ideafile 2007

03 Jan 2022 07:26 - 17 Jun 2023 08:29
Open in Logseq
    • Idea: collect all the random good ideas I have written down in here in a central place. Hm...I my ideas seem to revolve around collecting tree/web like knowledge bases and visualzing them in some way (yeah I know, like that's original): -- visualization -- wiki change visualization -- citeseer -- evotrees -- buzzword linkages
    • memetracker (buzzeword grapher). Did this years ago, never released, whattamoron I am. -- mine online conference proceedings (just titles) and track spread of buzzwords over time and "idea space". Could also use citeseer data (easier to obtain). -- query agents (for google or others) -- left/right wings on google news -- social awareness wikipedia, where you see other people browsing...actually that's a way-old idea, whatever happend to that? Same with third-party annotation software -- never really took off. Hm. I wonder if you could somehow SIMULATE it with blogs and google-backlinks. Hey yeah, you could probably hack a browser...wait, doesn't Alexa already DO something like this? Oh well. -- knowledge spreadsheet --
      • a programmable tool for exploring data, but a focus on interlinked symbolic rather than (or in addition to) numerical data
        • tool for managing queries and result sets (ie, the good parts of Afferent/Base/ontology tools) Papers/concepts: -- netarchy (v hierarchy, etc) -- "The mathematical structure of life itself: Christopher Alexander" (essay idea). -- "Recursion as the basis of morality" -- somehting along the Minsky MMM paper. Misc brainfarts: -- my own take on using infoviz in education -- you build YOUR OWN representation of some knowledge space, link it with others. Note that this is currently getting trendy, sort of (ethnoclassification?). -- "Project idea: building a higher-level rep language over OWL-like stuff (that is, something that can do at minimal nonmonotonic types of things like exceptions!). This may get the benefit of standardization with the power of better languages." -- LifePod. Give every child an ipod like device that records their life as much as possible (ie, you could record gps and full audio. The translate all the audio to text and make it searchable, then I ould find out what the hell was up with Glenn Terry, if anything. -- Idea for a website: "where do you rank?" -- you put in your income and other data, and also (optionally) your ranking factors -- how important income, assets, education, etc are to you -- and it shows where you rank in the US, the world, your locale... -- Random idea of the moment: playing with a spreadsheet with ben, tryingt to find perfect numbers. Of course it should be lisp-programmable, but aside from that, I thought it would be nice if you could simply format a numerical field as a vertical line whose h-displacement was proportional to its value (or swap h/v for rows; I was working with a column). This seems simple and fairly Edward Tufte-ish. It was OpenOffice so in theory I could go program this myself! Yeah right.

          -- self-organizing online reading groups. People make lists of books on their queues, then somehow (human initiative or even automatically) groups are organized, maybe filtered by criteria (age/ed/whatever) or clustered according to social network, and people can have a collective read and discussion. Perhaps link this to all the open courseware that's appearing. Alternate view: seminars. I want a group of people to discuss things with.

          Random idea for a paper: "Towards Empathetic Agents" -- but rather than claiming to know how to build such things, discuss what the hell that would actually mean, the ominous idea of social agents without empathy and its forecasting by SF writers like Dick...oh I don't know. I was reading this paper "Are Agents an Answer or a Question" by Joseph Goguen, sort of a shallow overview of agent issues, and was so inspired. -- spreadsheet math (a bit of curriculum) -- physical exercise videogame, sort of an update on the old Exploratorium installation, with body tracking, but organized to get kids moving... an old one Idea for a website: build your own future. Everyone gets to slide events around on a big timeline, build up your own image of the future and compare it to others. another old one Company name of the day: Tactile Semantics Well, not punchy enough, but something like that.... touch-know feel-think Like that. OK, other idea: my usual personal knowledge organizer idea. Keep all scraps of knowledge in a DB. Gather stuff from all over (Palm, etc). User-definable XML schemas. Sharing with other people. Grow into a network. Replicate PC world (start with personal empowerment, eventually grow into a network). Nobody does this! Dave Winer looks like he'd come close.

