Recent: 5 hours
Cumulative: 173 hours
September 9, 2009
I have so far been unsuccessful in finding a mentor, but thanks to the hours I collected this summer at APL, I still have time. The first portfolio evaluation is next Friday, but I haven’t learned the new system yet. There’s a meeting tomorrow at 2:30 to which I plan to go. I received a response from Mr. Batson a few weeks ago (I’m not sure if I’ve journaled since then). It was fairly short, saying that he was too busy for even the minimum – three upcoming book contracts, school, etc. I still haven’t figured out a good alternative. I’ve begun compiling a list of names: people I’d consider or who might know someone else. Firstly, I need someone with a name. I was interested in this project because I thought it would be interesting to look at story construction and maybe even try developing something myself. I’ve been thinking about it for a long time, even when I was doing science, but couldn’t consider wasting my time that way. I needed a tangible result, but now I think I have a plan that has a probable chance for success. I’ve never done anything like this before, but I’m glad I finally have an excuse to study it.
Experimental Design Ideas
General area of interest = art: story design and construction.
I want to test the usefulness of the template. The template may be applied to two main scenarios. The first is it will be used by editors to quickly and objectively identify errors with applicants’ works. The second is it will be used by authors to concurrently develop a proportional story that can then be developed into a novel.
I can test the usefulness of the template by evaluating its impact in these two potential applications. A good experiment requires many trials, which will be possible in the first case, but not the second.
How can I test that the template really works. The template offers a mathematical formula that is the minimum of the “ideal story”. My problem is that there are two variables. First, not every book is good, so the subject changes. If I had an established formula, I could compare subjects against each other using the scale of the formula. The problem is, that I’m trying to test the validity of the formula. I cannot test both the subject and the formula at the same time.
Maybe I need to do some research on how beauty is defined by science. In Science News I think, there was an article about an algorithm that identified desirable features in male and female faces. How did they identify beautiful faces? Human input. Many people were asked to somehow identify beauty and ugliness in the faces they saw. The algorithm found common features, for example, large eyes, tended to be more frequently seen as beautiful rather than smaller eyes. Thousands of interpretations of hundreds of faces were evaluated to remove outliers.
Another thing that must be considered is the background of the human taking the test. When political polls are taken before an election, it is important that analysts choose subjects that match the proportions of the population. The greater the number of subjects used, the more likely the result will be estimated correctly. In those kinds of demographic studies, the variables considered are: race, income, occupation, trend of spending, age, gender, state, stated political affiliation, previous voting record, etc.
For me, specialized characteristics that must be checked are their favorite genres, morals, characters, lengths, writing styles, age group, etc. This is getting complicated, and I'm going to have to be careful to make sure everything is controlled.
Recent: 12 hours
Cumulative: 185 hours
September 14, 2009
I talked to Dr. Amy Chai yesterday about my project to see if she was first a good fit for me, and second might be interested in being my mentor. I need to articulate my project in terms of the experimental design. I need to define what I want to test and how I will measure them. This weekend I came across a study in which human beauty of the face was evaluated and imitated by a computer. I will go into detail about what I thought of that particular source in its annotation.
Dr. Chai recommended that I assess the subjective quality of books based on the recorded statistics of national sales. I'm looking into that possibility right now, however my main concern is that advertising and name recognition will probably account for many sales rather than the actual quality of the product. I will probably need to assess duration: timelessness of the work as well. It just seems like the records simply do not exist.
Variables I will study:
Existence/Lacking of Story Posts. (Unit = Y/N)
Correct/Ill-proportioned Story Pillars (% error)
- [(Actual timing or word count)/(Closest Ideal in Range to Actual)] * 100 = % error
- Each post will be identified and given a numerical range of space within the work measured in either time or word count.
- Things that must be taken into account are duration and position in the larger story. Thus percent error will be evaluated for both duration and position. These two measurements will then be combined into a compound reading.
