thingster logo

What would you do if you knew?

Introduction

March 23rd 2004

We live in a world full of fascinating places and experiences that we'd like to share with each other.

Thingster lets you share information about your favorite places and experiences.

 

You can publish different kinds of things such as simple 'post it notes', events or images.

Posting something new is very easy and fast.

 

All of your posts are organized in your own area for future reference.

 

Thingster can also generate views of everbody's knowledge together.  (No image here yet).

Executive Overview

Present

Thingster is the first open-source weblogging service for locative media.  It was written by Anselm Hook, Ben Russell, Tom Longson, Brad Degraf and many others in association with Locative - a multi-disciplinary group of theorists, artists and engineers exploring the implications of attaching information to place.

Users can publish 'virtual post it notes' about any geographic location: a street intersection, a street address, a restaurant, a hiking trail or a geocache.

Speed and ease of use is a key feature.  The time from first seeing the service to making that first post can be less than a minute.

The reward or 'exit strategy' for a project like Thingster is social and environmental.  The hope is to enrich neighborhoods such that it becomes easy to discover local services at a lower cost and to create additional environmental awareness.

The hope is that people using Thingster should have a higher quality of life than people not use Thingster - they should simply be 'more fit'.

Future

This is a work in progress.  In the future this service is intended to act as a full-blown turnkey standalone community server with a strong emphasis on location and volatile and transient events. Future goals and extensions include:

Spirit

A question this project is trying to answer is this:  How can we shift our ethical foundations from one of simply protecting human life to one of protecting biodiversity as well?

4 pieces

There are four pieces that seem relevant:

  1. Our generation is shifting towards an age of 'always-on' always-logged' media where people of all ages are able to publish and share information about themselves, events that are happening, places and things of interest.  Upcoming services such as Nokia's Lifeblog and existing services such as del.icio.us are good examples of this social information sharing.
  2. 80 percent of human knowledge is about location. The way we think, the metaphors we use, our needs and values all come from our fundamental embodiment in the physical world.
  3. We are being deluged by a huge volume of digital ephemera.  Capturing, organizing and publishing digital information is hard.
  4. As a planetary civilization we are all facing frightening environmental challenges.

Knowledge Sharing

As individuals we all benefit from tools that organize knowledge; tools that help us log our discoveries and interests, that let us share information.

Emerging grass-roots tools for sharing geographic information have a particular value.  Such tools provide individuals with a way to share their own view of the world - not a corporate, orthodox or official view with all of the rough edges removed.

Grass-roots cartography - as it continues to evolve - has the potential to let individuals understand their surroundings in depth; to see the web of social, economic and environmental issues that tie their community together.

The better our local knowledge is the better our moment-to-moment decisions can be.  This includes where and how we spend money, what natural resources we know of, how we use them and what volunteer, recreational and social opportunities we pursue.

Some ideal examples:

Tools are not there yet but are improving.  As they improve individuals will have more power, awareness and hopefully wisdom.  Member communities as a whole may then wiser as well; better stewarding their resources.  A communities knowledge could be handed as a gift to each successive generation.

Environmental Challenges

The environmental challenges we face are serious.

These include issues such as global warming, depletion of fish stocks, destruction of old growth forests, water table decreases, toxicity of farmland, the destructive effects of noise pollution on animal populations, loss of species habitat, biodiversity and the like.

Buckminster Fuller recommended 'pursuing those goals that achieve the greatest good for the greatest number' as a way to set the most effective personal direction.  Here the environment is seen as a key element connecting all of the other diverse issues that surround us; it's really something that we all need to attend to urgently.

At the same time we're not going to be able to deal with this unless it is provably cheaper to respect the environment that trash it.  Any changes to our behavior have to be so clearly rewarding that they offer a real alternative to people already too busy simply trying to tread water in their daily lives.

Discontinuities

Civilizations often only ever get to face one discontinuity event; after which they are gone.  We've no memory or strategy for dealing with huge abrupt events because these events obliterate the civilization and the memory.

We already live in a world where events are unfolding too rapidly and on too large a scale to be recognized and managed by individuals. The real concern is that the entire system will crash out before we correct our behavior.

Boom and Bust

There is no particular reason why our modern civilization should be any different from say the boom and bust cycle of a petri-dish culture. In fact human social systems have crashed before - from small examples such as Rapa Nui to large examples such as the 'Dark Ages' where the Roman empire exhausted natural resources so thoroughly that it took a thousand years for the surrounding ecosystems to recover and for civilization to stir again.

