Being the miscellaneous section of the forum, I'll spill out some of the half-baked ideas I have bouncing around. This post will eventually suggest an idea for what I will call a Genetic Dream Diary. Just mentioning that because I think it will be hard to see the relation to lucid dreaming in the lead-up explanation that will begin with the next paragraph.

Part of my work involves assisting with research on genetic artificial intelligence algorithms. In short, that involves creating computer programs that learn by mutating, testing out mutations, and then creating new mutations, called offspring, spawned from the more successful previous mutations. By doing this the program finds better ways of doing whatever it's designed to do. It does this not by the scientific method of forming informed hypotheses and pursuing them, but by trying out subtle alterations to whatever has proven most successful. Most all of the offspring are unsuccessful. It's through large numbers of offspring each pursuing different mutations and new offspring learning from the most successful individuals that optimum solutions are found. What happens, and this is significant, is that solutions are found that are not intuitively pursued by even the most creative humans testing out their hypotheses scientifically. It's a matter of scale and not being limited by human mental paradigms, not that the computer is smarter than humans. Ok, here comes a more concrete explanation of what I'm talking about. Prepare yourself, because I am about to call us all a bunch of mutant offspring.

The lucid dreaming community is much the same as this. Dr. Laberge is the very successful offspring of those whose work came before him and we are all his offspring, each of us, intentionally or not, experimenting with mutations of what he found to be successful. So, the approach that you take to lucid dreaming and the conditions your attempts take place within make you one of many unique offspring. This is a good thing for a dynamic organism attempting to evolve toward a desired goal. I'm assuming that goal is to discover and develop more effective ways of learning to lucid dream.

Now there is a big difference between our community and a well designed genetic model. In a well designed model, data about the exact nature of each mutation is recorded, so that if the offspring becomes successful it can be used to spawn offspring. We, as individuals, on the other hand, have a hard time knowing what exact changes in our approach have resulted in success. We end up with lots of people who become proficient at lucid dreaming, but much of the confluence of variables that led to their success remains unidentified, even to the individual who has learned to become lucid at will.

For example, it's possible that simply thinking a lot about lucid dreaming and doing reality checks throughout the day for a long period of time could result in a constantly increasing number of lucid dreams. Now the person doing this would be likely to be trying lots of different approaches in addition to this. Let's say this person decides to eat French toast before bedtime. If they had an increase in lucid dreams after eating the French toast, they may believe that the French toast was the cause, so they continue eating French toast and their lucid dreaming rate continues to stay high. This person ends up telling everyone on the forum about how French toast works well for them, without ever seeing the actual reason for the increased rate. In the meantime, they are creating offspring that are pursuing a strategy that doesn't work, because people who read about it want to try it.

I'm saying that the complexity of learning lucid dreaming makes the evolution of its effectiveness move very slowly because individuals are continuously covering old ground and failing to recognize directions that lead to success and pass that information onto other members, although forums like this definitely help a great deal.

So, what I suggest to make this genetic organism more efficient is what I'm calling a Genetic Dream Diary. The Genetic Dream Diary would be a networked computer application. We'd record all the data from our dreams, including what approaches we are applying. All this data would be sent to a common source and made available for analysis. Trends would emerge and correlations could be found, including those that were unimagined. Successes would be recorded and information distributed to other individuals who could try out the combinations of variables that appear to work. Recommendations could be sent out and individuals could choose to follow them. In this way the dream community would become a highly efficient evolving organism. It would be an organic intelligence developing independent of (though assisted by) its participant's intention and therefore free of our limitations of creativity and subjectivity.

If only the SuperNovaDreamer application compiled data and sent it to a common source, there could be so much that could be learned from that. Something designed specifically for this purpose of networking our data could really create some interesting possibilities.

