 |
|
There are many ways a group (2 to N) can collaborate on writing projects. There are many models for that, here we examine one as part of our mission to document and contextualize literary curiosities. The following starts with a cut and paste from the Autobiography of Pain website dedicated to collaborative science-fiction writing, and then follows with comments about random collaboration.
Welcome to The Autobiography of Pain.
We are striving to be the world's first full length wikinovel. This novel belongs to everyone. The world is free to change, edit, or delete anything it wants. If you don't like the story or the characters, you can change it.
= How does this work?
There will be stages for each chapter, a specific amount of time allotted for the creation of each. Once the chapter has reached the end of its countdown, that chapter will be put into the vault and cannot be changed. You will be able to view the completed chapters in the vault, though. You can contribute by either using your creative brain to add to the content of the chapter/book or even use your grammar and editing skills to proofread and make corrections to the current story. We will give the credit where the credit is due.
=Isn't this going to get out of hand?
Probably. But that is what makes it fun. Sure, someone might come along and completely change the story. But that is what makes this better. No one knows how the story is going to go. It could be a comedy, a thriller, romance, or a complete mixture of everything.
Oh yeah. And if you don't like the title we have, tell us a new one. This is everyone's work, not just ours.
http://www.theautobiographyofpain.com/about/
== Excerpt:
"I woke up to the stale taste of vodka and cigarettes in my mouth. There was a miniature mining crew hacking its way through my skull. The deafening silence has returned. After a lifetime of abuse--mostly self-inflicted--I can't decide whether I'd rather live or die. Apathy.
Resigning myself to face the day (because it happens to be easier than planning death), I took a swig from the bottle of vodka I had stashed under the bed. My robe is stuck under Cherry. I forgot I hired her last night. Some people have no manners; bitch didn't have the decency to leave before I woke up. "What were you raised in a barn," I muttered. I nudged her and pointed to a roll of bills on the dresser. She swore (classy), tripped as she put on her fishnets and knocked the telephone off the nightstand on her way out."
=Comments about such project
This is certainly a way to write quickly, as well as to proof text. Note this uses a classical novel framework, it's still a 16 chapters novel, all that is original is how up to 100 people type in something in it. The general issue with growing the writer base for a single piece is that to support the collaboration you have to lock the current version to a single writer at once to maintain sense. So 1000 people could not really write a particular novel faster then 100 ones. In fact 100 people writing a chapter is surely counter productive without a few actually directing the process and the others commenting rather then actually writing parts of it directly.
The other value is such text type is not conversational. The end result is not email threads about something. The collaboration is a standard document that one person could have writen. What this is also not is a never ending branching hypertext novel, a set of literary tracks that splits out of a main text. In the role-playing tradition that would be based on a novela each contributor would expand an aspect, role-play the story from the point-of-view of one character for example...
Collaboration and this Book
The purpose is not encyclopedic or academic (in the sense of producing a textbook on something) but more like creating a knowledge slice so people can visualize a new literary genre. As the amount of content grows the objective is that everything naturally best fit to that slice more and more naturally. At some point each information presented can be turned to the question "how does that fit with the rest here?". After a critical mass then what this is becomes more and more clear and then ad-hoc collaboration much easier to manage. That is at some point the collection becomes cohesive as a collection that it becomes easy for anyone to see something and say this relates to this Book and how, and if you read the Book even where in it.
It's irrelevant if the source of a particular tidbit that fits in came via an email, or arrive magically via typing a set of keywords in a search engine or was found in another book. In the most traditional sense what I read collaborates with me by existing and being related to.
As discussed elsewhere, this is still a personal contribution, therefore I have read somehow all that I reference. Let's say I started to work on this project 30 years ago and only 100 logical Pages are in publishable form, how will I get to 1 000 000 ones in the next 20 years? That's 10 000 to 1 ratio somehow done. I guess then one would say that this is when/where and why collaborative methods become necessary. In order to achieve the discussed magnitude at some point, when a subset is sufficiently defined, large scale literature requires deployment over more then one stream.
Could this scale to a genre (large scale literature?). I think so, a bit like webmd.com might be a location for trustable health information, this could be a trustable place for intellectual exploration, somewhere between art and science - using a different method. The type of Book Collaboration that works here probably share some aspects with some so-called social network models such as Facebook's secret society hand shaking model in that collaboration is intentional (approved).
10 000 monkey typists
10 000 monkeys typing all their lives might generate one useful sentence. This to say, some approaches to grow the book by a factor of 10 000 so it delivers on it's title so that technical delivery could be achieved:
1: simply random character generation
http://en.wikipedia.org/wiki/File:Monkey-typing.jpg
2: a variation of random generation which would use the current text as input and this could further be constrained by some automatic grammar correction software. What we could achieve like that is something that visually looks real and would sustain real quick eye scans of any pages.
3: from there the next level of delivery could be achieved by scrapping the internet archives. We would be using a set of keywords that must be present in a document (eg 4 of the following 200) and some other statistics, probably seek for structured documents. We would define even more explicitely a formatting style and go... let the analysis run a few months over zillions of documents and come out with 1 000 000 short articles to bind together. The main difference with 2. is that now each article would be coherent. We might like that generate global statistics that match the writing style however we will certainly have deviate from the idea of a particular point of view about something and the overall relevance as a collection might not be much higher then a random search on any large collection of content.
How many articles have been published?
The next level in terms of feasability would be to collect all articles ever published that are not copyrighted anymore or are public domain. If you restrict yourself to read just titles and perhaps some outlines then in one year one could arrive at a pretty good prefiltering. If one worked on it 12 hours a day, 347 days a year and spent a minute per article, then one would review 1 000 000 articles in 4 years. Assuming the articles are prefiltered via search methods and such then this technique seem feasible with the proper toolset.
A number of questions: I presume all these articles are to be in english or french (perhaps italian and spanish) otherwise I am much like a monkey, how could I parse out a russian or chinese article? That is there might be 200 000 magazines (and if you extend to different internet only publications certainly 1 000 000,are also published in newspaper and in books ) yet not all of them I can minimally parse. If we limited ourselves to the english language (also to avoid translation duplication), the question woud be how many articles + book chapters have ever been writen. The value of the question is to tell us that our pre-filtering should eliminate 99 999 of every 10 000 non fiction article or book chapter ever written (more or less?). To rough it say there is 30 000 000 non-fiction books ever written in english (I made it up) each with an average of 10 chapters, and there has been the same amount of original articles then we have a total space of 600 000 000 items.
I might over estimate the amount of non-fiction books here (maybe the ratio is 3 fiction for 2 non-fiction for sales but unclear in terms of pure title count), there is certainly over 60 000 books published each year in the US alone and 10 000 magazines / publications that publishes more then one article a year. A non-fiction book might have 12 chapters in average. It seems 1 000 000 000 items (book chapter, magazine article, internet page on something...) could be our initial estimate. 1 000 000 is 1:1000 of that so is sort of the density of Dewey. Except we are resampling knowledge along completely different models.
