« February 2003 | Main | May 2003 »

Intelligent media: FOAF my shiny metal ICBM

I've been working on my FOAF file for a couple of days now.  Thanks to observations and commentary from #RDFig and #foaf, I'm now advertising both my nearest airport and my current location using the right namespace.  The FOAF explorer found the geo tags, and is now showing the nice green button for each of the locations specified.  Yay!

The links for this have been added to the syndication sidebar, which also depicts the RSS 1.0 and 2.0 feed links; the 1.0 feed is significantly enhanced beyond the standard template version.

The blog's also been tagged with the ICBM coordinates of my hosting provider, so you can discover websites within a few miles.  I see some; time to make new friends.

I'm open to further suggestions as to things that would be useful to add to any of the feeds; feel free to leave comments, as well.

Update: Turning the archives HTML page into an RDF feed was a brilliant suggestions; thanks!  RSS 1.0 Archives.  Apparently this is useful to anotation?  The entire content of my blog is in that feed, along with comments, foaf, and trackback data.

Cool.  Please don't subscribe to it :)


Reminder: Sharing music is for personal use only.

Do not show this message again

Rendezvous sharing, courtesy of iTunes 4.

Opinionated linking: Adding opinion metadata to hyperlinks

No, I'm not putting in a link to Orlowski's piece. You can get there through Jonathan. I don't want to reward Orlowski by pushing him up the blog rank. (This is why we need Kevin Marks'"vote" attribute that lets us specify that although we're linking to something, we don't like it.)
Stating disagreement could be done via several methods; I think a combination of two of these methods is the best route, and I'll state that route below. Following the policy of "practice what you preach", I'll be implementing these gradually here as well.

Two notes. First, I'm not focusing on disagreement; being able to say "i+agree" and "i+disagree" is the goal. Second, non-opinionated links should be interpreted as indicating interest, without opinion; if desired, we can explicitly state "neutrality".

Here's a link to the article in question, with the disagreement indicated in the URI and CSS.

Non-standard "opinion" attribute for A tags
<a href="..." opinion="negative">

This allows us to mark any link with an opinion; using a limited set of keywords ("negative", "positive"), we can indicate our opinion of a given link.

This method has critical flaws: "opinion" is not a valid (X)HTML attribute, ranking systems (like Daypop) discard the attribute, and the "opinion" is not presented to the either the user or the linked-to site (more on this below) -- they have to dig into HTTP referrers and into the HTML source to find out the opinions of a given link.

CSS "opinion" classes for A tags
A.negative { color: red; }
<a href="..." class="negative">

This allows us to mark any link with an opinion; using a set of keywords defined in CSS ("negative", "positive"), we can indicate our opinion of a given link.

This method resolves one of the above flaws -- the visibility of the opinion to the person clicking the link. Using CSS, we can indicate positive links with a blue solid underline, negative links with a red dashed underline, neutral links with a black solid underline. Whatever.

Unfortunately, it doesn't solve any of the previously listed problems of the "opinion" attribute; it's still not indexed by engines, and the linked-to site has to do context discovery.

Locations that aren't in the page
<a href="...#opinion=positive">
<a href="...#opinion=negative?...&...">

This allows us to mark any link that doesn't already have a #location defined with an opinion.

There's a couple disadvantages: the opinion won't show up in the server logs, and the user might not see the opinion in the address bar. One strong advantage comes to mind, though; I'm reasonably certain Google's rankings treat in-page links as distinct weights, given the advent of single-page blogger archives.

Extra CGI parameters
<a href="...?opinion=positive">
<a href="...?...&opinion=negative">

This allows us to mark any link with an opinion; using a CGI parameter, we can indicate our opinion of the link.

This method resolves one of the above flaws -- the visibility of the opinion to the linked-to site. Properly configured, they can break down the statistics of incoming links by opinion -- and perhaps The Register would understand more clearly that Orlowski isn't popular -- to the contrary, he's reviled. That he's on Daypop because we're angry with him, not because what he writes is well-liked.

A note about ranking engines is in order. I presume herein that ranking engines will not strip the CGI parameters from the link, as doing so might break perfectly valid links that require other parameters. As almost every link I've seen has no CGI parameters, I don't have much real-world data to evidence this one way or the other, but if the popularity of this idea rises enough, perhaps they'll be influenced!