Gordon:

I have several responses to your post, but this one is easiest to articulate in the amount of time I've available right now:

TLI's course in lucid dreaming begins with cataloging and analyzing dreamsigns. By establishing a boilerplate that new contributors (and all, but I'm thinking especially of newbies, and don't think I don't still consider myself one) could use to record their dreams--both lucid and not--and incorporating this data into a project such as you have suggested, we might not only speed up the evolution of lucid dreaming collectively, but be able provide personalized feedback and solutions to new contributors based on what we already know (i.e., without waiting for the results of the French toast test), getting them more immediately involved in TLI and--I would hope--lucid dreaming experiences. If I was in marketing, at this point I would go on about improved new-member satisfaction and retention, and an increase in TLI's appeal to non-members, because the required resources--human notwithstanding--are not likely to pay for themselves, but that's a different matter... anyway, I've been looking for another good reason to swap out my composition notebooks for a PDA [smile]...

Good night. I'm going to try to not dream about parsing data, now [smile].

Joshua

I think what you described is right on. Being a part of a larger project and knowing that your data is contributing to it would build quite a lot of motivation.

I've been contributing to the SETI@home project for several years now and all I've got in return was a certificate saying how many computer hours I've given them. What keeps me interested is the feeling of contribution (and the possibility of finding ET). This project would provide much more payback to its participants along with a feeling of participation.

I'm looking forward to your other responses to my idea.

As I laid down to sleep last night, just after writing my previous post on this topic, I began to consider these things:

I do want to note--for my own humility, but also because I predict similar responses from others--that, beyond the consideration of resources, there are caveats to what I think we are proposing in this discussion. The most obvious to me seem: It is possible--indeed, probable--that the very process of recording and analyzing one's dreams facilitates lucid dreaming. If that is so, then there could be a negative reaction to mechanizing that process. And speaking of mechanization, let us not overlook the romantic aspect of keeping dream journals and sharing our experiences with each other via prose and poetry. I would hate to see Keelin's kind and wise words replaced by the Batcomputer spitting out tonight's lucid-dream-induction itinerary. I am certainly not saying that these are reasons to not speculate or pursue a project such as this; I think--does it need to be said?--that it's a pretty neat idea. I'm just contemplating more if's, and keeping in mind that, while I read Norbert Wiener for fun, not everybody wants to live in a William Gibson novel [smile].

That said, if you're reading this, then you're most likely interfacing with a computer anyway. Please, dive in; the water here is warm [wink].

Still more, later.

Joshua

Funny: When I became lucid, this a.m., my dream was replaced by a character display--like a computer screen was projected onto the inside of my eyelids--giving data about the dream. I could change screens by concentrating on doing so. When I decided to enter into another dream, my lucidity dissolved with the display.

Gordon:

I have another, hopefully constructive, caveat for you:

I assume that, in recording their dreams, the dreamers will need to fill out a standard form from which the computer can glean information (if they simply entered a prose paragraph, you would need a massively powerful machine to identify and collate usable data). If this is the case, be careful of how you construct that form ' it needs to be simple enough to fill out literally in the dark, yet complex enough to absorb experiences and techniques unique to each dreamer. If it's not simple, then you run the risk of dreamers filling out forms long after their dreams occur, when it's easier to do so (and the dream is less remembered and invites unintentional conscious embellishment). And if the form doesn't have a way to allow unique methods or experiences to be noted, you might wind up with a very shallow data pool, limited to what the form permits. So, if the dreams are corrupted by conscious afterthought, or the data stunted by a limiting entry system, you might not get the efficient genetic organism that you seek. Instead, you could wind up with very large piles of questionable data that conform to whatever hypotheses were used to create the entry form.

I hope this doesn't sound too negative ' this is an excellent idea, and to see it applied to the advancement of lucid dreaming is most exciting! I would bet it wouldn't cut down much on the "French toast" proclamations, though"

Good luck!

Peter

Joshua,

I'm not concerned that the process becomes dehumanized. I think we're just all too human for that to happen. Besides, it would take a serious fear of technology to motivate someone to forgo the benefits and advancement that we're talking about and anyone who is using the internet to communicate is unlikely to be this much of a ludite.

Re: your data dream. I've had a variety of dreams that seem to be taking place on a computer screen, which is not surprising considering the amount of time I spend on the computer, both at work and home. I've even had a series of dreams analyzing the same data--so far, nothing worthy of bringing back to waking life, though.

