- Project Owner: wagle
- Current Needs: I didn't know about this contest until Saturday May 7, 2011. Time. I already have The Great Need to concentrate on it.
Our idea: Evolving TagSpace as an Approximation of Cognospace
In March 2003, after a long career as a problem solver, something happened to my brain. I can no longer solve new problems without getting very confused. I can talk about solving problems all I want. I now solve new problems by describing to the computer how to solve the problem and then having it walk me through it. The issue is complex, but my problem seems to largely have to do with my primary spatial short term memory not holding things for more than 1-3 seconds. I think primarily spatially, and have to translate to words as I can. I am (still) exceptional in spatial thinking, but maybe only above average in verbal thinking.
I must rely on and aid my spatial thinking. Its how I program and otherwise solve problems.
I use various computer programs as tools to help me organize, plan, and solve problems. My tools are currently primitive, so I am quite slow. But the tools also extend the size of problems I can “hold in my head”. I can tackle larger problems than what I used to be able to do.
It is a life goal to perfect my tools and to permit everyone to use them with great effectiveness.
My tools currently use crude approximations of a number of ideals. One ideal is what i call “CognoSpace” which would be a full computer representation of what goes on in my head put on the computer so I can reason about it without getting confused.
The best approximation to CognoSpace I have found are anti-ontologies in the semantic web. What I am calling anti-ontologies is the notion that some people seem to take that ontologies are too weak to express everything you want or need to express. I am currently beginning work on a survey of the ontology literature. I am planning a talk on “anti-ontologies” in 3-4 months to the local Semantic Web User Group.
“Tagspace” is a first “good” approximation of CognoSpace.
A recurring pattern in my current “computer-aided-thinking” tools is to have lists of tagged “items” and to use the tags to organize and process the items on a hourly, daily, or monthly basis.
Example 1: LifeBalance is one version of my organizer that uses a planner program in an odd and unintended way to process long lists of task-like items. There, each item has a single tag (a “place”), but those tags can include other tags, so some complexity of structuring is permitted. I have about 3000 items (“ideas”, “plans”, “chores”, “errands”, etc).
Example 2: TagSieve is another version of my organizer that uses Firefox to process “knowledge” in the form of 6500 bookmarks to the web, local files, and 500 or so entries in my personal wiki. Here, I have a simple bookmark-tag query engine that permits me to form queries using AND, OR, and NOT expressions on specific tags.
LifeBalance and TagSieve above both use simple approximations of the more complex approximation I call TagSpace. Both approximations have shortcomings I need to fix.
One thing I really need to do is to replace LifeBalance as my primary organizer/planner. One thing I need to do to do that is to understand TagSpace much better. And to have that better approximation available while I develop the replacement. This is basic bootstrapping.
In this project, I will extend the capabilities of TagSieve to permit more complex queries and to permit tag-structures that better approximate what is going on in my head.
My TagSieve tags are complex. Instead of being simple terms, they contain a sequence number to force particular sorting orders, a category indicating what general sort of item it is, and a description string indicating what specific sort of item it is, usually with the first word being the keyword being looked for by name by the human.
The next step is to enhance the queries to support queries by keyword, and to support queries containing wildcard regular-expressions on the complex tag names. What I worked on today (Saturday May 7, 2011 in Portland Oregon) was to understand the code well enough to add these two features. At that point I applied for an Intellij IDEA license for editing the javascript, since all the editors I had available to me on Saturday were too primitive. Two days ago, I was accepted for a free license.
My complex tags are specific to me, and so are unsharable. Simple parts of them are sharable in the short term. Recipes for forming complex tags is sharable in the middle term.
“Cognophysics” is what I call the idea that The Workflow should adjust to the current user to make him or her as efficient as possible. I prefer function over form, but cognophysics combines the two notions in a way emphasizing efficiency, which is what I want. What’s currently a common takss should become easy and fast.
Interaction must be fast even for processing large bodies of information (like 3000 tasks or 6500 bookmarks). For me, I must be able process an item before it drops out of my 1-3 second short-term-memory. Eye-candy that promotes locating an item quickly is good. Eye-candy that wastes time is bad. I like the idea that I understand Tufte promotes.
This project is relevant to “Applications for Good” for Education for two reasons:
First, I want to generalize what’s good for me to helping other people out. My current ontologies are very brittle. They are hard to change, and not particularly sharable. Being able to visualize and change the ontology (graph of the tags) rapidly would permit people to try things out and communicate with each other.
Second, at some point in the next few decades, computers are going to be smarter than humans. Rather than sit still, people should use computers to become smarter and more effective. My TagSpace idea is a tool for letting people process and review information more quickly and effectively. Linkspace and Cognospace would be even better.
Project resources:
My MacBookPro is my primary platform and was paid for so I could use and extend my organization and planning tools.
I’m used existing tools as best I can. Now is the time to extend them.
Tagsieve is a Firefox extension, and I have much of the existing work on that on github.
Two days ago I was accepted for a free Intellij IDEA license so I could work on this project.
My current organizer/planner permits me to do 30-40 hr/wk employment as a programmer.
I’ve ordered a javascript book that explains modern javascript that should arrive later this week.
Video:
Our team:
Perry Wagle, DeveloperOrganizerManagerDesigner
The authors of tagsifter and tagsieve. Neither of them are able to devote much time to the continuing development, but maybe my work and help will inspire them.
