FRDCSA | internal codebases | CLEAR

[Project image]

Architecture Diagram: GIF
Code: GitHub

Jump to: Project Description | Parent Description | Capabilities

Project Description

CLEAR is a great way to have books, manuals, websites, etc, read to you, allowing you to pause, quit, resume, and filter out nonsense. Clear uses the Festival text-to-speech system and ccp to do this. It is very useful for studying. For instance, while browsing a researchers website in w3m-el, you can select a region over all of their publications and execute M-x clear-queue-all-links or "\C-c\C-mc" to queue all of their papers. A message containing a command to queue the links is then sent from Emacs-Unilang-Agent to UniLang, which sends it to CLEAR. The queued links are stored to the current readlist. Whenever you have a chance then, you tell CLEAR to resume reading and it reads you books. It sorts the reading list by topical dependency.

The objective is to model the user's reading flux at the sentence level. Conceptual understand may then be modelled through analysis of the sentences. A lot of data on the entire process is recorded. The purpose is so that the computer has another modality, in addition to things like expected background knowledge, apparent knowledge, to model which axioms the user is familiar with. (Eventually the Textbook Knowledge Formation tool chain we are assembling with the Chess Analysis Knowledge Base project will be used to model fields axiomatically as opposed to the document level granularity currently used by CLEAR.) By modelling the reading flux, a much better mental model can be created and the system can behave contingently. This information is critical to many tasks, such as:

  • Assessing the user's background knowledge.
  • Verifying familiarity with some tradecraft.

The motivation for CLEAR is inspired in part by the following information from the JAVELIN project: http://www.lti.cs.cmu.edu/Research/JAVELIN/

Utility-based Information Fusion. Any item of information I can be assigned a value representing its utility to analyst A with respect to task context T and question Q. The utility value can be used to rank the possible answers in a manner inspired by Maximal Marginal Relevance:

U = Argmaxk[F(Rel(I,Q,T),Nov(I,T,A),Ver(S,Sup(I,S)),Div(S),Cmp(I,A)),Cst(I,A)]

Essentially all information items (facts, links, inferred relations, etc.) in consideration for fusion into an answer may be ranked in utility to the analyst as a function of:

  • Rel: relevance to the requested information
  • Nov: novelty (likelihood that the analyst does not already know it)
  • Ver: veracity (of source S) and support for conclusion I within S
  • Div: source diversity (analyst may want contrasting or reinforcing views)
  • Cmp: comprehensibility of information by the analyst (one can assume a uniform distribution until the system learns otherwise from analyst feedback)
  • Cst: expected cost (e.g. time) for the analyst to assimilate the information.

The way CLEAR does this is by monitoring the users readers, and deriving belief models of the user's state (such as awareness, eye-position, whether they were listening (in the case of auditory media) etc), as these reader applications were running. Currently only reading is implemented, as I don't have a suitable eye-tracker.

In addition to reading, currently CLEAR can quiz the user, as in this example:

/var/lib/myfrdcsa/codebases/internal/clear $ clear -q lists/camo.rl 
Readlist is lists/camo.rl
Initializing TTS engine...
festival: no process killed
server    Tue Dec 28 20:10:06 2004 : Festival server started on port 1314
Reviewing library/us-army-field-manuals/extracted/20-3/US ARMY FM 20-3 Camoflage Concealment And Decoys/ch4.PDF...
client(1) Tue Dec 28 20:10:08 2004 : accepted from pc-sonicparts

Question 0
Make optimum use of concealed routes, hollows, gullies, and other terrain
features that are dead-space areas to enemy observation and _____
% fighting
Incorrect! (firing)

Question 1
Although an enemy's use of radar and _____ aerial recon hinders operations
at night, darkness remains a significant concealment tool.
% visual
Incorrect! (ir)

Question 2
When natural _____ and concealment are unavailable or impractical, the
coordinated employment of smoke, suppressive fires, speed, and natural
limited-visibility conditions minimize exposure and avoid enemy fire sacks.
% cover
Correct! (cover)

Question 3
Designate concealed _____ for movement into and out of an area.
% routes
Correct! (routes)

Question 4
A trade-off, however, usually exists in _____ of a slower rate of movement
when using these types of routes. 4-18.
% this
Incorrect! (terms)
One can see how Question 4 is not as useful as the others, but overall, I find the system to be very effective in liu of a real system. If the shoe fits...