Putting it all together
<a href="...?i+disagree" class="positive">

Combining the CSS method with the CGI parameter method, we can implement a solution for opinionated linking, in a form that evidences the opinion of the link to Daypop, the person clicking the link, and the linked-to site. The CGI parameter serves to distinguish the link from other links to the same page, with differing opinions. As an opinionated link passes from blog to blog, its index on ranking sites such as Technorati will rise independent from the same link with a different opinion (or none at all).

Supporting XHTML: my:opinion
<a class="positive" ... xmlns:my="http://semantic.web/link/opinions" my:opinion="positive" >

torstens provided a nice way to add XHTML support to this; a namespace for opinion. I've integrated it into the master example up top, and I recommend its use highly in conjunction with the CSS and CGI query.

Continue reading "Opinionated linking: Adding opinion metadata to hyperlinks" »

Identity surfing, pseudonymity, reputation

Real reputation is emotional. The penalty for identity surfing is and should be leaving your reputation (friends) behind.



Truly anonymous identities, by design, cannot have reputation; there's nothing to distinguish one anonymous person from another one, thus reputation cannot result.  A very good example is Slashdot's "Anonymous Coward"; by choosing to remain anonymous, you evidence no identity whatsoever, remaining truly anonymous.  You can also choose to evidence a name, url, email, and/or Slashdot account -- however, the option to choose none remains.

There's another kind of identity, a "pseudo-anonymous" (?) identity.  It retains the aspect of anonymity that shields the participant(s) in that identity's communication from identification, while allowing the identity to be distinguished from other identities.

anon.penet.fi used to give out these identities; many Usenet posters developed relationships with others that were filtered through a pseudonym like "an256748@anon.penet.fi"; other anonymous remailers now do the same.

Reputation doesn't port to truly anonymous systems easily; if you're truly anonymous, you have no identifiable attributes -- including reputation.  This is both bad and good; you can't tell if the person who's posting now is any better than the one before, but they're secure from the prying eyes of whoever.  With a pseudo-anonymous identity, though, reputation asserts itself.

Given a uniquely-identifying characteristic, reputation seems to be a pervasive feature of social networks; either in the minds of the participants (the perl5-porters mailing list), or by a vaguely descriptive adjective (Slashdot's "karma" ranking), or by a numeric point ranking (eBay's "reputation" ranking); examples abound.

Creative domain licensing


Someone asked me if I knew of a creative commons no-copyright license; I wasn't able to give them an answer at the time, but I've followed up on it since.  Whoever asked, hope you're listening :)

It appears that the license I was looking for was the Creative Commons public domain.  It's not part of the radio button section of the license selection page, but if you look off to the right a bit, there's a link to it.

Apparently they keep it seperate because it's so distinctly different from the rest of the licenses; I suggested they add, say, "(that is, give up all copyright)" or somesuch to the page.

For the record, my content here at crystalflame is covered by the attribution license; I'm not much of a fan of the GPL-like viral sharing clause, though.

Semantic breakfast

These notes are taken from a breakfast discussion on semantic indexing with Ben Hammersley, and several other people. It's pretty haphazard right now; I'll clean it up someday.

If you're here from the Mailing List Bots discussion, searching for "metamantic web" ought to find the relevant section of the notes. This'd be the first google'able use of that phrase, too.

- R.

??? -- http://www.nitle.org/semantic_search.php

"surface occlusion, with a matrix of words" -- ...

very once in a while, after a sniff, he would actually read the letter and take a few notes. It turned out the scholar was studying the spread of cholera in the early American colonies. Letters posted from villages with cholera were disinfected with vinegar

-- http://stacks.msnbc.com/news/765360.asp

"prototype theory" -- the quickest example (good to know if you're thinking about indexes) -- the best example of the failure of ontologies is the platypus. Fundamentally, people do full categorization; they don't use abstractions and instances, they use examples and instances. The current ontology of animals is centered around animals. ... you have to make cleaving lines

prototype theory says "this is a clump -- my thing is like that clump, but it has extra features". the main problem is that you still need a type system. what you do is you have voluntary typing -- you don't have an abstract. you create the class by putting things into it. --- connects to Twingle! -- the "prototype theory" is a formal discipline for looking at that in a different way. I kindof look at things from an OOP pattern, and didn't