          -- Eclipse CVS revision visualization idea. Well, turns out Eclipse already has a feature for this, but it could stand improving (in CVS View, do Show Annotation on a file). The display could be better, code could be color-coded, etc. Hm, see this: Apr 2005 Idea of the moment: any modern computer should have automatic, essentially infinite, versioned backups! Like CVS without the hassle. I can imagine an add-on that would do this (and wasn't there one for the Mac years ago) but integrating with an OS is too painful for me to contemplate doing it. Additional things you'd like: condensing to diffs to keep more versions, not having all the versions be separate files (because that will confuse other tools), labelling important versions, managing groups. While we're at it, how about keeping track of all the various versions of a file that might popup on your machine from such processes as mailing a file back and forth so it can be changed? contemplating writing an essay on anarchism, structures of cooperation, and open source...and why Symbolics was better but couldn't last due to its dependence on the centralized state (DARPA)...I admit it, Eric Raymond is right. Unix rules. I give up. Of course, I could write such a thing but its not clear anybody would care. Idea of the moment: link geocache data to Google Maps. Surprised nobody has done that yet. Jan 2005 Idea: at a meeting, everyone has laptops. There should be a screen on the back of their laptop cover showing "stuff I want to share". Everyone's plugged into the projector and can take it over according to some protocol. Another idea (this was old, but seems to have disappeared): a pagewatcher service. Simple enough that there are many people doing it: Idea for a web-site-section: QUESTIONS:
    • how do snowflakes maintain their symmetry?
    • how and why do plants exhibit fibonacci patterns (I found the answer to this one)
    • how does a GPS unit figure out distance from a satellite? Aug 2004 Idea for essay: "Recursion as the basis of morality" -- somehting along the Minsky MMM paper. Oh good, here's some nutcases with similar ideas: Vague idea of the moment: the left can be divided into the rationalist (liberal) and spiritual (socialist) segments, sort of. Liberalism, classically, doesn't do much for the collective soul since it emphasizes individual self-realization. Socialism is more a religion-replacement, with a strong sense of the importance of the collective. We need something more like socialism, if not exactly that. note connection with other religion-spirit-mumbo-jumbo thoughts Oct 2004 Random idea for a paper: "Towards Empathetic Agents" -- but rather than claiming to know how to build such things, discuss what the hell that would actually mean, the ominous idea of social agents without empathy and its forecasting by SF writers like Dick...oh I don't know. I was reading this paper "Are Agents an Answer or a Question" by Joseph Goguen, sort of a shallow overview of agent issues, and was so inspired. Here's an idea from meeting with Shrager: DO google-news with a left-panel and right-panel, which contain ideological spins on the stories. Sort of Warren-Sack-ish. Wonder what he's up to.
      -- There are OWL files popping up (Swoogl will find them) and it would be nice to have a web page that could browse them...right now I think you have to download them and open Protoge, which is a pain. Oh duh, here: nice in a way, but doesn't even show you a hierarchy...

      - 1/15/07 LinkBack (actually implemented; see ubu/project).

      1/15/07 1) record all browsed pages, upload urls to a server 2) server takes these urls, gets the pages, analyzes them for content (simple form: use yahoo's phrase extractor service). 3) present interesting clusters to the users OR -- delicious is nice but its use overhead is too high, i never know when to bookmark something and I rarely go back to my bookmarks. I want it automated. I want something keeping track of what I'm interested in. Mid-range -- have a one-click "interesting" button like stumbleupon.

      1/15/07 Something that automatically tracks all your blog comments for various platforms, so you can keep track of ongoing conversations (maybe have notifications) and also lets you present your convesation profile to the world.
      3/20/07 Everything I write in my notes files should be automatically tagged with date (and, ideally, the reason I was writing it...or the context in some sense). Freebase:
    • screenscrape-by-example
    • utility to turn MQL queries into form-based pages
      • and then a Afferent-like builder for MQL queries. 3/22 Really need a PROGRAMMABLE collaborative, shared-world, game-like environment for kids. Still isn't one there as far as I know. Probably easiest to graft a programming (by example?) onto some existing world, like Second Life or Croquet... 3/23 [I'm getting hyperized by a couple of positive movements on the job front] Lickider! Intelligence amplification! Sharpreader is good but not good enough. How about something that could easily monitor a 1000 feeds, let you cluster and search things, persist things, easily see conversation networks etc....and somehow be linked to your own output, blog or notebook or whatever. Yeah! Some small concrete steps to improve on SharpReader:
      • automatic offering to bring in new feeds
      • firefox as browser
      • lucene or some other local text search engine
      • show you a list of what you actually look at (maybe even with reading time?), somehow be self-organizing