This system used to be called: Correctly Learning An Individual's Reasoning Visually Or Yielded Assuming Normal Cognitive Execution (CLAIRVOYANCE), but this is just sort of inaccurate. CLEAR is much, more, well, clear ;)

CLEAR is primarily useful for reading documentation, papers, books, websites, digital library content, and other sources that are of practical use to the project, so that we can get much more done, especially while resting or moving. A critical feature is the ability to read documentation for installed packages, because mastering newly installed software is critical to the FRDCSA mission goal. It also does very well with important documents, like the above survival and preparedness information (as the recent tsunami confirms). Lastly, and most important, it allows us to implement a system of instruction so that we can help to train project members in various capacities and verify their progress, without being burdensome at all - but rather entertaining. This can be combined with CRITIC for collaborative filtering based ratings to enhance the value of the information. CLEAR in conjunction with DigiLib is also used by job-search to ensure familiarity with position requirements. CLEAR uses functionality from CoAuthor to compose custom reading specific to users' tested proficiency and known reading history.


  • clear should say something like I cant read this document, there's too much code it in, or something.
  • clear should track how many times a given document has been read, and when it has, etc.
  • clear: see https://arbital.com/
  • Write a module for clear which detect code and other Non English and optionally filters it out, or mentions to look at the diagram.
  • Finally improve clear so it doesn't screw up saving things. That will be difficult but very much worth it.
  • Create a function for clear to indicate a text that is misbehaving with it, and put it in a regression test "library".
  • Add a feature to clear that makes texts that has been read more recently highlighted in a deeper red.
  • Fix clear which is really messing up badly.
  • Improve clear so that it handles better the storage of assertions about the document.
  • set clear to change volume as needed /var/lib/myfrdcsa/codebases/internal/clear/doc.
  • Figure out what is going on with pse2.el being undefined and also with clear messing up.
  • Ensure clear doesn't send any commands.
  • Add a button to clear to record that a document has been fully read. Have it have a 'are you sure' prompt.
  • Lower the volume of music when clear engages.
  • Make sure clear works with emacs -nw.
  • Add buttons to clear for going to next and previous page.
  • Implement a system for recording when we learn different Cyc concepts. Integrate into clear/Academician/study space.
  • Add a reread button guarded by a verify dialog to clear.
  • Use spread0r and write an Emacs version, and also integrate it into clear Tk version.
  • Create a function to read a SPSE2 domain using clear.
  • Fix clear up to when you click on a word pronounce for someone. Make font resizeable - and integrate TTS and literacy efforts. Install on several computers that we own, and use to teach children literacy. Have it correlate and do flashcard like training of words and sounds. The end. Bob's our uncle.
  • Create a new capability to say that a given goal is clear in the sense that nothing is obstructing it and it can be worked on immediately - do analysis through Prolog or Cyc or PrologCYC
  • Make the clear read list default to the pdf documents in my home and Downloads dirs in reverse temporal order.
  • Add KNext processing to clear.
  • See if we could have clear read into BBB.frdcsa.org
  • Add a feature to clear/academician to start reading from the page that the 'c b' command is sent from.
  • Re-add the feature for pressing y to prevent clear from reading past when you fall asleep, only, make it work with the headset buttons now.
  • Fix clear truncating last sentence.
  • Add the ability to re-read a document to clear.
  • Write clear functions for pausing, rewinding, searching etc from within Emacs. So managing clear is just like any other keybinding.
  • Fix the error that we get with clear not reading certain sentences.
  • Have clear store all instances of documents that it cannot parse for analysis.
  • Have the different clear instances communicate with each other.
  • clear should use kbfs to represent the state of a file, so that we can read a file and then read all changes between versions, etc.
  • Create a mode system using headset by using double press of middle button as escape or something for clear. Have the ability to change up the reading list, and all kinds of other tasks using the headset's 3 buttons, with double clicks for each of them or something.
  • Add a feature to clear that detects the reading level of the text and guages the reading level of the listener and then speeds up easier to understand text.
  • Using the new academcian/clear code for getting pages right, perhaps the system could also strip repetative boilerplate where it exists using code derived from AIE.
  • Add a close tab button to clear if it doesn't already exist.
  • Make an interface between academician, news-monitor and clear that records when we have read from an RSS feed, that we did read it.