- These percent error measurements will indicate the importance of each aspect
Under Emphasis of Story Pillars
The story template will recommend that a number (at least 16) elements be included; this is easy to measure using a simple Boolean evaluation. The proportion of posts is slightly more complicated. I must account for both the duration of the post and its placement relative to either the endpoints of the larger story or relative to the other story posts.
Recent: 10 hours
Cumulative: 195 hours
September 24, 2009
I’ve been trying to narrow down my project in terms of identifying the specific variables I want to test. I’ve only had limited success so far. Again, I’m encountering the disorganization of art and subjective opinion. My overall premise for the project is to determine the usefulness of the Story Template in diagnosing and correcting errors in story. I hypothesize that The Story Template successfully combines sensitivity and specificity into the most accurate method available.
Sensitivity will be evaluated based on the number of correctly identified ‘errors’ versus missed ‘errors’. Specificity will be evaluated according to how many times The Story Template algorithm incorrectly identified ‘errors’ as ‘errors’. Accuracy will then be calculated as a composite of those two assessments.
Error Does Not Exist
Algorithm Identifies Error
Algorithm Does Not Identify Error
Both ‘a’ and ‘d’ are correct. Both ‘b’ and ‘c’ are incorrect. Sensitivity is a measure of ‘a’ versus ‘c. Specificity is a measure of ‘b versus ‘d.
Sensitivity = 100 [a / (a + c)]
Specificity = 100 [d / (b + d)]
Accuracy = 100 [(a + d) / (a +d) + (b +c)]
Librarial Research Findings
- The differentiation between “classics” and “non-classics” is dependent on a subjective scale of common reader preference.
- The most prominent three bookstores in the United States are Amazon.com, Barnes and Noble, and Borders.
- Theoretically, the collective list of top book sales from these three retailers could indicate modern “classics” if the definition of “classic” is dependent on the number of book sales within a predefined time increment.
- However, lists of sales were unavailable even through purchase. There were other book lists indicating what the retailer deemed to be most popular, but there were no sources to support these rankings.
- Many items included on these popularity lists seem to be less recognizable echoes of other recent successes or because of celebrity endorsements.
Recent: 17 hours
Cumulative: 202 hours
October 12, 2009
Unfortunately, I haven't had as much time to concentrate on my project lately. I've been thinking about it a lot, but haven't written down as much as I should have. In the last few days, I set up my portfolio, decided on a definition of a gold standard (control group), the final list of variables to be tested, posted a preliminary experimental design and research portfolio, etc. Most recently, I've shifted back to the second component of my research: aka: the novel. I've finally pulled myself away from a brief planning chart, though it's still not complete, to write out a first draft synopsis. I really like certain sections of events so far, but others need a little more work. I have general ideas about what will happen, but I need to make sure everything is logical.
I haven't written much at all about my progress in terms of this aspect of the project that I'd want to post here. Some of my discoveries though:
- Stay away from man vs. self conflicts. They can be done, but for someone as inexperienced as I am, it's more prudent to stick with a more prominently defined antagonist. It's too easy to get this confused with the hidden need character strand. After many pages of free-writing, I came up with a perfect antagonist!
- Make the protagonist at the center. Especially in the first and second act, I've been developing a very interesting character arc, but it felt like the protagonist didn't have enough of a direct role in the plot. I'm still working this out, but have an idea for a parallel series of events to echo the primary plot points.
- In the first part of the 3rd act (of a 4 act), there is a sequence of events dealing predominately with the hidden need character strand. I coined the term "hidden need triplet" which has been approved and will be used in the final Story Template product.
- "Anything can happen - as long as it's introduced from the beginning". I've been establishing the physics of my story world, and I've almost escaped the fanatic fantastic stage.
- "Do not introduce any characters in the slide!!!" Still working on this part, because I need certain characters to appear by the end, but chronologically, they cannot exist or appear until a certain number of things have happened. This necessary sequence does not conclude unfortunately until the slide. I'm stuck right now. I'll need to either change my resolution or rehash the story world.