Today the same pattern is at work with rapacious corporations extracting non-renewable natural resources at too fast a rate for recovery. Planetary scale ecosystems ranging from deep sea fish populations to old growth forests are crashing. Carbon dioxide levels and resultant global temperatures are following a radically disruptive curve of increase. The value of dollars is being undercut by the weakness of the underlying ecosystems.

We are within decades of a brutal end-game scenario highlighted by competition over what little resources remain followed by the spectacular collapse of first and second world interests.

Concerningly it is unclear if humanity will have future opportunities beyond this ongoing systemic collapse especially given our nuclear capabilities and propensity to war over what resources remain.  It's not simply that we're undergoing a disruptive moment.

Unifying viewpoints

There are three kinds of parties involved in environmental change:

  1. A vanishingly small number of powerful individuals think (or are institutionally bound by their shareholders to think) that these issues are not serious.  In a sense these individuals represent an always-present dissonance in humanity.
  2. There are also a number of scattered environmental and social welfare organizations, at best contributing to a destructive tug-of-war over remaining resources.  These organizations lack coherence, each one splits a potential constituency, and there is no overarching model or view of issues across them.  In a sense each one is a special interest group - just like their opponents.  Particularily concerning is their rhetoric of social engagement.  Marches, elections, feel good slogans, environmental posturing and the like are only marginal ways to effect social change.
  3. At the grass-roots level are the consumption patterns of billions of individuals.  Many environmentally friendly solutions are perceived of as being more expensive and are simply a non-starter for these individuals.  Any solution offered to people would have to be demonstrably cheaper than the solutions they are currently exploring.  That means showing them the true round-trip cost of their decisions and showing them the actual state of their surrounding ecosystems.

What is needed is to organize with teeth.

To unify environmental organizations, to improve their law-making powers, to use the law to bring opponents to the table, and to argue with analytical precision based on factual data.

Ultimately everybody has the same stake - and participants on all sides are victims of their positions; everybody is racing to stay in place.

Tools that show the status of ones surroundings may act to unify perspective; at least that is the hope.

Effective change 

We have a physical capability to manage planetary scale resources.  It seems plausible that we can act in a globally coordinated manner.  However unified action on this scale only comes from awareness of the issues coupled with a sense of urgency.

Individuals are today employing many strategies to protect themselves from this coming challenge.  One strategy: that of building up financial buffers is not going to be very effective when the environment is too degraded to act as a backing for that wealth.

What being proposed is that we need tools that create constant and persistent awareness.  And that (if the assessment is correct) that the users of those tools will be more effective and simply more fit than resource abusers.

Technology behind the Web

Today there are two distinct kinds of web. We still have the traditional web, primarily populated by static web pages meant for human consumption, and we have the new 'semantic' web, today largely consisting of blog postings and being machine readable. This new web is itself undergoing continued rapid evolution.

Semantic Web

The fundamental feature of the new web is that it separates the structure of content from the presentation of that content. New-style content is now enriched by metadata that can be used by machines to help categorize, present or otherwise perform a variety of previously difficult organizational roles. For example a series of blog postings can now be trivially sorted by date, or filtered by degree of separation of the authors from oneself - two simple examples that would be exceedingly difficult to do with traditional web media.

Volatility

The new web chases a volatility curve. Use of the Internet can be framed according to the transience and volatility of the traffic on it. On one end we have traditional static content that rarely changes ( such as a biography or a tutorial ) and on the other end we have real-time chat such as IRC , AIM or SMS messaging. Blogging tends to be situated closer to the volatility edge than traditional content. Where google for example collects web-sites on a 3 to 4 day basis - blogging aggregators such as Feedster and Technorati tend to collect web-sites on an hour to hour basis.

The more volatile the digital posting the more difficult it is to track and the more relevant it can be to moment to moment situational awareness.

Standards

There are two competing standards for powering this new machine readable Internet:  RDF and RSS.

The semantic web will hopefully someday be powered by RDF ( Resource Description Format ) however today it is powered by RSS (Really Simple Syndication). RSS is an early and non-compliant version of the RDF grammer that has achieved massive popular success unfortunately ( 95 percent of blogs are RSS powered and are not RDF compliant). RSS is fairly limited in not supporting the ability to extend the metadata and having a fairly weak model of relationships over the internet. RSS may actually always remain dominant for blogging with RDF taking over the other possibilities for structured knowledge sharing.

