimize the text as literature etc. But nothing is farther from the truth. (Mez for example isn't broken code but displacement/insertion techniques - shifters, slippages, of language that are semantically marked.)

In general -

please note it is not the code that is broken - but the interiorities of more or less traditional semantic worlds - sememes as well (in other words, distortions of world-making).

That is a great way to put it. That when so-called natural language is, for example, 'mezangelled,' then "the interiorities of more or less traditional semantic words" are "broken." The code that breaks into the words is not necessarily itself broken. (Although let's face it, if it doesn't get recontextualized pronto, it ain't going to compile.)

But why code as opposed to other (chiefly linguistic) matter - from other registers of language or other practices of natural language use (other tongues)?

Because code is often substructure or protocol or generative. As substructure, it underlies, for example, the very distribution of the piece. As protocol, it may underlie the very typing or production of the piece. As generative, it has produced or partially prodced the piece.

I think the question of why code? remains. You are just saying, because it's there. What you say is also true of the procedures, the 'dirty' structures in Joyce. (Btw, is Joyce dirty? generally speaking, in your sense - abject dysfunctional - I think it's a real question.) Total and global or other-tongue syntax is dysoperative at any time in the prodcuct ion and gen aspir ation to this piece, any event writing, cleaning, even. (Very easily accessible without recourse to code-code.)

No, I'm saying not because it's there (it's always there) but that it's presence in one form or another in the work problematizes language, structure, and body; it also questions production itself (producing or programming code). It's a writing which in part litearlly in-forms itself; as such it also problematizes translation since to some extent it translates only into itself.

I am asking, why these media for these projects? Is it because no matter how dirty and abject and dysfunctional you get with codework, you stay clean? When does any of this start to matter?

I don't think so at all, although on the surface of course. It's clean on the surface because all ascii is clean, well-defined, as is the digital domain, with its reliance on potential wells of code/noise itself. But in terms of interpretation, the discomfiture and abjection can be strong. I'm thinking for example of the pieces Kim McGlynn and I did using ytalk (Perth to NY) a few years ago.

The digital always possesses this 'turn' of cleanliness - to the extent that 0/1 permits no undesired noise and is infinitely reproducible - and filth - to the extent that the media literally become dirty, outmoded, useless, and no longer run. So there's eternity on one hand and tremendous fragility on the other.

The point is not to valorise or downgrade (aesthetically/socially/politically) some text because it addresses or incorporates code, or subgrade it because the constituent code is broken or operative. Brokenness need not ally with value in any neat way. [See Sandy Baldwin in the current Cybertext Yearbook]...

Here is the issue; your 'addressing' or 'incorporating' (btw I've read Baldwin and was just down in Morgantown) implies a separation which for me - like the activities in the _chora_ are both problematized and inseparable.

As far as brokenness is concerned, I am working out of Winograd and Flores (re: Heidegger) here.

The questions are more to do with: what are the properties and methods of code as such, and how do (and how could) these contribute to language art making? What are the specificities of code that will allow us to derive textual objects with distinct characteristics? Or allow us to extend the Class Text and/or better understand its underlying abstract Class?

But whose questions, John? These are yours. When you say "more to do with" - this is your approach, not mine. When you cay "Class Text" again you're operating with the notion of "clean code" ("specificities") which may well not be the case. Look at Kenji's work or nn's writing.

So yes, these are my questions, and I am messy-dirtily wedded to a relatively clean project, but I do think these questions may help me answer why such-and-such media for such-and-such project.

Of course I have no disagreement with you here.

I think the world is dirty, people are dirty, in the sense of Mary Douglas or Kristeva (Powers of Horror); code in a formal sense is a defense against that. So I'm interested in the interplay of world and code, which is to say in that liminal area between consciousness and formal systems - which takes into account desire, sexuality, stumbling about, etc. And I'm fascinated by plasma, which at least on a theoretical level, can efface all codes, not even leaving debris behind.

(I want to point out also btw - in relation to the Cybertext book - that Jim R's work is quite clean, but he's not the only practitioner; I was doing codework in 71 and later wrote a number of programs in 76-78. Some of these are now in the internet text. And I was _late_ - Fernbach- Florsheim was doing things in the 60s with computers/code. Etc. etc.)