-- http://hook.org/anselm/thirdparty/categories.htm

trying to categorize posts -- semantic tower of babel. a system that became a MT plugin --morelikethisfromothers -- it goes out to anybody who links to your post -- goes back to the site that's linking to you -- finds the linking post, and looks to anything which is on that blog is -- the author by dint of linking creates

-- metaphor theory by george _______

"society of mind", people using metaphors to bridge from one domain to another. it has a problem, in that people don't create fine enough categories "metamantics" -- problem with morelikethisfromothers, for any post that's in my category, what is the relevance of its connection to your categories?

"metamantitian" or "metamanticist". TBL always talks about the semantic web -- maybe we should call benhammersley the "metamantic web" [quietly propogate this meme, credit to andy@hook.org, zero google hits as of]

"curated collections" -- we don't have time, too much stuff, too many ontologies.

-- http://www.soundratings.com -- audioscrobbler

Continue reading "Semantic breakfast" »

Free all radios for great justice!

GNU radio is a free software toolkit for building and deploying software radios; for learning about DSP and communication systems; for creating new kinds of radios, modulations, protocols, development environments; regulatory hacking; exploring.

I'm in it especially for the new kinds of modulations, by the way.


Software radio is:  Get the software close to the antenna; turn hardware problems into software problems; defines the waveforms transmitted; demodulates the waveforms received.

The equipment is off-the-shelf hardware.  Conventional radio stuff?  Spectrum monitoring?  Multi channel, mode, morph on the fly, better spectrum utilization, cognitive radio.

With a software model, you have one RF frontend and you suck down an entire spectrum of broadband; it lets you do distributed, parallel multiprocessing of an entire range of frequencies.

AI!  Smart radio detection.

What kind of hardware?  Narrow:  Soundblaster, narrow band RF front end (NFM, USB/LSB, AM, ...); audio output from receiver (packet, PSK31, ...); sub 50 kilohertz.  Wide band: high speed A/D (20M+ samples/sec), wide band RF front end (TV tuner).

[time passes]

Politics: MPAA / CPTWG / BPDG "Broadcast Flag".  FCC software radio in general; FCC free software radio.

Mapserver grass

We're the pitchforks and torches for the smart mobs.

Each map has a purpose; each map tells a story. We need maps that tell the relevant stories.

We want to be able to tell stories about things that happen in a variety of different places. If you want your story to be heard, you have to tell it in the right way.

We're mapping the wireless revolution! One of the things we're trying to do is build a fixed-point wireless network out in the hills (and the trees) of 802.11b. It has line of sight requirement.

[the nocat explanation was awesome]

802.11b requires line of sight; who can see whom? If you can see it, you can shoot it.

Who can see whom? We'll plot links by hand, get elevation contours, etc. The numbers are insane, when you scale to 100 people. We're trying to bring people together, to promote self-reliance. Why do we need a multionational corporation to connect a community together? Communities gather around their stories.

We were trying to get Mapster to do its thing, and it just wasn't being easy. [What blue lines?] The big GIS map just isn't our map! It obscures what we want to see. [Can you see the blue lines?]

This is Rich's story, but it doesn't quite go far enough. [GPS route-maps].

We want systems for the management, acquisition, and presentation of geospatial information.

* Acquiring and exploring geospatial information * Creating a narrative

Key point: Maps are a narrative.

As the weblog phenomena demonstrates, everyone has something to say. We'd like to enable people to tell their story (and enable us to tell ours) using maps.

There's one similarity between them all: Stuff, attributes, locations.

Open source software; freely available data sources. We resisted going with the super-topo software because we couldn't just give them away.

We turned to GIS. [diversion: I have a tendency to reinvent the wheel from first principles]

Layers; data sources; coordinate systems; projections. We need to project a spherical surface onto a flat plane. The compromises you make change your maps; when you try and exchange data, you havce to mkake sure that you're coordinating the tradeofss properly. [note: suggest comparison to Photoshop, in addition to transparencies and hat suggestion you made? SimCity! demonstrate, with a running simcity instance]

Data sources.

* TIGER. Collected by the US census beaurus. Includes rivers, political boundaries, zip code boundaries, etc. Available on the website; often inaccurate, wildly incomplete.