An RDF document declares a series of triplets or statements about resources on the Internet. A single triplet is like an atomic unit of thought in the RDF nomenclature - it declares 1) a resource 2) a relationship or predicate 3) and a value. This parallels the english "subject, verb, noun" sentence structure and also parallels the convention in object oriented computer programming of declaring "structures" or "classes" that define different kinds of properties about a given object.

When an engineer or domain specialist seeks to create a metadata definition for some new domain of interest they can peruse well known RDF predicates to see if the kind of relationships that they want already exist. Alternatively they can define new RDF predicates and define a new RDF namespace to capture and share those predicates with others. The idea of namespaces allows a single RDF document to refer to many collections of useful predicates when describing a set of assets. Examples of popular RDF namespaces are RSS itself and FOAF ( friend of a friend ) which is used to define common properties of people including an ability to enumerate peers and relationships - something that can be used to help filter out noise for example.

New Directions

RSS/RDF have also helped promote new ideas and discoveries:

One is the idea of using a RESTful gateway for all machine driven queries - this means both using HTTP and using certain conventions such as persistent URLS with certain phrasings. One benefit of REST is that traffic can be directed over port 80 which is typically visible to the Internet on most gateways - most other ports having been blocked in a possibly misguided attempt to improve security of internetworks.  Another benefit is that often humans can form their own queries easily using a conventional web browser and this helps with development and testing.

As well there has been a shift towards using cellphones and mobile devices to access the net.  Multi-modal interfaces are now available over a wider variety of viewing tools.  Tools exist to reformat content for cellphones or to reformat content for disabled persons.

Standalone blogging applications are becoming popular.  These are dedicated to quickly managing structured content as opposed to traditional web browsing.  It perhaps indicates the end of the 'single tool does all' philosophy of the web browser.

Many natural systems follow power laws where instances of objects exist at vastly differing degrees of size. This is quite noticable on the Internet with some blogs receiving orders of magnitude more visibility than other blogs. Reputation and trust filtering will help create a possibly more equitable distribution of user attention where the most popular content will be the most meritous as quickly discovered by a group of observers rather than simply the traditionally most loudest.

Currently the idea of privacy or limiting the scope of publication of a blog posting is non-existant. It may (arguably) be possible with better trust filtering to actually publish something to a limited scope audience. More likely however what will happen is that human conduct will become more informal and colloquial.

Blogging directions

In the future even the most basic blogging tools will probably offer a comprehensive range of services. They'll allow authoring of a wider variety of types of objects and will generate a wider variety of views of those objects. Typical uses will probably include:

An example of a new process management tool based on blogging is BaseCamp.  It's a good demonstration of the power of logging in general and how logging can help organize.

Group citation sharing is another new intriguing application. Del.icio.us, probably one of my favorite new services, is a hosted service that provides a peek into the modern zeitgeist by sharing bookmarks. People effectively rate links by choosing to publish them or not and can integrate views over shared interests using a novel 'tag' based mechanism. All Consuming also explores the citation space by discovering the most interesting books (at least to bloggers). In a sense these services are almost like a virtual pokemon game - where contributers are trading a currency of the most interesting links with each other - there is a gamelike aspect at work where participants are somewhat playfully exploring their favorite interests. They also act as discovery tools where individuals often end up discovering other individuals with similar interests based on a mutual intersection of citations.

Aggregators

Aggregation is another result of a semantic web. Feedster for example provides searches of recent blog posts. Mikel Maron's 'World as a Blog' provides compelling visualization of geo-tagged blog postings. Maciej'es Blog Census shows total blog populations broken down by language. Aggregator based models can perform powerful synthesis and visualization of content - and we're seeing only the beginning of this. One can imagine a corporation for example that doesn't actually pre-define project schedules, components and dependencies but simply aggregates all of the blog posts of its developers and extracts a laundry list of known project components being worked on and those project dependencies.

As a contrast to aggregators there are large 'hosted service' models such as Tribes, Friendster or Orkut. These services are somewhat dismaying to open data proponents not simply because they are hosted but because of a commercial model that seeks to protect the content as an 'asset' and thus partition that content off from the broader community. Colloqiually this is referred to as a 'walled garden'. Also as a contrast to the aggregator based model are traditional services such as Craig's List - which while providing excellent community value is both a hosted service with some commerical goals and doesn't use machine readable web standards as of yet.

Future

The individual desktop will probably move seamlessly out to the web; with web gateways and local dedicated windowing environments folding together. Any kind of structured content managed in special proprietary applications will probably yield to the kind of universal solvent of the semantic web.