  • Have a user awareness component that can tell whether I was paying attention to clear as it was reading it to me - to constrain when it says "(done (read ...))". Make a special predicate for auditorily read to.
  • Add to clear A title, which replaces "Tab for 2" which displays the name of the currently being read text.
  • Improve the functioning of clear's and Academician's recording of when a document has been fully read, and especially the DESTROY saving capabilities of clear.
  • Have clear use ghost buffers to record buffer file referents.
  • For clear/emacs integration, have it bind searching order something like saying, move reading to this point.
  • Add clear to be able to read emacs info pages.
  • Integrate NLU with clear for reading a given buffer to the user, marking that it has been read, etc.
  • clear up the confusion in existing code about 'CycL String', as it's used in two ways, one, the original, better named 'CycL-like String', and then the second, an actual 'CycL String'.
  • Start recording which articles from news-monitor I have read in full, as in clear.
  • Make it so that if we don't clear the context, it won't reload the data.
  • Integrate SPSE2 planning domains/study domains with clear for reading list management.
  • Add the ability to clear to handle reading multiple books at the same time, switching focus between them as needed.
  • Have this rotate, so clear should read books for part of the day, old unilang messages for part of the day, text files from projects for part of the day, etc. Then we should watch a movie as recreation for part of the day, and so on. Develop a scheduler for these kinds of things. Use PDDL possibly. Have it know to not repeat the same subject. Have it directed by research interests in academician.
  • Maybe have clear read our unilang logs to us, selecting which ones are system messages versus ones written in English for us.
  • Add to clear the ability to use the G930 buttons to control the process.
  • Fix the way with KNext it wraps some statements onto two lines, although it should be clear because it's nice enough to add a trailing period.
  • Add emacs key bindings to clear.
  • For PDFs that are being read using clear, have it OCR the text and then highlight the text on the image.
  • Add to clear/Academician the ability to assert that a given chapter has been read, etc. Extract the knowledge about chapters and so on from texts on literary analysis
  • Tie clear with academician for recording which pages have been read. Then have clear send commands to Emacs-Client in order to turn the pages so to speak of Academician mode.
  • Add the ability for clear queue to pause the currently being read item and resume with something else sent to it
  • Get clear, queue to work
  • Interface academician with clear, so that when you add something as an interest it schedules reading for you.
  • Use the bookshelf of paperless office to represent the readinglist of clear. Tie that all in somehow. Make different repositories.
  • Change the "Are you paying attention?" question in clear to a topic question from the document.
  • Fix chm support in clear
  • clear visualizer or documentary-generator could use the debian package zoomer, which generate a video from a picture by zooming from one point to another
  • Integrate the paper reading system from academician with clear.
  • Write an interface for clear that allows you to download books through digilib from various sources, like Wikipedia and Gutenberg
  • Take the job-search system and have it learn which types of jobs to apply for, have it figure out when I'm not looking at it and only train on examples where it was clear I was reviewing
  • Fix problem with clear not working correctly with "queue" etc.
  • KBS, Org-FRDCSA-MarketingManager-Source-LinkedIn clear-context
  • KBS, MySQL2:freekbs2:default clear-context default
  • KBS2, clear-context
  • KBS, MySQL2:freekbs2:formalize clear-context
  • KBS, MySQL2:freekbs2:test-mysql2-vampire clear-context
  • KBS, MySQL2:freekbs2:test-mysql2-vampire clear-context.
  • KBS, MySQL2:freekbs2:test4 clear-context
  • evangelist is related to clear Visualize
  • systemx should use clear to verify that people have read certain books.
  • Fix teh pause mechanism of clear.
  • Have clear automatically speed up reading as the user reads more.
  • Maybe switch clear to an XML or SGML like format...
  • Have clear parse paragraphs too, and do other kinds of analyses...
  • Add a feature for clear to convert a text to mp3.
    ("comment" "103928" "or ogg")
  • Add a feature to the clear/frdcsa firefox extension for it to read the selected text aloud.
  • Get Brian using unilang so he helps develop clear.
  • I mean work on clear tomorrow.
  • Have to get clear to read info pages.
  • unilang client or the like should be able to run shell commands. Example is if clear crashes, and we want to restart it within the window, without having to restart everything.
  • Be sure to read all of Justin's writings sooner or later. Once we have a good clear KB, that knows what we are reading.
  • clear should have a function for re-reading the last paragraph.
  • clear< exit
  • clear should work for source code files too.
  • That last thing works with training materials too. Have clear functionality as one type. Improve that type system.
  • For clear, write citeseer interface, and citeseer dependency.
  • Should port the TTS component of clear to manager and have clear operate the manager::Dialog API)