Hey, we wasn't trying to do a hard-core history of programmatology in the Cybertext Yearbook and I do know that even you were late. Rosenberg deserves the priority he deserves, as we all do.

what is code in the first place ? - obviously it can either be a PROGRAM which PRODUCES a residue - or it can be a CARRIER of meaning. this depends on the semiotic encodings as well as the PERFORMATIVITIES at work. a code may or may perform - in the sense that it may or may not create a result that one might characterize ontologically and/or epistemologically as AN OTHER.

And because a very prominent feature of code is its operation, the "program that produces a residue" focuses (for me anyway) critical attention. We are more familiar , in this context, with "carriers of meaning," however slippery and shifty. re(ad)Joyce! (I know you always already have.)

Yes I have, but the carriers in this case are structured or dirty structures - very different. You're coming at this through both literature and a 'clean' notion of code (see above); I'm not. For example, Perl poetry is operable, but the residue is pretty much irrelevant - yet as far as I'm concerned that's a terrific use of code. As is the figlet program. I don't distinguish - which is why my list in the first place is highly inclusive, not exclusive.

What could be cleaner than Perl poetry? The (vr-valentine) Card with a Perl Endearing.

Perl poetry is definitely clean and restrained - my point is that the poetic content is secondary to its running. It's by necessity a tour de force.

I've never been sure why Joyce comes up in these discussions; it's hardly code or hyper etc. etc. Surely there are other examples? There's always the pseudo-14th-century writing of Chatterton, for example, which did have roots in dictionaries, translations. And for strict coding, there's early television (late 19th-century) with its scanning and spark-gap images...

again - think of code - in the sense of morse code - as a _mapping_ between two or more strata. think of code - in the sense of programming - as a _performative_ between one stratum (that of base more or less sequentially accessed) and another (that of residue or results which may or may not appear dynamic or interactive). (see Eco, Theory of Semiotics.)

the POLITICAL economy of a text, performative, or performance in relation to the OTHER that appears.

CODE / PSEUDO-CODE ETC. IN CODEWORK

0 Texts which problematize the relationships among language, subjectivity, symbol, meaning, body, organism - texts which take nothing for granted.

1 Texts which are literally performative.

2 Texts which are program output. Here we might distinguish between texts which are sutured in relation to the _standard text_ - and texts whose content and form clearly reflects both programming and program _interference._

3 Texts which in part present the programs or controls that have created the remainder - for example a text which might include k:1 banner as part of the content - an ascii banner.

4 Texts which are clearly modified in their entirety - for example vowel substitutions, formations using awk, grep, sed, tr, etc.

5 Texts which present a _history_ or series of operations - texts which appear to require reconstitution or recuperation - these may or may not be accompanied by their _history of transformations._

6 Likewise texts which are incapable of such reconstitution or recuperation.

7 Texts with content 'picked up' and placed against other content (for example spam text interspersed with other text) in order to expand or render a sememe problematic.

8 Texts which rely on and relay through distributivities such as SMS, email lists, IRC, etc., as a means of dispersion, engagement, intervention, political or other action, etc. And texts which report back on these texts.

9 Texts working from MOOs, MOO language, MUDs, and so forth - texts generated in chatrooms with one, some, or all characters under the control of the 'author' or 'author-collaborators.'

10 Texts which are reportage from commands, such as traceroute or dig or @lastlog or Kismet output - texts which examine the network and its tethering.

11 Traditional texts, traditionally-written texts which report on other texts, other codeworks, which mimic codeworks, and so forth.

Code runs and conceals itself. Code that runs generates text over durations. Code that runs guarantees that language art cannot bracket its time-based dimension. It plays and plays out precisely and particularly in the 'Not to mention ..."

Code doesn't necessarily conceal itself. Code doesn't necessarily do anything you say it does. I wouldn't use the phrase 'language art' myself - I think prions are also code, DNA is also code and code is not necessarily language. I'd have to go back over my Eco for this.

Materially, my point is that code does do exactly what I say. You cannot read the code that is running as it runs. It was not coded for you, it was coded for the system (for an alien or underlying culture). The code may display some manifestation of itself - some instantiation of its own archive - as it runs, but it cannot display the code structures that are running. This point has a bearing on the dirt question since, for me, it introduces a possible site for something like an unconscious in the otherwise hypertransparent arena of codework and net.art. I'm no where near as well read as you in the necessary literature but would be interested to know what you thought of 'Inner Workings' http://www.dichtung-digital.org/2003/3-cayley.htm

I'll go to the URL... The running is always clean, even if there's a core dump. One might say it runs as it runs. But not all codework has code in it or is running; I wanted to make that point in the typology. The work presented may well be the residue of running code, or as in Mez, something else entirely, a diacritical structure imposed on 'english.'

