On ‘A File Structure for the Complex’
July 19th 2008 10:53 pm
Nelson, Theodor H. “A File Structure for the Complex.” “The New Media Reader.” 2003. 134-145
While reading this article I kept having to remind myself that Nelson is not talking about the Executable and Linkable Format but his own format. About half way through the article I decided his Evolutionary List File should really be called the Extremely Limited File, or the Enigmatic Laborious File.Correct me if I’m wrong, but I got the impression that Nelson didn’t see too much farther than the immediate future and was all excited about giving historians something to pine over.
To me, it seems as though the file structure is quite limited to text, and that Nelson didn’t see into the future possibilities and potential the computer offers, and therefore ELF is ‘extremely limited’ as I put it. This is not to say it’s completely useless for other things, but it doesn’t take a genius to see it was designed with text and the sort in mind.
It also seems enigmatic and laborious as I put it. The historian isn’t going to care how the data is stored, only that it is. The supposedly easy to use structure of ELF is probably not going to be easy to use for most other people. Have you tried learning Lisp? I haven’t wrapped my head around it completely due to lack of time, but it’s definitely a programmer’s language. ELF seems like a low level Lisp, and unless things are explicitly identified as “manuscript” and “footnote”, historians and the like aren’t going to find it very easy.
I could be wrong however, but a few simple searches on Google and Wikipedia turn up nothing on “evolutionary list file”, so I gather nobody cares.
Nelson probably should have titled his essay “A Complex File Structure for Historians”
About the author... Daniel Huckstep is a software engineer(ing student) at the University of Alberta. He enjoys all aspects of computers and their software, reptiles, guitar and music creation, reading, and macaroni salad. 
Wesley Cooper
responded on 22 Jul 2008 at 10:15 am #
An interesting start to a critique. There is a passage near the end of Nelson’s article in which he contrasts what he’s doing with the more limited role of programming languages like Lisp. You may still be right, but in a fuller version you’d want to address this passage directly. Also he’s explicit that the files he has in mind should store music, images, etc.
darkhelmet
responded on 22 Jul 2008 at 11:00 am #
He says “It can be extended for all sorts of purposes, and implemented or incorporated in any programming language.”
Well, yes, that’s entirely true. You’d have to extend it for other purposes because unless you have documents and other text which CAN be organized in a list fashion nicely, it won’t work. You can also incorporate it into programming languages as either a data structure, or the premise on which the language is built, such as Lisp, where everything is basically a list or an atom (lists and entries in ELF).
As he also says, ”[e]volutionary file structures, and the ELF in particular, are designed
to be changed piecemeal by a human individual.” That’s the kind of limitation I am talking about. He didn’t seem to see past the initial ideas of computers, and with that statement makes it seem as though they would be suited to store data, but not manipulate it or process it. I don’t think anybody will deny that when you use a computer to do something, you tell the computer what you want to do (mouse clicks, keyboard, etc) and it does it (prints your document, open a web page). There is no human manipulation of data on the disks directly.