  • Figure out what is wrong with clear, why it wouldn't read something like the bible even though there was no voy: RESPONSE: myfrdcsa/codebases/internal/digilib/data/gutapp/kjv10.txt > No More Documents
  • do topic mapping, based on current readings and zipper things together. clear should detect useful leads in documents and store those for later use and analysis.
  • Neat if clear could pronounce bold words stronger, etc.
  • /var/lib/myfrdcsa/codebases/internal/unilang/start audience broker clear corpus cso ELog OpenCyc pse unilang-Client
  • A basic technique clear can use is largest subsequence, by analyzing a tree of existing sentences to determine whether the given sentence has already been read. You can restrict the dictionary to make it more efficient.
  • Roger, all clear, planetalk (for all).
  • Here are my goals for tomorrow. I would like to get clear doing some basic generation, but I would most like to focus on getting it so I can knock tasks off the task system.
  • Create web browsing score for how much communication is being done. Model web reading using clear topic visualizer. Penalize bad words, etc.
  • /var/lib/myfrdcsa/codebases/internal/boss/perl-codebases/clear/Feedback.pm:use perllib::UI; contains the desired feedback code for clear.
  • clear can talk to broker then.
  • I am thinking I would like to work on clear some, to get the clear things ready. Although, more important would be the schedule maker.
  • clear should also model a person's vocabulary (useful for foreign language acquisition, etc.)
  • clear should compute an enormous meta-reading model consisting of all readable text on the drive.
  • clear should consult all/Termios/picform/bard for understanding writings.
  • clear should speed up gradually over time if wanted.
  • Read Emacs lisp manual in full w/ quizzes. Expand clear quiz functionality. Use named entity recognition, not just keywords, etc.
  • Have clear read man pages too. Have a whole priority system.
  • I should read all the linux howtos now. Write something simple today, that makes sure clear is in basically constant utilization. This will of course include writing a compressor for voy files.
  • Adding support for info to clear.
  • Get everyone using clear and get quiz functionality working. We will also have a site to submit good reading materials to.
  • Usually when clear won't read it is because it thinks it isn't text. Have it default to reading it.
  • clear should handle pdf.gz
  • The fundamental clear datamodel has to change to incorporate things like paragraphs.
  • clear should be able to read man pages, perhaps simply by running "man "
  • clear should read the name of any new files it starts reading.
  • Obviously clear should be mutlilingual
  • The problem with using the /tmp/clear method is that we don't know what is being read, should come up with something akin to the website stuff.
  • We can start up a clear server that reads books to people over the network, that way we don't have to distribute our files, and we can also record what everyone is reading.
  • We need a planner for clear that plans what to read to you and when.
  • Fix this damn problem with clear not reading things that are downloaded.
  • Add per-paragraph and per-chapter jump hotkeys to clear.
  • clear y
  • Set the path for clear data to prevent it from putting it anywhere.
  • Or they could use a standard length like "AUDIENC: unilang: clear--: BUSROUT: etc
  • clear can adapt to the user's changing information wants, for instance one sec they might be interested in such and such and the next, something else.
  • clear should do code detection, etc.
  • Here are some ideas about clear, obviously it needs to analyze whether you've already read crap.
  • Fix the clear bug that makes it jumble words together.
  • Need to find some way to make clear run smoother.
  • clear should be able to jump back 5 lines...
  • clear could use language detection software to do this.
  • Add a speech interface to clear.
  • do accessory tasks to work, like reading w/ clear, requires less expense...
  • clear i
  • Solve problem with unilang-client quitting when using clear.el stuff
  • Switch clear to Icalendar times.
  • For clear, write citeseer interface, and citeseer dependency.
  • clear should have a function for re-reading the last paragraph.
  • clear should speed up gradually over time if wanted.
  • Get a 3d card and create a cool system for clear that uses web cam based eye-tracking and 3d graphs of content areas and images (images could be pulled from documents when appropriate and displayed in the knowledge display.)
  • clear should also model a person's vocabulary (useful for foreign language acquisition, etc.)
  • Figure out what is wrong with clear, why it wouldn't read something like the bible even though there was no voy: RESPONSE: myfrdcsa/codebases/internal/digilib/data/gutapp/kjv10.txt > No More Documents
  • Can start and stop clear from mamager.
  • Have to get clear to read info pages.
  • Can apply clear topic jisting to unilang messages.
  • In order to have searching through what clear has read, set up an information retrieval system.
  • I haven't been reading much with clear lately and it shows, since I was barely able to understand it.
  • I mean work on clear tomorrow.
  • The fundamental clear datamodel has to change to incorporate things like paragraphs.
