Platform Application Ideas
From n² wiki
letolizel // will filter/split off another page soon for most promising ideas
[edit] LazyWeb NG
Similar functionality as Ben Hammersley's lazyweb.org, but need to find an alternative to trackback pings
[edit] SemWeb Music Store
(was Pet Store)
Typical storefront app, along the lines of the J2EE Pet Store http://java.sun.com/blueprints/code/
See: SemWeb Music Store
[edit] Blog
simple blogging tool
[edit] Wiki
build a simple wiki based on very loosely-coupled components, including:
(Creole) Wiki syntax to XHTML converter XHTML to Wiki syntax converter (via XSLT?)
[edit] Bliki
combine the above
[edit] Person lookup
take foaf:mbox/foaf:mbox_sha1sum return foaf:homepage/foaf:weblog/foaf:PersonalProfileDocument
[edit] URI Lookup
Uriqr kind of does something kind of related to the Person lookup above, taking a literal and finding URIs of things foaf:name'd or rdfs:label'led with that Literal, ranking them according to usage data provided by Sindice. It kind of works already, but slowly, and without a sufficiently comprehensive dataset at the backend. I'm interested in using a Platform-based store of FOAF data to power Uriqr, which is hopefully coming in the near future.
At a more generic level, I can imagine many interesting Person-oriented lookup services being powered from a similar dataset.
[edit] Bookmark Miner
upload Firefox bookmarks becomes (maximally) indexed in service
ALSO aggregate del.icio.us feeds, do lookup/data extraction on link targets
[edit] BugTrac
Hook up :
http://simile.mit.edu/repository/RDFizers/svn2rdf/
(that's in Python, has XSLT)
[edit] news translation helper
take feed of BBC news (any feed) filter for non-common words map to wordnet provide wikipedia-like user disambiguation of senses
[edit] Feed2Mail
A service which subscribes to a feed on your behalf, sending you an email for each new post.
Components:
RDF (+HTTP) to email
RDF doc POSTed, contains
[
a m:Message ; m:to _:recipient ; m:from _:sender ; m:subject "The Subject" ; m:message "Here is the message"
]
Message is CONSTRUCTed using a SPARQL query against model containing an AtomOwl representation of the feed.
timer for polling:
get feed.atom from source
atom2rdfxml.xsl
sparql service - construct
rdf2mail
[edit] GRDDL transform lookup
registry given a string e.g. "text/html" look up how to get to GRDDL
[edit] "I'm Portable" widget-utility
A blog widget that offers personal profile data and it in multiple formats (FOAF, hCard, vCard...). It does nothing more than take a single source of profile data and make links to the result of running those through conversion.
I'm betting something extremely close to this has already been done, here are what just suggested it:
1. Something-to-RDF converters have been under discussion recently (mostly re. attempts to collate what's available - see RDFizers, ConverterToRdf, MicroModels
2. some online conversion tools are available - see Babel (various things to RDF), RDF Converter (between RDF syntaxes). There are several microformat-to-traditional format converters too, along with at least one online XSLT service.
3. as Brian Suda just noted in a LinkedIn invite, it's a real pain that social network data isn't portable across systems (and microformats and/or RDF could help)
4. online services rarely grok RDF (nor should they have to)
So what about flipping 1 & 2 around for RDF-to-Something conversion? Here's how it might tie in with 3 & 4.
Direct serialisation (Raptor?) XSLT over RDF/XML SPARQL (one or TwoPhaseSparql) programmatic
[edit] Maximal HTML page interpreter
triplr++ - tidy, grddl ,all links, xmp from images, feeds etc...
[edit] Deep top 100
(application of Maximal HTML page interpreter etc)
go through something like Technorati's top 100, first pass grab the feed and FOAF data of each site, make an aggregator (boring!). Second pass - address what the blogs are about, e.g. scrape Gizmodo, gather microformat reviews of the gadgets described. Accumulate all of this stuff in a triplestore with SPARQL endpoint, do a nice facetted view, etc etc.
[edit] MyKnowledgeBase
Upload/link to:
FOAF profile firefox bookmarks subscription list
all available data is scraped/parsed page text indexed (Lucene + some RDF : including tagging)
available through views, e.g. Longwell
I'd be a little concerned that going from JADE to session beans with JMS would only be replacing one over-engineered solution with another (I'd try simple HTTP messaging for starters)
prior to 1) th
[edit] Test harness
Build an online test system
Creole Wiki converter tester as demo?
user supplies:
URI of test requirements manifest (this is the agent definition) URI of conversion service
Ditto for GRDDL
[edit] Text Editor!
If the shared-VM comms are adequately fast, why not?
use a graph model see http://www.finseth.com/craft
[edit] ultimate conference panel
???