* GNIS. Official repository of domestic geographic names information; contains 2 million physical/cultural features. Lat, long, elevation.

* Other USGS sources. DLG, DRG, Orthophoto Quads, BARD (Bay Area Regional Database). Geo-tiffs, with layers including the information. A lot of

* DEM - available from USGS, elevation at 10 or 30 meter resolution for the entire US; can be converted to contour vectors; we can calculate line of sight!

* Make your own! GPS; Netstumbler; Journals; RDF...? Emerging standards on XML style annotation of geographic information?

* TIGER Mapsurfer; URL. What's cool: Instant, attractive maps. They don't tell our story. They can't provide the source code. The test application was never completly documented. The software is missing large portions of source code, and is hardware specific.

* GRASS; Geographic Resource Analysis Support System; collection of UNIX programs with a Tcl/Tk front-end; Open Source! Does everything! Hard to use! By doctoral students, for doctoral students. Everyone who uses grass is either in the field getting paid to do this, or they're PhD's or candidates, or they're on the mailing list or developers.

* FreeGIS: cool. * OpenGIS: cool. Industry consortium devoted to open standards and freely available data. Open specs: Simple Features Specifications for SQL (geospatial database). Geography markup language. Werb Feature Services (retrieving features); Web Mapping Services (retrieving maps). * PostGIS: adds geographic objects to the PostgreSQL object-relational database. Useful as a backend spatial database for GIS applications, like Oracle's Spatial extension. Follows the OpenGIS simpl

* MapServer: it's a CGI. GIS browser. Capable of reading data from GRASS and PostGIS. Mapserver knows layers. Mapscript bindings for Perl, Python, Tcl, Guile, and Java.

WifiMaps: They appear to be based on MapServer, at www.wifimaps.com.

Personal Telco Project: Node maps via MapServer. Nodebot - an IRC resource. Reports nearest PTP node locations via IRC; uses XML-RPC to PHP MApScript. GeoWiki. Python MapScript plug-in for MoinMoin. Link to http://personaltelco.net/

NoCat Maps: Node database, geocoding, elevation profiles, http://maps.nocat.net/ -- all you need is a computer and an antenna. Demonstration: Possible links from Walter Greenwood.

Schuyler: My internet traffic travels six miles, five hops.

What it doesn't account for: Curvature of the earth, fresnel zones, ground clutter. The software can tell you when you can't, but it doesn't address buildings and trees.

Where we're at: Frustating, complex, confusing, weakly documented array of nitty gritty. Building databases is a black art. Visualization tools that allow us to play "what if" easily - the spreadsheet analogy. Web services that expose public databases and facilitate development and man-based applications.

Map supermarket prices on a geospatial map. We need public databases. Let a hundred map servers bloom! Give us ideas. What I'd like is for us to have people tell us what would be cool to do with maps.

Suggestion: GPS trace my own movement for the course of a month and see the size of the city i live in?

The Keynote formerly known as Smart Mobs

A year ago he was watching the conference remotely blogged by Cory, and decided that he had to be here watching it.

"You can create tools that amplify collective action". Stuff is happening that will continue to enclose our rights; innovate around it.

"Collective action". People choose to participate, voluntarily. Enabled by communication methods. They're enabled by literacy, by technology, by social agreements.

The Web. Enabling technology; built without constraints, enabled by and enabling collective action.

"Early signs of technocollective action". P2p. Folding@home. Wikipedia. Seattle Wireless. Social software. "Ten years from now, when we have billoions of people walking the streets, each of them wearing or carrying comptuers, with [greater processor and] with wireless bandwidth [...]" What are we going to do, when we have that much computation power available?

WiFi is an industry example of collective action. Any form of wiki, mailing list, those are tools for collaborative action; without them acting together.

"The use of mobile communications to keep those elections honest", in reference to voting in other countries than the US. Coordinating actions using texting. Moveon.org? Okay.

Power is not just about elections. I think the big power question we face: "Are we going to be consumers, or are we going to be users?"

We were treated as consumers until the PC gave us the power to take an active voice in what was formerly a passive distribution method: movies, music, etc. The media industry is clearly aware that the Internet wasn't created, and is being run from the user's end of the game, not the corporate.