  • clear
  • Create an interest model for the user that helps clear know what to read.
  • My total was $148.52 and the clear said your change is 1.48, to which I said, no, that's what I just gave you!
  • clear should handle pdf.gz
  • Neat if clear could pronounce bold words stronger, etc.
  • Adding support for info to clear.
  • Integrate facetracker with clear if facetracker works well
  • If we want to have clear do this, we can use pitch to emphasize certain words.
  • Therefore top priority is getting clear working, so that it comes up with a read list.
  • clear should work for source code files too.
  • clear should determine how useful reading a document is based on the amount of mathematics, programs, tables and pictures, i.e. things that aren't useful in a serialized reading way.
  • Have a facility for reporting texts read visually to clear.
  • clear can use link grammar parser to determine whether upcoming sentences are problematic (i.e. not really sentence data).
  • clear should have support for multiple pedagogical modes.
  • clear did not handle a document with spaces and commas correctly.
  • clear should have functionality for training people up to speed.
  • Mom wants to have a clear subsystem that explains things to her.
  • clear should correct things like pro-\nvides
  • Add something so clear can be launched from opera or whatever
  • Obviously clear should use the Emacs wikibrowser
  • Add support to clear to support clear, y
  • clear should support rereading.
  • clear should incorporate a way to get ratings for books used.
  • clear should detect examples and have the user work through them.
  • Chris points out that MediaWiki tracks who views which pages, and you can use this with clear as a module to determine who has read what pages (actually itsn't only seen whcih pages, make the inference).
  • clear should be able to read code properly.
  • The manager::Records::Session system can zoom to the nearest complete clear screen in order to begin to reconstruct a given location.
  • Catalyst provides certain abilities that would be very nice for clear.
  • Write my own gaelic language text to speech module, add it to clear, that wraps text in gaelic into a form that is usable by a custom dictionary using the English language models.
  • Take important todo items, rewrite as questions, search google and read results throguh clear.
  • Fix bug with clear are you paying attention.
  • This should be used for clear.
  • clear should have the ability to summarize what the user has read and what their areas of expertise are based on that reading.
  • Redo things like clear queue into the freekbs format and have that processing take care of it, that way, we have a coordinated way to handle processing.
  • If you see 'Not enough arguments for format string', just clear the freekbs-stack.
  • This thing I've made for Sorcerer sure looks useful, it is being used in Sorcerer, broker, should be used in job-search, and could be used in busroute, clear, corpus, critic, cso, digilib, and verber.
  • clear can talk to broker then.
  • /var/lib/myfrdcsa/codebases/internal/boss/perl-codebases/clear/Feedback.pm:use perllib::UI; contains the desired feedback code for clear.
  • /var/lib/myfrdcsa/codebases/internal/unilang/start audience broker clear corpus cso ELog OpenCyc pse unilang-Client
  • /var/lib/myfrdcsa/codebases/internal/unilang/start audience broker clear corpus cso ELog manager OpenCyc pse unilang-Client
  • Read me things in the background through clear, like audience emails.
  • Roger, all clear, planetalk (for all).
  • clear should consult all/Termios/picform/bard for understanding writings.
  • Create a clear.el function for reading all selected files in dired.
  • clear should compute an enormous meta-reading model consisting of all readable text on the drive.
  • Add auto-detection of when an html page needs body text extraction to clear, along with prompting the user
  • Use clear to read licenses to web services and software.
  • clear should be able to select music on the basis of the mood of text... and provide a soundtrack...
  • clear should better handle re-reading stuff, than just deleting the existing voy...
  • Fix problem that prevents playing all clear queued links
    ("completed" "77973")
  • Integrate perllib::BodyTextExtractor with clear to remove junk from html pages when reading them.
  • Fix the problems with clear-queue-all-links
  • Figure out how to listen to clear when using Justin.
    ("solution" "39905" "obviously clear must be installed on Justin")
    ("completed" "39905")
    ("solution" "39905" "Just git clone and set up another frdcsa instance on justin")
  • Figure out a way to listen to emails using clear even when on Justin.
  • Note that my software, in using names like /tmp/clear.txt is liable to have data stolen. Change these names.
  • Move clear rssreader to a better namespace
  • Fix bug with number changing after asking "Are you listening" in clear.
  • Obviously we want to change the date format in clear from the silly date stamp to a microseconds since so and so.
  • Allow q to work on clear.
  • The sltion to moving my stuff into CVS or subversion is clear. Simply start a new dedicated project server (with more space).
  • Also, I need to finish the clear package. I do not need to create a live CD, unlike my previous thought just now. I have only to release a package of it.