[edit] Atomagotchi
fed when you add to the sum of web knowledge
uses an aggregator fed by any feeds identified using FOAF profile : personal blog, revyu, wiki edits, del.icio.us
produces feed describing current state
use foafster images - svg
state machine
has litter which needs emptying? (bad markup?)
has its own foaf profile can know other atomagotchi
states:
happy, sad, cross and concentrating Hungry - cross Eating - concentrating Full - happy Sleeping -
think bubbles
game tests: http://www.gamasutra.com/features/20070212/garneau_01.shtml
[edit] SW Pattern Repository
This blurs into Best Practices and Deployment a bit, but then what doesn't... see DesignPattern, WebPatterns.org , Web Design patterns a couple of starters:
- FilterDown - interface with RSS/OPML
- TriplestoreWhiteBoard - the agent trick revisited
PS. there are already some patterns on the ESW Wiki but they don't appear to be in a Category
[edit] Job Mart
Job/studentship mart (for SW?)
[edit] Feeds of my Friends
Given URI of FOAF profile, display aggregated feeds from every foaf:knows person
here's an idea I'd quite like to see - kwijibo
[edit] Space and Time visualisation app with faceted browsing
[edit] Ingredients
- Google map widget (or similar)
- timeline slider widget
- Longwell style widget for constraining types and properties
- Data with time/date and geo-co-ordinate properties
[edit] User experience
- I select the class and property types to filter the data
- I adjust the map to the zoom level I want
- I move the slider to the time period I want
- pins appear and disappear on the map as I move the slider and adjust the filters
I suspect that much of this could be done relatively easily with existing javascript widget libraries (eg: from YUI). I think it could be made to look good and be fun to play with, and could moreover be an interesting and useful visualisation tool. Because the application is not domain specific, it would also be a good demonstration of some of the value of using RDF (ie: whatever your data, you can just plug it into something like this and it will work).
[edit] Possible scenarios
- election results (party affiliation of constituency representatives) and (eg) hospital openings and closures
- the locations of european universities and printing presses 1450-1700
- geographical distribution of people who know people (it's a shame in some ways that FOAF only really expresses the present tense)
- which chain supermarkets opened where and when.
[edit] RDF2OPML
Not an application as such, just provide a fairly generic SPARQL query + XSLT to give results that can be viewed through tools like Grazr (http://grazr.com) and Hyperscope (http://hyperscope.org/), plus one for pulling out subscription lists.
[edit] "Democracy"
(Need a working title!). The democracy-related things Keith's been talking about - see http://project.knowledgeforge.net/semwebparlparse/trac
[edit] Tomboy backend
Online store for the notepad/wiki see http://en.wikipedia.org/wiki/Tomboy_%28software%29
[edit] Pastebin
Source code snippet upload area, see http://en.wikipedia.org/wiki/Pastebin , http://pastebin.com/ , http://bitworking.org/news/187/Robaccia-Pastebin
[edit] MS Access-a-like
Easy form builder etc. in the style of everyone's favourite DB. See also : http://www.snee.com/bobdc.blog/2007/07/semantic_web_project_ideas_num_5.html
chrisc: This could be easily be built upon the recordstore API that is inside Talis Engage.
[edit] Playlist manager
build something from XSPF: http://lists.musicbrainz.org/mailman/listinfo/playlist
[edit] Software auto-update
library to auto-update software products, with tools for creating/checking md5 hash etc
[edit] Migraine Management System
Provides individual advice for dealing with migraines & similar headaches
Known data base - stuff on symptoms, possible triggers (especially dietary)
Various individual data input channels - questionnaire, guided diary (especially for meals)
Wikipedia http://en.wikipedia.org/wiki/Migraine http://en.wikipedia.org/wiki/Cluster_headaches
[edit] FOAF, DOAP etc Validator
RDF/XML syntax validation, plus some rules to see how well the data fits the expected profile
[edit] Social Network Portability
FOAF-based solution to this: http://microformats.org/wiki/social-network-portability
[edit] URN Resolution Service
For folks like HCLS, see http://www.faqs.org/rfcs/rfc2169.html
[edit] "Citizen Analysis Tools"
A combination of TheyWorkForYou, UpMyStreet, and the data integration power of the Semantic Web. Which is the best area to live in in a city, based on a combination of cost, amenities, green space, crime, schools, transport, etc? Plus loads more stuff based on local council data, etc.
[edit] Site Cloner
Crawls an existing site, clones content with same URL structure. Ideally it would also be update-aware, keep in sync.
Applications:
- moving from one blogging tool to another - maintaining old links without maintaining old software
- live site backup - if the real site goes down, flip the DNS while you fix it
