Tag Archives: Korsakow

Writing, Zeno, Pitch

I am writing an essay about interactive documentary. Actually Korsakow. It is very very late. It is remarkably recalcitrant. I am so very nearly there, that moment when the finish line becomes a demonstration of Zeno’s paradox as it seems to move further away the closer you get. It is so close. Here’s a snippet.

My elevator pitch would be that Korsakow is software for authoring generative, associative, and processual films. These films are complex, possibly autopoetic systems that rely on patterns of relation to emerge for author and users.

Really Real

A common scenario witnessed this week in classes with media students. They add stuff to a new project in Korsakow, add keywords, do lots of building, and then go to save their project. The program doesn’t like that. I’ve written an explanation for them that most (the majority) won’t read, let alone understand. The problem here is thinking that the digital is supposed to be friction free (I blame Apple of iOS for this), that it is immaterial and, well, should be designed and made so that I don’t have to know anything about it, or what I’m doing. Digital natives my arse. Just because you can drive a car doesn’t mean your a mechanic, and just because you can use seven programs at once and know all the best shortcuts in Final Cut/Premiere/Photoshop/gMail you’re not digitally native. You need to think like the machine (like a good mechanic), and you can’t do that if you haven’t got a clue about the machine. It’s a thing. It does stuff. Real thing. Real stuff. Not virtual. Not immaterial. (Which is the cardinal sin that the NFBs big arse ‘capturing reality‘ has made, but that’s for another day.)

labs, this week

Launch Korsakow. “Ok…”, drag some video in.
“Mmmm, Ok, that wasn’t so bad.”
Double click. “SNU editor? Whatever.”
Make some changes, add some keywords, link to an interface, and a thumbnail (“What’s the drama with this?, this is easy”). Repeat.
“Cool. Oh, better save the work.” File, Save.

“Shit.”

It’s all gone. “F$#kn rubbish piece of software.”

If you don’t want this experience then open Korsakow, and SAVE THE PROJECT BEFORE you add media, and you’ll be good. (Subject to good housekeeping.)

If you want to know why, continue reading (you are asking the software to do something impossible).

When you drag or insert media into your Korsakow project you are not putting the video ‘in to’ Korsakow. Korsakow just remembers an address of where your video is and so when you do an export it uses this address to go and find the video, and then transcode it when you export (publish) your film. This address is called an alias, and Korsakow relies on an alias, which is in effect a link to the real file, to find it.

Lots of programs use aliases as a way to have multiple copies of a file without actually having multiple copies of the file (incredibly useful when working with video), and it’s built into the Finder of OS X (File – Make alias, aka Command L). That’s pretty straight forward. Korsakow does this so that it doesn’t have to swallow video, which introduces all sorts of storage and data problems which would make the program slower (since it has to process all that video) and less agile (as keeping the media outside of the program makes it much easier to change the media as you go).

How do aliases work? By addresses and paths. All the files on your computer, much like everything on the Web, has a file path, which is its address. On a modern Mac this begins with / and then is a series of folder names and eventually a file, so a word doc in my user account on my computer might have the path /Users/amiles/Documents/somewriting.docx (That example isn’t strictly accurate but it will do.) This is its address.

Now we saw that Korsakow needs to know and remember the address of where your video is when you add it (which is why you don’t change the names of files, folders, or where they are on your computer after you add them to Korsakow unless you want export hell). I imagine it does this in the simplest way, which is to remember where these media files are in relation to where it is. On computers we call this a relative address. In other words if it knows it is here, and the video files are in a folder called ‘media’ next to it there, then Korsakow only needs to know an address that says ‘media’ to be able to find those videos. Now the program could me made to remember a full path, all way to the top of your computer, but then if you moved your project to another computer all these addresses and paths would break. By remembering where the media is relative to where the program file is makes it easy to move the project to other computers.

These relative addresses on a computer are just like personal directions you give someone in the street. When asked how to get from Carlton to Brunswick you don’t tell someone to first go to the Melbourne General Post Office, you say go up there, turn left, then right, etc – these are directions relative to where you currently are.)

Can you see the problem? If you haven’t yet saved your Korsakow project file (the file that ends in .krw that stores all these aliases) then how can it know where it is? And if it doesn’t know where it is how can it then find a way to where the video is? Now yes, just crashing is basically a bug. But it is happening because you’re asking a program to do something impossible – to think about the relative address to somewhere else without first telling it where it is. The fault here is with us, not it.

Korsakow and Dense Nodes

In network media I discussed small world networks, dense nodes, and so on. A korsakow film is exactly this sort of structure. Below is something I wrote in 2011, reposting here as it should help people to understand Korsakow films as an architecture and structure that you do things with, and an architecture and structure that is, in its very DNA, the same as the networks we are working on. It’s sort of a mise-en-abyme moment really.