  • Certain tasks need to be done for next years computing workshop. We should brainstorm about what we can do to help improve certain things. I didn't do any real lesson planning due to time and effort constraints and this was my main difficulty. I think clear has applications of course, but to be an ITS its going to have to know more about learning styles than it presently seems to.
  • Some features to add are as follows. clear can propose a list of things that the reader may have read and the user select this. In other words, there is a system for selecting which books, websites articles etc have already been read. There would be ways to determine whether a person was familiar with a particular text simply by observing and testing their knwoledge.
  • Just thought that if you aren't paying attention to clear you could say so in such a way that would cause it to repeat the last read section.
  • Have to get clear.el loading
  • Some ideas here, if the user is using clear, it should start by trying more frequently to question whether they are listening at the beginning, and then relax this somewhat as they demonstrate that they are there. This is excepted when it is night time.
  • Here are some various ideas for clear. After determining that the user is asleep, or has not read the book for a long time at any rate, don't restart at current line, restart earlier.
  • Add are you awake feature to clear, every now and then it asks you, are you awake? You just have to hit enter or something. If it times out after asking a few times, it pauses until c is pressed after all, if a any key is hit it restarts the are you awake question, or Are you paying attention?
  • clear can use summarization software to summarize a document and read that summary.
  • DocTrain (as in doctrine and Document Training) is a good name for clear or a component thereof.
  • clear can make documents readable within Emacs. Perhaps it can generate a buffer in emacs, and then follow the point along the document as it reads it. By moving the cursor, you could start it reading elsewhere. It could highlight what it has already read to you.
  • clear can use bibtex entries and USBN book number or whatever. Some dependency information can probably be extracted from Research papers by citations.
  • clear should also record which movies you've seen.
  • clear should be able to determine whether a given book is below your level.
  • clear can also almost behave like a hypertext reader - but linking over conceptual connections.
  • My todo list for today probably involves adding features to clear, and then packaging it
  • For clear, we should have the ability to jump to the next match, so I don't have to guess the line number to sync to a book page.
  • I sense the need for a new system - clear determines whether you have read the appropriate stuff - but have you acted on this knowledge? There is a need for a "project manager" that will help with hands on development of various systems. This ties in nicely with our project since we could use this to get people resume valuable experience as well as build us systems we need.
  • That's one thing is I haven't been studying formal models enough lately. Like this PI calculus stuff looks interesting. We want to map it out and use clear to ensure familiarity. Another thing we could do is spider citeseer again. Perhaps slowly over time.
  • clear can use the students to some extent to point out whether books have overlapping material.
  • clear should be able to read man pages, perhaps simply by running "man <programname>"
  • Could set up "clear radio", stupid but possible.
  • clear should be able to read man pages, perhaps simply by running "man "
  • clear should select questions that the user is not likely to have known before reading a text, using a variety of methods. For instance, if the user has read certain books, don't use questions are answered in those books.
  • clear should be able to record the time's that various people spend reading a given text.
  • AI is clear as a bell to me.
  • clear can attest to the user having read certain documents, and by association with clear's subject matter indexing, we can at least come up with ideas for the user's skills profile, which can then be translated into a resume, or what not.
  • Add the ability to spider with www::mechanize to clear when wget fails.
  • clear should print a nice list of books titles/references which someone has read.
  • clear should have the ability to read a website "Cover to cover", for instance: http://scheinman.org/
  • clear should categorize resources.
  • clear should allow multiple user profiles.
  • should set up pbuilder and also get clear packaged
  • clear needs to record all voys centrally and compressed
  • Note that with the dependency information for educational materials all in order - we could begin classifying the knowledge axiomatically for two purposes - first, so that we can program systems that reference and make use of this information, and secondly, so that we can use this for clear. We can reference works by ISDN, etc, and use that among other things to begin indexing books. That way - we have an unambiguous identifier.
  • Package clear by the end of the week - and get all of its dependencies in Pbuilder.
  • clear should adjust speed automatically, for differing types of input, etc.