- I have a brilliant idea for "the crazy plan" section!!! It took me absolutely forever to figure it out, but it fits so well; I'm glad I waited. I still need to work out the darkest moment and help from outside. I have a few ideas, but they seem somewhat contrived. I'll keep thinking.
November 4, 2009
This week is the end of the quarter. Consequently, I've been going down a checklist to make sure I've gotten all the deadlines instead of using my time for actual research. I'm not sure which is harder: doing the experiment or writing about it. I feel like I should be talking to my mentor more, but at the same time, I like the independence. Like the last two years, my experiment is my own, and though I've consulted Dr. Chai more than I did Mr. Steve Walker (advisor for last two years), she is not as involved as the traditional use of the title "mentor" might suggest.
I like what I'm doing this year; story structure is interesting I think - it's abstract. the only problem is that everything is just so subjective. On the plus side, there's a lot of freedom, but on the other hand, I'm afraid no one will take this seriously. I believe that the Template is a good product, enough so that I'd take the time to prove it. I believe my thesis. I'm doing brand new research, and I have a probable chance at getting it published. Why shouldn't people take it seriously? Art is unstable, but for high school, I can afford to try it.
The closest comparison of the Template to another task is the sketching of a face. There are set proportions of every face, though every face does vary. The eyes are an eye's width apart. Eyes are in the vertical middle of the face. The bottom of the ears should come to the nose, etc., etc. Likewise, there are certain common elements of story that may be referenced as a general rule, though quality of the story, like a face, also depends on the creativity of its features as well.
Story can be divided into four equal acts. In each act, an equal range for the number of chords. Then a known range for the number of scenes within each chord is also defined in the Template. The Template then describes the twists and turns of a story, sort of like the Hero's Journey does, but not quite so rigidly. I hate that I can explain what I know using specifics! It's impossible to understand my experiment without first understanding the premises of the Story Template, but unfortunately I can't denote that information, because it is still proprietary at this point.
There are pillars of a story - four abstract components that must be attempted to distribute creativity evenly. The best stories exceed in all aspects, but they prioritize, deriving genius from only one or two of the four pillars. Then there are character strands and subplots . . .
In the context of my research here, I started data collection a week or so ago. I've decided that to make the process more efficient, I'm going to access copies of each subject's screenplay instead of summarizing and analyzing myself. I still need to do the timestamps by hand, because those aren't available in the incriments I need, but that's alright. As I expected, each subject requires a long time for analysis, even though I'm just sticking to movies. I hope I can get enough data.
There's another seminar tomorrow, and we'll be talking about thesis statements and how to write a paper. Right now, I could easily write an essay on my reasoning behind my thesis, but I still don't have the data to form a conclusion. I really do need to get to work; I keep getting further and further behind, but I'm getting tired. I expected to be more than half way through data collection by now, but I'm not even close. Thanksgiving break is next week, so I'll be able to catch up then.
After I get all this research done, I wonder where I'll submit it exactly? I'll send a proposal for a supplemental manuscript to be published at the same time The Story Template is released. That will definitely help my resume, if I can make it happen. I'd like to submit it to a competition as well though - something comprable to science fair, but appropriate for the topic. I have considered twisting the focus around to appeal to scientific audiences, but decided that I'd never be able to compete. This is art, and unfortunately, no matter what I do or how I say it, scientists and engineers would look down on this idea. I'm applying a scientific approach to the analysis of something subjective. This isn't behavioral science; it isn't any kind of science of which I'm aware. Nevertheless, I don't think it's bad. I can't care what anyone else thinks. This is a good premise, and I will prove it.
I'm not shuffling around someone's lab, connecting breadboards and screwing with stubborn computers. I'm not working on a government-funded project that will contribute to national security or advance medical understanding. I'm not dependent on someone else. I don't have an incomprehensible 20-word title or sponsers or clients, but I did choose and articulate my own research. I created my own thesis, developed my own methods, am collecting my own data, and will publish my findings as only my own. I will not have the opportunity to explore art as I am able now without the cost. Science and math are the way of the world, because they are practical. Art can be successful, but it carries more risk. I'm only playing the lottery when the cost is small.