The whole model that the web is shifting towards a collective vision that is tightly coupled to the local reality; not simply a media artifact or the position of a single demagogue. What is happening is a shift of personal publishing to take over the role of traditional 'mass media propaganda', paper media communication and knowledge distribution overall.

Ultimately synthesis, modelling and visualization of structured content on the web should begin to demonstrate properties akin to that of the nervous system of an animal; an invaluable tool for analysing and understanding our world and possibly the foundation of things we cannot yet imagine.

Technology behind Thingster

Development patterns

CMS

Thingster is a CMS with these goals:

Publishing/Posting

Posts may be:

Posts may indicate:

Posts may be published to friends, an extended group of peers or to the whole world.

All posts are simply logged chronologically and can later be revised if desired.

Other publishing features include:

Discovery

Thingster offers a variety of discovery mechanisms:

  1. Other users may discover your posts by looking at a cartographic projection of say an urban city center or a hiking trail they're interested in.
  2. Alternatively discovery may occur by filtering against a variety of criteria such as the degree of separation that they have from you, your trust rating and/or explicit search terms as well as location.
  3. A tags based categorization mechanism is provided to allow posts to be dynamically categorized into topic maps.
  4. Thingster also has proactive software agents that can discover and alert you to complementary interests such as concerns that interest you, buying, selling or trading services.

Signalling

Signalling mechanisms are actually special code to help broadcast signals to peers and neighbors - roughly signalling can be compared to google-adwords.

Users can signal important events or phenomena to their local community; occasionally bursting through the normal social barriers if need be.

Explicit signalling facilitates a peer to peer discovery of local micro facts and services - sidestepping a role traditionally reserved for mass media and advertising.

Money carries this role in our society.  For example LETS currencies will connect together small merchants and buyers in a town via a published list of merchants that accept those currencies.  Often the role of a LETS currency is catalytic - after a while it can then go away - once the information has been dispersed.  The problem with money based advertising is that it makes it hard to signal free things or things for trade...  it shifts all actions towards profit.

Views

These are the kinds of views supported or to be supported:

Assets

Asset management and publishing has several aspects:

  1. Assets are tracked using an industrial strength spatial hashing database (postgis).
  2. Assets also use RDF semantic web standards to allow distribution and federation of multiple servers.
  3. Third party services may crawl the Thingster databases.
  4. All of the Thingster processes can be run offline; with web content being served statically if need be.
  5. Pluggable software agents implement the advanced services such as matching agents and themselves employ semantic web standards.
  6. The service offers web, email and cellphone gateways for reaching assets.
  7. On the web the service can generate a variety of rich media displays such as cartographic maps, gantt charts, calendar and timeline listings of assets.
  8. The software is licensed under an open-source BSD license in order to maximize security and trust of the software itself.  I want to avoid a 'walled garden' effect of assets all being in one place owned by one entity by encouraging others to have their own servers.

Database and Hosting

The server is Jetty backed running on OpenBSD and talks to Postgres for storage.

The server is setup with support for multiple simultaneous disjoint instances of the server running on different database back-ends at the moment.  Used to serve multiple domains simultaneously from one host.

Support for a filesystem backed or PERST backed database is not yet done but is intended so that Thingster can be used on a client box.

Search

Generic search engine but can filter by location.

Ontology Management

Using RDF standards Thingster defines a number of new predicates that it will accept as an aggregator from other sites.

Images

Special support for generating thumbnails and extracting EXIF data from jpeg images is provided.  Geotagged images are automatically associated with their location for example.

Mapping

Mapping is an incomplete component (mostly due to time not technology).  Currently Tiger 2003 line segments are supported with support for VMAP0 to be done.

Mapping is probably the least complete portion of the package.  Chris Holmes of the Geoserver project has offered help and support where possible and the goal is now to switch to using Geoserver when time permits.

Currently the system has its own totally custom map renderer called 'Panterra' that is intended to be released as a separate package at some point.

Here are comments excerpted from the current mapping solution:

Panterra is an toy svg map renderer for large volumes of content built on top an embeddable spatial persistent datastore.  It is capable of seamlessly rendering the entire continental USA from the freely available 2003 Census TIGER/Line data . It is by  anselm@hook.org written in April 2004 over a period of two days as part of a continuing exploration on ways to do fast and easy, seamless and pretty client side map rendering on embeddable devices.  It is being released publically in the hopes that it will encourage other people to play here as well.