  • Add features to save the voy files in clear's repository if the directory isn't writable.
  • Add features to clear to enable it to "spider" websites that use "next".
  • clear should scale to user's speed.
  • Second is that clear's quizzes should adapt themselves to determine the users ability
  • Various thoughts there. One is that we can have clear add a book to the classification hierarchy whenever it reads it.
  • Areas for improvement with corpus. It's quite clear. The procedure I came up with isn't efficient but it is easily adaptable to an efficient procedure. Two things - don't do n^2, rather only look at reasonably related documents - this will cut down dramatically on run time. Second, make more efficient the sharedterm code, and lastly profile to see where spending all time.
  • clear needs to take into account in voy files, which user read it. It should probably keep a central voy file repository.
  • Note that even though clear did not successfully parse the ps file, pstotext succeeded. And also note that when I manually converted it it appeared as data and so was not run.
  • Should package clear.
  • Need to create new software that models Ryan's interaction, in addition to an RSR module. In other words, he might develop coping strategies if I am not adaptive to his needs. For instance, he said he is not interested in reading. I should show him clear and see what he thinks. I should record his reaction, and base decisions off this. In fact, software to do this would be quite simple. So that is what I should do. Maybe I should write it by tomorrow.
  • I just thought of a way that clear could read through hyperlinked books. It could spider them and determine the order from the order of the links.
  • Some various functions we might implement - clear (read buffer), or clear read man page.
  • Getting clear working is important for our allies.
  • Need to assemble my digital library. Also, need to put in rating system so that clear automatically builds the library from what I am reading.
  • coauthor's dependency system can be used with clear in order to build mental models. Clearly, coauthor and clear are very closely related, which is very funny since I didn't even detect that at first. We need to get the coauthor system's ability to generate dependencies between written materials working so that - 1 - I can generate a doctrinal hierarchy for people to test proficiency in.
  • Should add mental model to clear to determine whether analysts should be reading particular items.
  • Mention that clear now reads RSS feeds.
  • clear should have facilities for letting the user know how much further they can expect to work with their current book.
  • Fix g 0 bug with clear
  • clear the sleep queue before reading aloud - so that accidental key strokes from night down hurt it
  • Here are some important tasks - find good C/C++ books and begin reading them using clear.
  • This is much nicer since clear doesn't do a bunch of BS here.
  • Add a feature to clear to skip able of contents.
  • Add speed control to clear.
  • Note that Justin's friend Aaron was interested in the clear system, for reading. This means he can provide us books that we should read, at least for understanding how to go about doing things. As for us, we should always keep multiple levels of books, those that we tell non-members to read, and those that we read ourselves. Perhaps we should invent a language for our own communication.
  • clear should quiz the user for an after report of any document it has read, and if the user is unavailable, queue the review.
  • Need to back my stuff up onto DVD. Nothing fancy, just clear out some of the directories. Build index of all DVDs.
  • clear should identify mnemonics and drill them into the user.
  • Remote control for clear.
  • clear should have breaks.
  • Annotation and bookmarks for clear
  • clear should prompt to determine whether person is listening to it.
  • I need to configure clear to read the emacs manual.
  • Put planet.debian.net on clear.
  • Send arpradha@andrew.cmu.edu information about clear.
  • It is worth noting that, after hearing David (Levitt), and to a lesser extent, Guy (Goldstein), discuss zionism with Tom Cauchois, it is clear from listening to them that they can be controlled, more or less, by skillful PsyOp manipulation of the symbols and themes of zionism.
  • Have clear monitor reading better by writing squid element and use this to enforce training.
  • check out what is wrong with clear, why it wont read urls after downloading them.
  • In order to get clear working, I'm going to have to do certain things like ensure it has all of its run time dependencies and I'm not interested in that.
  • Be sure to post clear on freshmeat, and sourceforge?
  • The group building system. The group building system is a protocol for groups of people to become large problem solving teams. The protocol is as follows. It is clear that if two people are working on contradicting or inconsistent goals, they are not likely to satisfy them - at best only one of them will be satisfied. Therefore, this algorithm dictates that whenever two groups of people meet, they determine which group will be dominant. The goals of the second group are subordinated and integrated into the goaling system of the first group, but the command structure, in the form of a tree, is retained. In this way, large groups can form that solve many problems. Although the individuals at the lower ends will not (seemingly) benefit as quickly, the fact is that as a group they will be far stronger and will thus achieve more of their goals than if they were not to benefit as a group.