I realized that the Dream Diary wouldn't have to be an application, but could be a web page. You log in, check off the relevant data in the form that appears above the place to record your dream, then fill out your dream diary for the day. Wow, this just became a whole lot more possible. Of course, only you could access your diary contents, but the data would be compiled with that of everyone else. (I think combining the data collection with a dream journal would increase interest and participation.)

Data would appear on a page along with personal descriptions of approaches that individuals were using. Charts would be nice and everyone could see the data. Everyone would be free to imitate successful individuals and the results of that would appear as new data. Thus, a more efficient and organic evolution. "Hmmm. The people who are automatizing their reality checks are showing a quicker increase in their lucid dreams, I think I'll try that, too." : )

Its clearly time to start brainstorming on what kind of data should be collected.

Dream recall--How many dreams did you recall from the previous night.

Lucid dreams--How many lucid dreams did you have the previous night.

Sleep duration--How much sleep did you get.

Method of Induction--Self descriptive, but should include details.

Lucid dream history--An estimate of what was happening before you started recording everything.

Just this much would be incredibly interesting and useful.

Time for me to go to bed. Any contributions to this idea?

Gordon:

I think I'm thinking along the same lines as you. The way I see it just now, there would exist a web interface with input fields for standard queries such as you have indicated: How many dreams recalled? How many lucids? WILD? Did you use a device? 61-point relaxation? Etc. Then there would be a place to describe the dream(s) in prose, maybe with a checkbox to allow sharing the dream(s) with the forum. The interface could display statistical information: how many dreams recalled, lucids, how long, etc. over time. All of this can be accomplished by a typical database and web interface, and should not be too difficult to implement (given the resources).

Now, a wee more difficult to implement: I'm also considering--at a 500-foot level for now--the possibilities of pattern and trend analysis. The way I see this is something like: A pattern-recognition program--that parses the prose and compares to other parsed prose and input values--determines that enough people ate French toast before having lucid dreams that it is worth testing as a hypothesis. The next time someone logs in to the interface, the suggestion to eat French toast is presented along with an acceptance toggle. If the person accepts, then the next set of data they upload is treated especially for the French toast hypothesis. Or, alternately, the suggestions could be something arbitrary or semi-arbitrary, such as "think about the color blue," and the data of everyone who agreed to do so would be analyzed in light of what effect the color blue had on their lucid dreaming experiences. It might help if we could boilerplate the prose even a little, e.g. by inserting tags around reality checks.

A little easier: still with the pattern-recognition, but just looking for the occurrences of possible dreamsigns, which goes back to my first post in this discussion. Within the first several uses of the interface, the program should be able to suggest things just for this one user to look to for reality checks and lucidity induction.

That's all I can squeeze in during my short break from work. Thanks for keeping this post going, Gordon and Peter. Later.

Joshua

Joshua,

From what I know of computational linguistics, its not possible to parse prose into anything that would be very useful without putting more work into it than its worth. Like Peter said, I think it we'd have to feed the information directly in with checkboxes and pulldown menus in order to make it useful.

More later. I'm at work now too.

True that, which is why I began in the direction of what seems tangible--really just continuing in the direction that had already been pointed to. The rest was part of the brainstorm, allowing for possibility--extensibility--but not depending on it. It's fun stuff to kick around, and you never know when someone's going to publish something that brings that sort of thing much closer to our reach...

But my subsequent brainstorming has been concentrated around the region of:

1. What can we know about a dream and the relevant events surrounding it? (No epistemological intentions [smile].)
2. And of 1, what can be communicated via checkboxes and pull-down menus, etc. (i.e., easily)? (Also, how can checkboxes, pull-down menus, etc. facilitate communication of difficult information, such as prose.[1])
3. And of 2, what could improve the lucid dreaming experience--in any way--of an individual and/or community?
4. And of 3, how do we usefully present the information to the individual/community (stats, graphs, summaries, etc.)? (The sequence of questions is not rigid.)

That said, I'm still gathering ideas.

(Footnote 1: E.g., the ability to post a free-form, dream description to a particular forum discussion, or to send the dream description and relevant other data to one of TLI's research projects.)