A significant idea that has a lot of relevance for things like the internet, hypertext, and social media (which are all forms of distributed networks) is the idea of a ‘small world network’. This is related to the famous experiment by Stanley Milgram about there being a maximum of ‘six degrees of separation’ between any two people, anywhere in the world. A small world network assumes lots of a small number of connections between individuals (nodes, clips in a k-film, links on the web, people you know), but with a few individuals who have a lot of connections. In relation to social networks these links are not about how close you are to others (whether geographically or personally) just that you know them. The existence of only a small number of people who know a lot of other people (who have a lot of connections) makes it much easier to get from one group to another, from one individual to another. The key features here are that these connections (how many people you know) is not equally distributed – I know 100, you know 200 – and that to get from one individual to another you do not need to know all the connections, all you need to know is somebody that you think will be closer than you are.

So, what does this have to do with k-films? Quite a lot, since keywords create (in k-film land) small world networks. Clusters or clouds of clips that all know about each other since they have common keywords. Now, imagine a work which has several such clouds. This is like a party where there are four, no let’s make it five, distinct groups of people who know each other. Now, to find someone in my group who knows someone in another group (in other words someone who could easily sit in one or more of the groups) is quite easy and this is how the two groups can be connected. The person I know in that group over there can introduce me to everyone else in their group – I just need one point of connection to be able to join them, it doesn’t matter that I only know one person.

Hence in my k-film with my clouds all I need to do is make sure I have one clip (node, SNU, pick your term) that has lots of connections to the other clouds. To keep my now rather dodgy analogy going, the person who knows someone in three, four or even five of the groups at my party. This node might have no limit to lives (it will keep appearing) and also have plenty of keywords so itn is a point of connection to all the other clouds and nodes.

In practice I might have most clips with limited lives. As I view the work I am caught in a cloud, but as I view material and clips ‘die’ this special node (what I’m currently calling a dense node) will appear. If I select this then because it has links to all the other clouds I can now get access to these other clouds. In this way I’m able to make sure that all the parts of my k-film can be connected. That’s one half of the problem. The other half is to figure out how to film or make this content in such a way that it makes sense, visually and contextuallly, so that it works as this dense node or hinge between these other parts. This depends very much on what these other clouds are about.

Need an example? I might have material that I have grouped (made as clouds) around night and day. I might then have to fllm something at dusk or sunset and use that as something that connects night and day and make this clip my connector. I might have inside and outside, light and dark, blue and red, and so on. In each case once I recognise what the terms of my structure are I can identify something that falls between them, and this is the one that I can use to join these two clusters or clouds together.

Monogamous Media and Korsakow

Korsakow is a system that encourages promiscuous connections between video clips. This doesn’t mean lots of lots of connections (though it might), but it does mean that this clip might hook up with that clip, and that one, and that one. And, not or. A way I characterised some of the confusion, and perhaps anxiety, about systems like Korsakow in a symposium today was around this idea of how promiscuous things become. In linear media it is all monogamy, or more accurately a literal serial monogamy, where that clip will always end at exactly 12 seconds, forever (at least until death does do us apart) and then this next clip will always and forever then appear. And play. As makers we don’t seem to want that order to be fucked with. (And lots of interactive documentary is as strident about these strict connections as heritage media with its single channel of constrained cause and effect always-just-this-way.)

My students, in spite of the nonsense of being ‘digital natives’ (now there’s a misnomer) remain highly invested in this notion of monogamous media (reified as the truly important story I have to make in just that way) and remain bemused, startled, threatened and at at times troubled by the looming promiscuity of connection they are being introduced to. We need to learn how to dance with these fragments of our and other’s media, letting them touch and slide and swoon and caress and bump and grind and abut. (Some students, however, are relishing it.)

Future time based media forms and practices are all about this dance, and while I’m not much of a dancer myself, these young people need to learn to get over their reserve.

A 2010 K-Film Explaining How to Make your K-Films

I’ve made a k-film that is very linear. Well, not linear, you can listen and view the clips in any order you like, but it is not so much a k-film as using Korsakow to collect some screen casts together where I go over some things to think about in relation to making your k-film projects. This was made using Korsakow 5.0.4, and I’ve included a playhead in the interface which shows you where in the clip you are. The advantage of this is that it immediately tells you how long a clip is, since if the playhead progresses very slowly then you know it is a long clip (and vice versa).

So, the k-film. It is commentary on some of the things to think about in your final projects. I made it as a k-film because it was just much faster to talk about this stuff and record it than trying to write it all out. YMMV. Click the pic go there.

Knotes