  • could configure clear to use kbfs and then in this way, it would be capable of correctly processing documents (as in identification via metadata)
  • check whether clear.sourceforge.net is taken
  • Make open source certifications based on this clear technology.
  • Develop the ability to automatically infer plans based on statements. For instance in the above statement: <<<Develop radar architecture using Rational architect.>>>, it is clear to us that at least two tasks have been defined. Develop the radar architecture. and Download and install Rational architect
  • Equip clear with the ability to walk websites.
  • Develop the ability to automatically infer plans based on statements. For instance in the above statement: <<radar architecture using Rational architect.>>>, it is clear to us that at least two tasks have been defined. Develop the radar architecture. and Download and install Rational architect
  • clear should use Citeseer somehow.
  • Have clear send commands to festival to speed it up or slow it down based on novelty estimates.
  • Automatically determine if I have read a webpage or not. Feed squid through clear.
  • clear< s
  • clear< s
  • As radar finds systems that interest the user, it presents these systems as though in a flight simulator - perhaps using critic, and feeds documents that should be read to clear
  • clear< p
  • Naturally clear should provide the ability to search backwards through what has been read, in order to answer questions and so forth. For instance if I read something to the effect that this truth is as clear as the sun, searching for this should get me the results I want. This reinforces the notion that clairvoyance should keep a central kbfs implementation of voy files, and run against that, as these can then easily be equipped for searching by a common search interface.
  • clear< p
  • Also, we want to change the library directory of clear to be under /var/lib/clear/library or something, or even kbfs
  • useful for ITS, clear http://freshmeat.net/releases/183336/
  • A major problem is that with systems like critic and clear we still wind up deleting the information because they either lose information or are not sensitive enough, respectively.
  • Voy files ought to be considered to be kept locally, along with file signitures, and with software to detect similarity (to prevent same document from being read and reread.). This will encourage the creation of a digital library. Consider renaming clairvoyance to clear. Is everything clear? Are we clear on that? This should be /var/lib/clairvoyance/voys.
  • (Almost clear proof that I'm not acting as intelligently as before I got this job, I accidentally rewrote a section of code for matching descriptions to Debian packages, and my latest version performed at least ten times worse in terms of accuracy, if not more.)
  • Here is something to consider. Looking at the way these things operate, it is clear that so called strategic - i.e. non-technical facts about communities is fundamental to security plans. Therefore someone wishing to improve security need only institute certain barriers in the communities behaviour itself. For instance, promote certain behaviours. This is evident in my interactions at home, whereby a preprogrammed set of rules took over and what we saw happened.
  • F-16 pilot: Finds snake, drops two CBU-87 cluster bombs, and misses snake target, but gets direct hit on Embassy 100 KM East of snake due to weather (Too Hot also Too Cold, Was clear but too overcast, Too dry with Rain, Unlimited ceiling with low cloud cover etc.) Claims that purchasing multi-million dollar, high-tech snake-killing device will enable it in the future to kill all snakes and achieve a revolution in military affairs.
  • Let me make myself absolutely clear, under no circumstances will I tolerate the anything but immediate progress towards AI. I was not bnron, and given a mind, only to let other peple without clear understandings or plans affect the course of my progress on solving problems that affect people, and in other words, I will in no circumstance operate under bounded rationality.
  • it is absolutely clear that no appeal whatsoever is ever likely to gain the favor of anyone from home, etc.
  • Dad has done less than to help again. EVidence of this is that I said accidentally that "when I die", indiciated regression to earlier patterns of thought. Now, some things are possible now that weren't before, perhaps as a result of this design of mine, but I don't think it has gone as well. Although, the money from parents is useful in getting what I need. But this illusrtates a flaw in my thinking. I should have asked Al for some money to get all set up. But I think I am pretty much in the clear. I should also call and get my paycheck. I'll ask Al whether this may be used to finance my outstanding needs.
  • liek to keep my mind as clear as possible -dad
  • Take parents to see the closed building so they get a sense with what they can help with. Make it clear to them that I am not going on if AI doesn't happen.

This page is part of the FWeb package.
Last updated Sat Oct 26 16:50:52 EDT 2019 .