Yes, those are certainly THE directions we want to go.

Here is a description of a semi-controlled approach of to a genetic LD community.

I'm imagining a BB forum, much like this one and each of the threads is an approach to inducing lucid dreams. When a person joins signs up, they begin by entering their LD rate, even retroactive data if they have it. (This is the data that all other variables will be compared to, to look for spikes or increases and what may have caused them.) Each thread would represent an approach or combination of approaches to LD induction. A person takes part in a thread when they begin trying an approach and data on their rate of lucid dreaming is collected alongside their data concerning the approach they are trying. Aside from data on the approach, qualitative data in the form of a journal that describes, in prose, whatever comments each person wishes to make about progress, or whatever.

A newbie to the forum logs in for the first time and opens forum's first page. After being prompted to enter data on the history their rate of LDing, they see a list of approaches that are currently showing good results. (Approaches with good results would attract new converts.) The list has charts that compare standard induction technique's improvement rates with those from the threads.

The newbie clicks on "Automatizing Reality Checks" and opens the thread. There they see a description of how to carry out the approach, how to input the data, more charts about how others are doing and the journal entries of other members. The charts show a bump in the average LD rate of the users that are using that technique that started a few weeks after they began using the technique.

They begin using the technique, record their data and take part in various threads.

After they have been using this for a while, they read that the most successfull person in that thread is combining "automatization" with "french toast". They check the data and it appears that those who combine the two approaches are even more successfull and they decide to take part in the "french toast" thread as well. Then they suddenly explaim, "I just mutated!"

This could be getting ahead of myself, but here is an idea of some of the things that could be tested out with a system like we'fve been discussing. This is a list that I threw together while thinking about setting intention for prospective memory tasks (prospective memory training is also involved in my work). I was thinking about prospective memory tasks other than for lucid dreaming when I made it, so some may not apply.

It started out as what I unofficially call the BFD principle of prospective memory, but began to expand beyond that. The principle goes like this--the bigger deal made about setting the intention of the prospective memory task, the more likely we are to successfully accomplish it. I'fm suggesting that the following will affect the likelihood of the success of prospective memory tasks. Any of these could be mutations that individuals could test out.

--Subjective importance of the target behavior. This could take the form of consequences or rewards. Stronger associations with the consequences or rewards would result in more success. --Intensity of the intention. --Written is better than verbal. At least, it'fs proven so for me. --Length of time spent on setting the intention. --Thoroughness of the intention --Repetition of reminders of the intention, external or internal --A variety of statements and reminders of intention --Continuity. Do ten reality checks during the day—the tenth one intended to be done while asleep. For the longest time, I was actually feeling that when I went to bed, my reality checks were finished. --Predicting/visualizing cues and context during intention setting. This seems to be an important element in MILD. --Time between the intention and the cue --Number of context changes between the intention and the cues --Depth of involvement in the distraction (Sleep can be pretty distracting, can'ft it.) --Habit. How much experience do you have with successfully carrying out the task? --A variety of sensory associations w/cues. This could probably be connected to visualization. --Personal control of cues. Cues you create yourself are more likely to be carried out.

These are out of your personal control, but still have an effect. --Number of cues --Predictability of cues --The lack of ambiguity of cues or context --Familiarity with cues/context

I may be away from the computer over the weekend. Still thinking, though.

I realized this morning that I was stuck on separating quantitative from qualitative data, and how the former--however comprehensive--didn't seem to answer questions posed by the latter, most importantly the overall success of a lucid dreaming experience. I've begun to solve that problem by allowing for the dreamer to quantify their dream qualities. To illustrate, and I'm paraphrasing some of Gordon's material here (hope you don't mind, G., and hope you don't find this too elementary):

Each thread would represent a technique ("a method of accomplishing a desired aim") for improving lucid dreaming (I am assuming that induction is just one of many things that can be improved, even though it seems to get--deservedly--the most attention). Let's say we've identified three measures of the overall fitness of a lucid-dream-improvement technique: control, retention, and recall. For every dream submitted to a thread, the dreamer would rate each measure, something like: worse(-), same(0), better(+)--requiring base values, maybe the dreamer's typical ratings; or a rating from 0 to 1, 0 representing absolutely no control, etc., and 1 representing absolute control, etc. (Rating guidelines with examples could be provided for the dreamer.) The most successful techniques would be combined with others to create more complex techniques (offspring) that would be rated by the same criteria as the parents.