This was made possible by leveraging  Trevor Smith's TIGER/Line Parser and Konstantin Knizhnik's PERST package and  Jetty .  Panterra has no license - but some of its pieces are alternatively GPL'ed and BSD'ed.  Developed and tested on OpenBSD 3.5 with apache-ant-1.6.0 and jdk-1.3 native and Windows 2000 with Eclipse and Microsoft J# Dev Studio.  No JUnit tests available right now.

Note that panterra is a 3d mapping engine - not 2d.  It is derived from code that I wrote for a spinning globe a few years back.  Also note that Batik is not used since Batik is slow.

Restful API

The system offers an http based api.  Users can build their own user interfaces if desired.

The list of commands currently is:

Template Driven

Templating is proving to be a good pattern.  While not terribly powerful (being only able to do substitutions instead of 'real' work) it does allow a separation of the code from the layout and does allow the site to be themable.  Performance may still be an open question however as compared to raw HTML pages.

The template system is provided by Apache Velocity.  It can peek into system state by using the Session in HttpContext[/] handle which returns accessors for files, folders, current path, current request, form parameters and the like.

All events are captured by a root template called "index.vm".  This index rejects a few requests such as those for gifs and jpegs so that the broader system can handle them.  Other requests are treated as either being 'user scope' or 'global scope' and are routed to the appropriate child templates.

A display of the files in a folder is accomplished by looking at the mime-type of an object and then letting an appropriate template handle the response.

The way you set the metadata for a new or existing file is you simply start calling session.setMeta() - and when you are done you call session.commitMeta(). You can get at metadata for an existing file using session.getMeta(). Behind the scenes the system will save metadata to a file in the same location as the real file (if any) with the same name and with .meta on the end. This can be edited with a text editor if desired. If it is a new metadata file then the system always grants it a unique name based on the current date. It is ok to save a metadata file even if there is no 'real' file associated with it - in fact most things in the system are pure metadata.

Template tags currently include:

Other uses

The engine itself is generic - not particularily specialized to dealing with geography.  This is evidenced by three distinct uses it is receiving already:

  1. as a photoblogging tool
  2. as a geoblogging tool
  3. and as a book blogging tool

Other kinds of future applications that it could be used for include:

  1. being used as a personal organizational tool
  2. or personal exercise and diet logging tool
  3. a bibliographic tool for categorizing books in a library
  4. a tool for police officers to create a digital forensic reconstruction of a crime scene
  5. as a mapping knowledge server for environmental activists
  6. as a 'bottom-up' project schedule dependency generator for a corporation.

The big picture is that Thingster is just a blogging tool - and the hope is that it can extend its services to cover a variety of use domains.

Packaging and Distribution

TBD

Appendix: Far-field Renderer Pix

Probably the major work piece of Thingster is a pretty mapper.  This is still ongoing and here are some far-field pictures.  While these were generated on the server side as static images the more recent versions are now using SVG on the client side.

  1. Asia including a distant view of Japan.
    http://thingster.org/map.jpg?cx=117&cy=33&cz=0.3
  2. http://thingster.org/map.jpg?cx=137&cy=39&cz=0.1
    Japan in a bit more detail
  3. The Philippines
    http://thingster.org/map.jpg?cx=124&cy=10&cz=0.05
  4. North America focusing on the coastline of California overtop San Francisco.
    http://thingster.org/map.jpg?cx=117&cy=33&cz=0.3
  5. San Francisco area coastine (fairly coarse resolution map and no street data).
    Showing nice countour lines for the GSHHS coastline data.
    http://thingster.org/map.jpg?cx=-118&cy=33&cz=0.008

 

Appendix: Links

  1. http://del.icio.us/geo
  2. http://del.icio.us/anselm/geo
  3. http://www.indyvoter.org
  4. http://www.pervasive.dk
  5. http://www.mrl.nott.ac.uk
  6. http://www.localfeeds.com
  7. http://space.frot.org.uk
  8. http://www.headmap.org
  9. http://www.locative.net
  10. http://www.geourl.org
  11. http://p2pmap.org
  12. http://mapbureau.com
  13. http://www.blogmapper.com
  14. http://www.geosnapper.com
  15. http://www.terrafly.com
  16. http://proce55ing.net
  17. http://upcoming.org
  18. http://meetup.com
  19. http://mapbuilder.sourceforge.net
  20. http://www.geoserver.org
  21. http://gistoolkit.sourceforge.net
  22. http://jetty.sourceforge.net
  23. http://www.intertwingly.net/wiki/pie/Motivation
  24. http://www.kayaktrips.net/geo/
  25. http://beta.plink.org
  26. http://www.local-news.net
  27. http://www.worldinternetproject.net/
  28. http://www.mozilla.org/rdf/doc/aurora.html
  29. http://www.mozilla.org/rdf/doc/api.html
  30. http://www.mozilla.org/rdf/doc/overview.html
  31. http://www.cs.rochester.edu/sosp2003/papers/p125-ghemawat.pdf
  32. http://www.ai.mit.edu/projects/infolab/ailab 