I wouldn't say unconscious here, at least not necessarily, since the unconscious is also the site of the repressed, i.e. the 'dirty,' and the code runs clean and linear (more or less). (In that sense every program, like every dream, is successful, no matter what; it does what it does.)

Code takes literal time to run and as such it takes time to produce a readable display not to mention that it can defer reading, withhold it, structure and cultivate the time of reading. And yes I accept with no qualification that

All art is time-based btw.

but all art is read within institutions, and literary, even poetic (for-bog's-sake) institutions currently operate with a dominant notion of writing as deferral, as atemporal (to an extraordinary degree) for critical purposes. For me, writing in networked and programmable media challenges this in a direct and very clear/n way. Although, hopefully, things will get messy.

Here I agree with you; one thing about institutions, however - I think codework (and the name tends towards inclusivity as I said somewhere) at least so far has avoided such; if anything, for example on Poetics, it's somewhat of a gadfly or irritant - it refuses to go away, refuses to acknowledge that it's _not_ poetry or for that matter that it _is_ poetry - in a sense it's got the same sort of empathetic relation to language one might find in the Upanishads...

Some of the codework I do takes advantage of lag (in email or quicktime .mov), and some doesn't and some appears instantaneous ...

I feel a real difference between us is that I am writing from the position of dirty code, world-code, which may or may not operate, and that may or may not be the point. For example one piece I did involved reversing all the < and > on a specially written webpage. The result is chaotic, dys- functional in many ways, amazingly functional in others. At West Virginia, I re-morphed/mapped motion capiture sensors, transforming the body into a signal or searchlight system (see my heap.mov at http://www.asondheim.org ). And so forth.

It seems to me you're interested primarily in clean and generating concealed code - I have no problem with that. But I do feel you dismiss (even the word 'pseudo-code' is dismissive) everything else that's going on; since you're an editor and critic in the field, it's problematic for me.

I do want to be clear that I'm not, in any of this, trying to work out some way to dismiss any of what you do. I've got no problem and great respect for your performances and interventions. If I've used 'pseudo-code' it's not meant to be dismissive, simply to signal that the code in question would not operate.

A sample of morse code doesn't operate either. It's readable, and the code-forms, for example, Mez' are readable as well.

Pseudo implies a relationship to truth - i.e. a 'pseudo-intellectual' as opposed to a 'real intellectual.' It's a questionable term, I think, although I'm not sure what else might be used - something like quasi-code, or vary-code.

If I have any genuine critical - as in valorising - point to make it is in relation to work that is presented as more or less as composed writing in whatever media that simply incorporates code elements. Such writing gets no credit for doing this per se. To be interesting it would have to do more: e.g. get down and dirty; try to be as clever and interesting and involved as Joyce; or, even better, show us something about the properties and methods of code and/or language, more than the simple fact that you can kludge them together in the same word, sentence, paragraph, world.

Well here's the heart of the matter. I don't think the people I respect (this is a strange term here) do such kludging - that's what I was reacting against. I don't think I do, or Mez, or solipsis, or noemata, or nn, or jodi, or yourself for that matter. And much of this work isn't text based at all - Kenji and nn and myself have all done cdroms related to our other work, I've done video, some people have used flash. I think of codework as multimedia. But kludging would be, at least to me, fairly boring (although kludging within the same world is interesting).

12 Not to mention all those aleatoric texts, stochastic or chaotic texts or imagery, multi-media codeworks, generative works, generative works fed into themselves (resonance-work), specialized editors which refuse the WYSIWYG...

I feel a bit of an interloper here - not a nettime or anylist subscriber - so sorry to take up space and time in this forum. I'm afraid that normally the way I live (zen hermit up big-ass mountain) doesn't let me into such exchanges. Forgive me if I logoff now, with many thanks to Alan for his ever-incisive dirty (I mean that in his good way) words.

Hardly an interloper!

and thank you as well John - I think the discuss helps me tremendously (and others, if this arrives on nettime) - it's long overdue at my end -