Here is a new list of questions:

What measures should we use to determine the fitness of a technique? Any of the three I suggested above may be valid, but I intended them only for the example.

What system of values do we give to the measures? Can different measures have different value systems? Perhaps induction is always 1 or 0--lucid or not--or expressed as a frequency, whereas other traits such as dream recall are typically fuzzier (pun intended [smile]).

What techinques do we begin with? Those listed in TLI's course? What about stuff from the forum, or new ideas as we evolve (see question about mutations, below)?

How do we determine which techniques to mate, i.e., how successful does a technique need to be to qualify as a potential partner? Certainly we would want to combine techniques that complement each other. Can we have dominant and recessive techniques? How complex do we allow combined techniques to grow?

How would we introduce mutations--i.e., methods not present in either parent technique--into the offspring?

Joshua

Now maybe I am getting ahead of myself, but...

Because each technique seems to improve specifically one attribute of lucid dreaming, e.g., reflection-intention focuses on induction while spinning focuses on retention, perhaps we should class techniques by the measure(s) they intend to improve and allow the dreamer to rate more than one technique--one from each class--for each dream. So rather than each technique being rated for its overall fitness including measures it doesn't really qualify for, it would be rated for one measure of the LD experience's overall fitness, or more measures as applicable. Techniques w/in a class could still be combined together, becoming new techniques.

I imagine, then, that the dream journal interface would allow for several techniques to be selected and rated, and would link out to pages presenting the techniques similarly to how Gordon described threads. The interface might include something like this (w/ a better layout, thank you Edward Tufte; also, note that I omitted several fields to shorten this post):

Technique Class 1: Induction Select a technique: [select from list] --Lists all avalilable induction techniques. Maybe uses most recent, automatically. Also, could be option to select random, most popular, and/or most successful technique. Go to technique's page --Opens page displaying description of and collected data about the technique; also lists descriptions from other dreamers' journals. Get technique's form(s) --Opens form for recording data, if applicable, to be printed. These could be coded, but would require a lot of additional resources, especially if new forms are needed for new techniques. Rate technique for this dream: --Using this technique, induction has become... [] Much worse [] Worse [] Slightly worse [] Same [] Slightly better [] Better [] Much better --These would perform mathematical operations on the dreamer's current rating for the class/technique, which is how we would analyze and chart and graph the fitness of the class/technique and the fitness of the dreamers' LD experiences.

Technique Class 2: Recall --Same options as for technique 1.

--et cetera for all classes.

As in my previous post, I use classes INDUCTION and RECALL only as examples. Maybe they're valid classes, maybe not.

I would want the interface to be comprehensive w/o being too complicated. It doesn't take a Luddite to say, "I don't have time to fill out all this stuff."

Joshua

These are MY answers to the earlier questions. There must be many other answers as well. We're still in the divergent phase of the project. : )

Q: How do we determine which techniques to mate, i.e., how successful does a technique need to be to qualify as a potential partner? Certainly we would want to combine techniques that complement each other.

A: If all goes well, we don'ft determine which techniques to mate. If we are able to provide enough data feedback in accessible form, individuals will choose them on their own. It'fs the scale and the lack of control that brings results that we can'ft foresee. When individuals see what'fs working, they will copy it. It'fs kind of arrogant to think that "gwe'h would be able to guide everything better, or come up with more creative innovations, than the collective intuitions of all contributing members of the LD community. Design the system well, as simple and elegant as possible, and then step back and watch the magic. : )

Q: Can we have dominant and recessive techniques?

A: I'fm not sure how this could be applied, but it would be a great title for lists of techniques that have been proven to work and those that don't. Something that is built into some genetic algorithms (a kind of artificial intelligence) is "edormant memory'f which remembers something that has had limited success in the past and is brought out again later. Characteristics that work in the natural world are always worth a try.

Q: How complex do we allow combined techniques to grow?

A: We might make recommendations based on future experience, but that too would be up to individuals. People have different capacities to deal with degrees of cognitive load. I can try three different techniques at once, but am more comfortable with two. However, once techniques are habitualized, or automatized, adding more won'ft tax my capacity—it'fs a very individual thing.

Q: What techinques do we begin with? Those listed in TLI's course? What about stuff from the forum, or new ideas as we evolve (see question about mutations, below)?

A: People are testing out any number of techniques as we speak, or I should say type. We just provide a format for them to record, compile and share their results. (Results they are probably understanding for the first time because they are can be seen in contrast to other results.) Successful individuals and their offspring should fuel things from there. (Laberge'fs books provides enough to keep the entire community busy AT LEAST until things get going on their own.)

Q: What measures should we use to determine the fitness of a technique? Any of the three I suggested above may be valid, but I intended them only for the example. What system of values do we give to the measures? Can different measures have different value systems? Perhaps induction is always 1 or 0--lucid or not--or expressed as a frequency, whereas other traits such as dream recall are typically fuzzier (pun intended [smile]).

A: The measure of success depends on the goal. Rather than having individuals give their subjective interpretation of success, an increase in their rate of LDing, or whatever the goal is, would objectively show the degree of success. Rates of success for each thread would be available in charts and individuals will be attracted to techniques that work. Individuals decide the fitness of a technique by selecting it. Natural selection, In order to know how much of a spike or increase is meaningful; we'fll have to see what degree of increase would be normal regardless of technique. Also, compiling the results is important because in order to know that success with "ethe French toast'f technique was a coincidence, you have to see other results with the same technique.

Gordon:

These replies are pretty close to my own inclinations, particularly in letting the participants create and combine techniques (usually when I say "we" I am referring to the collective participants and including myself in this group).

An increase in LD freq would be an objective--and useful--measure of the success of induction, but how do you objectively measure something such as a dream manipulation technique? Or a retention technique such as spinning?

I like the idea that the most fit techniques are simply the most used, most naturally selected, but what data would be used to entice participation? Certainly testimonials by successful users, but what that could be charted, graphed, aggregated, combined, etc?

Back to work...

Joshua

Note to the web developer lurking on this conversation: No doubt you've visualized how this site would look and have a prototype in your head. You're about to be the first designer of a Genetic Community Website, what is destined to be an important and groudbreaking addition to the Internet age. : ) Let's see it! Your design team is waiting for a link to the prototype.

An increase in LD freq would be an objective--and useful--measure of the success of induction, but how do you objectively measure something like a dream manipulation technique? Or a retention technique such as spinning?

I think Laberge compared Spinning to Hand Wringing or falling down in a straighforward comparison. In the same way, people trying out the techniques would record their results. How many times did you try it? How many times did it work? How many times did it give you 30 seconds or more of lucid dreaming? I wonder how Laberge measured it.

I like the idea that the most fit techniques are simply the most used, most frequently selected, but what data would be used to entice participants? Certainly part of that data would be testimonials by successful users, but what else?

Three lines on a chart, one line represents average improvement, another represents improvement with the relevant technique and a vertical line showing when the technique was applied. That would be enough to convince me, especially if it represented several people who've employed the technique.

I wonder how Laberge measured it. The charts for recording this stuff are in TLI's course workbook. Flipping through this, you'll notice that the charts are different from technique to technique. That's why I was looking for a way that each participant could rate their success that would be the same regardless of technique, so that we don't need to accumulate different sets of data for every technique developed. If we allow participants to create their own techniques, then--aside from them being able to upload a record in docu format for everyone to use--who's going to extend the interface and database to allow new sets of data to be entered?

I hope I'm not annoying you, Gordon [smile]. I work in computer system development, officially for quality assurance, but also in functional analysis, which is why I tend to ask a lot of questions, and some very specific ones. And I ask a lot of questions in this discussion b/c--while you're waiting for a web developer to provide a prototype--I'm preparing a set of functional specifications in my head and some in notepads. So, continuing my line of thinking from before...

Let's say someone wants to add a "French toast" technique to the project. They go to a page that allows them to create a new technique. Probably a publicly editable section for the description, execution tips/tricks, and goal(s).

1. How should we handle the data that needs to be recorded for this technique--what time of day or how long before sleep did you eat? How many pieces? What kind of bread? Free-range eggs? Vanilla? Did you add butter? I Can't Believe It's Not Butter? Maple syrup? Et cetera (yeah; I realize this is a ridiculous amount of data; just illustrating the point)? My suggestion is to treat this data outside of the genetic system, as auxiliary/support data. More about this, below.

2. There are at least two things that I would want to know about this new technique: First: is it surviving? If not, I probably won't bother looking into it further, unless I have a morbid curiosity. Second: how successful is it? (I don't think 1 and 2 are necessarily the same; a technique can be successful yet abandoned for a more successful technique.) In the case of the former, if the measure of its survival is the number of participants using it--utilized techniques "survive" and those that aren't "die"--then we can simply maintain a count of the number of times dreamers use this technique, which could be graphed over to time to see the rise and fall of its survival--its life cycle, which would give us its age--may be useful data later. In the case of the latter, as Gordon said, "The measure of success depends on the goal. Rather than having individuals give their subjective interpretation of success, an increase in their rate of LDing, or whatever the goal is, would objectively show the degree of success." I agree with the fist part of this statement. My comments regarding the second:

a) As I began w/, above, it is theoretically possible that each goal will have a different set of data that determines its success. In reality, there will not be a different one for each technique, but there will still be several, and a system such as this is not likely to be auto-extensible enough to allow for new sets of data that are not congruent to existing sets--hoping someone will prove me wrong [smile]. Unless the participants are all developers in an open-source environment, I suggest a homogenous system of rating the success of each technique. Also, I may not be concerned about such details as yet; all I want to know is the degree of success. Are participants very successful? Slightly successful (indicating a mutation could increase success)? Which brings me to my next point:

b) I don't have a problem w/ using subjective data, if it's used appropriately. It makes use of the fact that our brains are already wired to perform functions that are too difficult to embed in the computerized part of the system, such as draw conclusions from or make decisions about complex sets of data. Also, using subjective or fuzzy values has a more natural feel than objective/crisp values (yeah; I know that subjective and fuzzy are not synonymous). If I ask somebody, "What induction technique should I use?" they are more likely to reply w/ "Technique X; I've had a lot of success with it," than "Technique X; out of 41 dreams I have experienced 45% lucidity 41.63% of the time or 17 times and 95% lucidity 21.95% of the time or 9 times." Granted, this latter statement can be very useful, and we may want to see this data in a user-friendly format, but if we cannot find a way to provide such a format for every variable that this statement could include, then I would be satisfied to see the combined results of the former statement. Also, the question "Why is Technique X successful?" is not likely to be expressible by objective data alone.

One more thing while it's fresh in my mind: if the French toast technique has the same goal as, say, the automatizing reality checks technique--which is, say, to increase LD frequency--then I am less likely to combine these techniques--maybe I would, but I could also abandon one for the other--than to combine techniques w/ different goals, such as eating French toast and spinning, which are not likely to be abandoned for one anther, but for something else entirely w/ a similar goal and a higher success rate. Of course, this is all dependent on the individual's cognitive load, as Gordon said. a) Here again, w/ a homogenous rating system I could easily combine the success rates of eating French toast and spinning, or compare the success rates of eating French toast and automatizing reality checks. But I also recognize that the output data for all techniques could be homogenous, while the input data is different from technique to technique, so... b) If we are going to include more objective data than this into the system--instead of keeping it outside as auxiliary/support data--, then again my suggestion of technique classes or types, each w/ its own set of objective variables that would determine the success of the technique. When a new technique is created, its class--which reflects its goal--would determine what data needs recording and make an appropriate place for it in the interface/database--I'm not even sure this is doable--that we can make such determinations up front about the required/useful input data--just recognizing the possibility. Also, the classes could be used for grouping, combining, and comparing results from one class to another, e.g., induction techniques vs. other induction techniques, or vs. control techniques, perhaps the latter classed or sub-classed into dream manipulation and self-control techniques. These lines of demarcation were pulled right from TLI's material, and as long as they're natural and fit--and can be altered if not--then I don't see them adversely affecting the freedom of the project's participants to explore new ideas.