Appendix: Contributions and History

The project is driven by a small community of geo, 'social metadata' and 'citation space' enthusiasts at the moment. People interested in contributing should contact us with patches or suggestions.

Contributers whose ideas or work used:

This project has been an ongoing goal in one form or another for perhaps 15 years.  Thingster as it exists now is about a year of part-time development. The core ideas predate the Web and initial attempts at implementing this service had to consider concepts such as 'presentation of maps, newspapers and classifieds on bbs networks' that luckily have been completely solved.  SimCity was one of the key original influences as well as Jean Quimet of the Green Party and a book called 'Anticipatory Democracy' by Clement Bezold suggested to Anselm by Jean.

More recently the ideas were dramatically influenced by the emerging of blogging, and the emergence of semantic web standards.

The thought of using the service as a general method of exchange and the name 'thingster' itself come from the Ben Russell at the headmap collective - a predecessor community that sparked the locative media labs community under which we all roost today.

Since this is BSD licensed this is a good way to continue funding the project on a long term basis while permitting commercial uses - commercial uses can split off the code base and don't have an obligation to push their commits back into the open source tree.  Others are welcome to take this code and encouraged to do same (although contributions back into the code base are of course welcome).

Appendix: Comparisons

There are a number of other overlapping services and blogging tools:

  1. Blogging
    1. Moveable Type being one of the grand-daddys of blogging tools.
    2. Typepad is a hosted service from the authors of Moveable Type that does support FOAF.
    3. Blojsom is a good example of an open-source cms for blogging that users can host themselves.
  2. Citation blogging
    1. Delicious is probably the best example of social mindshare over a specific topic - in this case urls.
    2. Music Brainz shows how a community can discover and rate topics...
    3. NY Books does the same thing for books; helping people help each other search through a large domain space.
    4. All Consuming takes things a bit further using an aggregator model to synthesize a book rating scheme from blogs.
  3. Schedule Blogging
    1. Basecamp is one of the first corporate knowledge blogging tools; able to generate schedules from blog posts.
  4. Health Blogging
    1. VitaSign is an excellent example of the use of logging as a powerful tool.
  5. Collaborative games situated in real world
    1. Asphalt Games by Michelle Chang and Elizabeth Goodman shows an area that a tool like Thingster could help with insofar as managing user accounts and tracking state and showing scoring as well as map visualization.
    2. (There are lots of other location aware games out now too - not cited).
  6. Geo
    1. Fundrace is purely a data-dump of knowledge dataset - not a collaborative tool
    2. Geowiki uses a wiki based model which can be difficult to manage having no metadata associated with posts.
    3. World66 uses a wiki based model which is somewhat problematic but it does seem to work well for them and has lots of great content.
    4. Urban Tapestries is an incomplete work in progress - looks promising - unclear if free.
    5. Maphub is a work in progress but fairly similar to Thingster and does have lots of nice text describing it.
    6. Onemap - possibly a best of breed mapping solution but not finished yet.  Based on the interesting work of the geoserver and geotools community.
    7. Geourl - not around right now but a good example of an aggregator based model.
  7. Images
    1. 93 Photo Street is Trevor Smith's application based photo manager but does extend out to the web as well.
    2. Geosnapper
    3. Flickr is another application based photo-manager but also having a social aspect with its own protocol (not web).
  8. Social (not blogging though)
    1. Tribe
    2. Friendster
    3. Orkut
    4. LinkedIn
    5. Plink
  9. Signalling and Trade
    1. EBay
    2. Craigs List
    3. gmailswap.com
    4. A plethora of dating services
  10. Event focused
    1. Localfeeds
    2. Meetup
  11. Politics and Activism
    1. IndyVoter
  12. General / Sweeping
    1. Nokia Lifeblog

Appendix: Content Management System Features

Here is a site comparing blog features between various blogging applications.

http://www.asymptomatic.net/blogbreakdown.htm

And here is how Thingster compares at the moment:

Other features that are worth mentioning: