libby.miller@bristol.ac.uk 310800

What is this? 

It's a playpen for exploring how to display and use resources sdescribed
in RDF, written by many people over time, including
libby.miller@bristol.ac.uk, daniel.brickley@bristol.ac.uk and Joel Crisp,
with lots of borrowed code (see the README in rudolf-project/lib, and
thanks to Marin Poulter and Jasper Tredgold in particular).

It's therefore a mishmash of different tools written because we hadn't
tried doing something yet, or because it seemed like a good idea at the
time. There's also a lot of crufty, pre-alpha code in this distribution. 
We've got to the stage where we could rewrite it well now that we know
what were're doing, but at the time we didn't. 

Ok, what is it exactly? 

You can divide it up into 4 parts

1. a displying part: displays sitemaps rdf data and rss1.0 (usually). See
the bottom of the README file for more info on sitmaps syntax.  It does
this by building a tree structure out of the data. You specify in
rudolf-project/config/testconfigure.rdf which rdf resources you want to
display, or you can add or remove boxes (channels, rdf resources) using
the interface tools, or you can just append

host:port/rudolf/mydesire with ?uri=urlofchannel

2. a searching part: cross searches various data sources, most of them
Whoiss++ databases, but also SQL data sources of a partular type, via a
servet interface which outputs RDF in rss or sitemaps format. The urls for
this are

host:port/rudolf/query?term=...  host:port/rudolf/rssquery?term=... 

you can add youyr search targets to the sample file

rudolf-project/web-htdocs/examples/services/sbigs.rdf

or create a new one.  Note that the sql targets do not actually use this
information - they search according to their local parameters, as
specified in

rudolf-project/web-htdocs/config/dbconfig.rdf
rudolf-project/web-htdocs/config/dbconfigwr.rdf

Note also that theoretically you culd search targets with other protocols,
such as Z3950 - but not yet. 

The actual search code launches theaded searches in parallel. 

3. a recommender part: the idea of this part is to have a recommendations
database, which acts like a communal online bookmarks store. There are
some cute ways of inputting data such as url, title, description using
javascript bookmarklets, which can be dragged on to your toolbar and use
javascript to grab informaton about the webpage, including highlighhted
text as description. I didn't write these tools: see www.bookmarklets.com,
martin Poulter m.l.poulter@bristol.ac.uk, and Dan Connolly (see
http://www.xent.com/current/0863.html). 

These tools enable you to put the data in the database quickly. The system
uses postgres, but could be adapted to se mysql or similar - all the
database confuration information is held in the rdf file

rudolf-project/web-htdocs/config/dbconfig.rdf

You can also get it out again in various ways, again to display as rdf
sitemps or rss channels. Rss1.0 is only possible for flat files (and
searching) at the moment, e.g. 

host:port/rudolf/rssquery?term=fun
host:port/rudolf/rssbydate?date=mo/da/yr

channels: 

host:port/rudolf/bydate?date=mo/da/yr

host:port/rudolf/byperson?mbox=mailto:libby.miller@bristol.ac.uk (gives
all that person's recommendations. May be a very big file, too big to
parse by the SAX xmlparser) 

host:port/rudolf/recommendations?mbox=mailto:libby.miller@bristol.ac.uk
(gives that person's subjects, a smaller file) 

host:port/rudolf/recommendations?subject=economics (gives all the
recommendations under that subject - not by person, unfortunately) 

host:port/rudolf/recommendations?resource=http://www.bized.ac.uk/ (gives
all the recommendatins about the resource) 

not done: a proper att/val query servlet

4. an rdf harvester and basic browser: this part is for gathering and
viewing rdf. 

Put urls in rudolf-project/org/desire/rudolf/webring/scutterplan.txt, i.e.
on a separate line

./scutter url

put a configuration file in rudolf-project/org/desire/rudolf/webring/conf

in rudolf-project/org/desire/rudolf/webring, type:  make scutter

and it will harvest this rdf data.  Then you can look at it by pointing
rudolf-project/web-htdocs/config/dbconfigwr.rdf at your database and
running the servlet

host:port/webring/FetchRDFServlet

This interface consists of several text-based queries that won't work with
your data, plus a facility for adding data to the database (it also goes
in the scutter list) and for displaying it (handy for making sure that the
RDF parses). Finaly, at the bottom of the page you should be able to
browse your data, startign by picking a subject or a predicate from the
form list it should have found from your database (if there is any RDF in
it). Queries of this show arcs in and arcs out for subjects and objects,
and all triples with that predicate for predicates.

5. other stuff

This is a playpen, and there is Java code in this distribution which is
not used, for example in
rudolf-project/org/desire/rudolf/tripleset