Wow; that was a mouthful. Later.

Joshua

Hello,

You'fre certainly not annoying me. Did I come across as if you were? I didn'ft mean to if I did. You'fre addressing lots of issues that need to be thought out. It'fs the old divergent/convergent tug of war going on, I think, which creates a good balance. I'fm remaining vague and you'fre going after the details.

From your post, I can tell we are definitely converging on quite a few points. Of course, something like this could become an enormous project, but I think we could also create something that would still function really well even within a forum like this one.

The question I want to ask right now is how simple can we make this system without compromising its ability to encourage a more efficient evolution. I think it can be quite simple. I think it would be best to pick one goal (ex. rate of lucid dreaming) and one technique (ex. Automatization) and build a page centered on that. A pilot test would probably answer many of our questions. I think a thread should minimally contain the following:

-Description of the technique for people to use -One thread, one technique or combination of techniques. -Some kind of data input method -A simple representation of results -A forum for recording qualitative data and communication that isn'ft necessarily quantified

I think that with these elements present evolution of successful techniques would be efficiently facilitated. More complexity could be added later to tweak the effectiveness.

I think it would be good to start writing up kind of manual that brings together what we'fve established. Then we can work on amending that instead of the rather non-directed way we'fve been going. I think that will also make it more coherent and other people will be able to join in without needing to read everything that came before.

I'fll begin putting together a thread, minus the automation, and run it by you as I make it. If anyone contributes (I think they will. Dean has already tried my automatization method and is reporting his results now in the Reality Testing thread. ), I'fll compile the data by hand and post it at the top of the thread. From there, I think the direction will become obvious.

I'fm giving a seminar next week which will limit my activity here a bit, but should be able to pop in every other day or so.

Gordon:

You're certainly not annoying me. Did I come across as if you were?

Not at all [smile]. Being a nuisance is something of an occupational hazard for me, and I've learned to anticipate certain reactions to my interrogations and proposals concerning another's brainchild, particularly if I've some enthusiasm for it.

I look forward to participating in your trial thread.

Joshua

I have opened the trial thread under Research, Theory, and LI experiments: The Automatization Technique.

The Automatization Technique

It's kind of buried down at the bottom below lots of abandoned threads, I hope it gets seen.

Gordon,

How is this coming along? I have not forgotten about it; indeed, I was recently moved to reconsider it at some length, and to write my considerations down (or out, as the case may be here in cyberspace). Here is what I have so far: http://biligene.home.comcast.net/partone.htm. It still needs some work and some polishing--the style is pretty plain, as are some of the expressions, but at least I have excuses for the style: lynx and vi. [smile] Still, I would appreciate your thoughts, as well as anyone else's.

Joshua

Merry Christmas everyone,

Here is the dream where I solidly interfaced with reality, and how I came to call it Prime Infinity.

June 12, 1988 Dream 3--

in an attic room, ## had left comic books and coats, the comic books were DR. SOLAR: THE MAN OF THE ATOM, and I forget what else, there was supposed to be a treasure there too. Later as I was angrily chasing a midget, I was looking for him in some {cylindrical} cardboard cans (the classic routine, I saw him jump in one, but when I got to it, it was empty when I looked in it, when I looked up, he popped his head out of another one, and I ran to that one . . .) and then finally when I upended one-- a crown fell out-- it was adjustable and marked on the inside: LOMBARD, I put it on and looked into a mirror, and I had bright pink lipstick on, and was wearing a hot pink strapless gown, the crowned seemed able to grant wishes . . . when a young lady that I liked came through, she wanted to talk to me, and she wanted to know how I felt, I told her that she was overly articulate, and she didn't act normally, I told her that I used to do the same with even bigger words (thinking anthropomorphize was exactly descriptive of what it is) and she didn't even seem to notice my outfit, but was interested in me, and my opinion of her.