Users who actively helped make the PC what it is today "actively shaped what could be done with the medium".

"... have to work for Disney, if you want to innovate, if you want to change the nature of the medium." Compromise of the end-to-end principle, routers that discriminate. "Balkanization of networks".

It's time to get involved in helping innovation. Get involved! GET INVOLVED! Feel like you don't have a voice? Contact me, I'll find a way to make you heard.

Make a human communication media, with a portion of the spectrum available for spectrum experimentation. "We're locked down with telegraph keys", limited by the legal and political restrictions to a technology that doesn't serve us as well as the ones we know we can create.

Invent micropayments! "The recording industry claims that they compensate artists". "70 million people voted with their modems: That's how they want to distribute music in the future." What none of the p2p music things have done is find a way to give consumers a fair price for the music, so music lovers can negotiate directly with the artists.

"If the recording industry and the movie industry were to go away tomorrow, the quality of music and movies availabe to us would go up immediately."

A self-organizing network is one that doesn't need to be pre-enabled by anything. There'll be billions of people communicating with devices; they need to be able to network, without interference.

"Trust mechanisms are pivotal". Trust is key. Reputation evolution?

[time passes]

Will this system remain open? Will we retain the ability to mark up anything?

Proctor & Gamble is going to be deploying RFID tags at the case & pallet level.

Think about designing whole systems: build in room for future innovators. Build in the ability for the network to grow beyond what you can think of. Link to others! Link to competitors! Find ways to enable individuals *themselves* to build new links -- and new types of links. They'll do things we can't dream of.

"Is there a default privacy switch on that technology? Can you turn it off? [...] Does it default to off? As we know, only geeks mess with defaults. [...] How can we use our technologies to seperate ourselves from technology, if that's what we want to do?"

Surveillance symmetry! How can we watch who's watching us, or prevent them from watching us? The problem isn't transparency, it's the monitoring asymmetry: you can't tell what they're doing with the information they collect about you.

Continue reading "The Keynote formerly known as Smart Mobs" »

Stupid bagels

Eating bagels with David Isenberg, Esther Dyson, Geoff Cohen. Interesting people rock. Discussion ranges topics wildly as usual; we've been through simple networking, the simplified POE server/client, etc.

Nanotech is going to need simple networking, because a billion nanobots need a way to self-assemble a network, and you can't hope for complexity to work as well as simplicity. I mentioned POE's simplified server/client TCP components, as well.

I know two Russian jokes now! Now, three! No, four! Five! AUGH!

Esther Dyson's this awesome person, friendly looking and with a cool personality; today's quote, "her mother gave her a kilo of butter". A conversational thread observes that in old Russa, it was more important to have things than money.

The bagels are good.

Blogs seem to be useful for light-to-moderate level content; useful for tracking partial thoughts, or throwing them out to the public at least. Also useful for blogging :)

Live site feedback! My real name is now actually present on my blog; before, it was only coming up in Google searches.

Talked briefly with Lisa Rein, from the Bookmobile; apparently there's talks in progress with publishers to get non-copyright-freed content published through the booikmobile. Apparently they're also working on getting children's content publishes, even through copyright would normally interfere. Woot!

The Register is talking about us, so we're apparently deserving of a "counterpoint". Interesting read. Did someone really say "The agenda is so boring that I just can't be bothered [...]"? Gah. (someone else notices this, too.)

From another thread, the Internet is a meta-community in need of a low-impact governance; overlaid on top the the existing individual-site governances. "The partipants in the Internet have been ignoring the governance of the Internet" -- Isenberg.

Who could you take in a fight? Submit answers via email or IRC!

"Let's Get Killed" is an interesting music album, made out of recordings of New York; thanks to JC Hertz from Wired, also enjoying a bagel. I already have three of his songs from other albums, all chillout. Cool.

The table changes every time I leave it and come back. Weird.

Is Rendezvous service auto-discovery and -broadcasting beneficial? Is it useful to advertise services when you're in such a large group that you can't just ensure distribution to people close by, because you're close by a thousand people. Hm.

My Photo

Recent Posts

Powered by TypePad




  • Antispam
  • Cloudmark
  • Shadows
  • Styles
  • You were here
  • floating atoll


  • Search


  • Ads