Introducing Myself

Better late than never!

This contribution follows attending ISKO 8 in London last week and 
meeting Kathryn La Barre, Aida Slavic, Claudio Gnoli, Douglas 
Tudhope, Ceri Binding and others and being reminded of this list by 
Aida. I'm clearly out of the loop and need to make this
connection.

My realisation of the relevance and importance of facet analysis and 
faceted classification came in the course of evolving prototype 
interfaces to improve access to databases.

The history of these prototypes can be found at 
http://www.view-based-systems.com/history.asp 
We created a Research Centre – CeDAR – the Centre for
Database Access 
Research at the University of Huddersfield as a focus for these 
efforts.

The evolution from expert system intermediary to the explicit use of 
facets in the interface culminated in HIBROWSE – High resolution 
Interface for Browsing and Searching – in a British Library
Funded 
Research project which used a subset of 600,000 records from Elsevier 
Science's EMBASE database – I've just revisited the
Screen Cams 
available via the history page:
http://www.view-based-systems.com/papers/hbscrcam.htm 

The presentation at the IFLA 97 General Congress in Copenhagen 
credited much to Ranganathan and the CRG – who had been there
before 
but had not had the advantage of our electronic tools to put some of 
their ideas into practice for retrieval.
 http://www.view-based-systems.com/papers/pollifla.htm 

We had applied these ideas to a variety of databases – A
Directory of 
Hotels, a Company Directory for IT (another ScreenCam):
http://www.view-based-systems.com/papers/datasit.exe 
and Student Records – which provided clues as to how the
application 
of facets at the interface was generic for all kinds of database 
content – as long as there were attributes that provided
hierarchical 
views – although we have no problem with long lists with what we
are 
doing now.

We were continually seeking partners to take this technology to 
market – without success. One potential partner did however say
to a 
key potential client:
"If you would like a view-based interface to your directory then
we 
can build you one" – without getting back to us after we had 
demonstrated prototypes to them! 

In the end I left the University and formed View-Based Systems Ltd 
and moved to the Yorkshire Technology Park – in "Last of the
Summer 
Wine" country – for those who have caught an episode of the
longest 
running BBC Comedy Series.
http://www.view-based-systems.com/contact.asp 
Our first contracts took us into the UK National Health Service with 
applications relating to Continuous Professional Development and 
Workforce Planning that embodies View-Based Searching Technology
– as 
it continued to evolve.

More generics were addressed as we had to deal with statistics that 
related to such measures as 0.75 Whole Time Equivalent – the
human 
resource measure within NHS Contracts meaning you have 75% of a 
persons time – rather than working with simplistic head counts.

Although we have licensed these systems to a large number of NHS 
Organisations the application developments relating to forecasting 
requirements for staff deflected from our efforts at doing view-based 
searching against different  databases.

Our multiple facets in the interface were reduced to one view at a 
time – given the low resolution of devices in UK Hospitals when
these 
were first introduced (would you believe 640x480 in places) – 
however – we also came up up a generic interface that lead us to 
present VBS Browser as a general purpose Internet Database Browser
– 
and even presented this at the Online Information Meeting Exhibition 
in London in December 2003:
http://www.view-based-systems.com/vbsbrowser.ppt (a self-running 
powerpoint)

All this is being done using standard SQL with SQL-Server 2000 –
and 
delivers a Windows application that can connect to databases across 
the internet – a Web Browser delivers the details of objects
using 
one .asp page per object – i.e. no web site as such.

This didn't exactly excite anybody – so on we progress from
one thing 
to another.

The connection to bibliographic databases was however maintained with 
a PhD programme of research that sought to deliver facets from 
deconstructing Dewey Class Codes – the last remaining CeDAR
project:
http://www.view-based-systems.com/VBSHelp/VBSOPACHelp1.htm 

This project required a usability component and serious examination 
of this brought us back to a multiview inrterface – we also took
note 
of what Marti Hearst has done on Flamenco and Eero Hyvonen did with 
Finnish Museums online – delivering a pile of facets as the first 
point of interaction.

The latest example of doing this in the library context can be seen 
at:
http://www.view-based-
systems.com/CataloguingandIndexingConference.ppt 
which was presented at the Bath Cilip (UK LA and IIS) Cataloguing and 
Indexing Annual Conference at Bath University and, cheekily, at the 
ISKO 8 conference when a spare slot became available through non-
attendance of a speaker.

I have looked at some of the exchanges relating to relational 
databases on this list and can only say that with IBM, Microsoft, 
Oracle and others seeking to improve the performance for RDBMS –
and 
ongoing hardware improvements – any special proprietary
structures 
that developers feel are required at the moment may not be 
justifiable in the "near" future.

The comment on Adiuri Systems Web Site referrng to scalability in 
what they call Adaptive Concept Matching that:
"Others have attempted to provide ACM-like functionality with 
relational databases and SQL queries. However, for large collections 
of documents, this strategy becomes impractical as the number of 
calculated document classification states increases in a geometric 
manner, as the size of the document collections and concept 
hierarchies increase. The ACM based approach is scalable solution for 
searching large collections of documents classified against multiple 
concept hierarchies." 
Is interesting.

I've questioned them on this as there is no indication who they
could 
be referring to – and can only presume they haven't figured
out how 
we do what we do with relational databases that means there is no 
geometric increase in computation!

We have now updated our web site to incorporate a background page 
with links to publications citing other examples of the use of 
similar approaches:
http://www.view-based-systems.com/background.asp 
We've also added a link to this discussion list.

One of the latest additions to this has been the link to
http://www.shopping.com 
showing that these techniques have arrived in mainstream applications.

The efforts by teams such as at Endeca and i411 are impressive –
as 
are the prices they quote – and it would be interesting to see
where 
View-based Systems Limited would be were we to have the kind of  
venture capital underpinning that these organisations have had.

This could be where we launch into a variation of the Monty Pythons 
Yorkshiremen sketch – there were just four of us working in an
old 
textile mill – having to write our own code using Visual Basic
….. 
which I'm sure someone could say - … you were lucky –
there was just 
the two of us working in a garage …. This could be where our
strategy 
on leaving the University fell down – we should have rented a
garage……

Back to reality – we would very much like partners and
collaborators 
across the globe – travel is essential for broadening the mind
– and 
partnership and collaboration is the only way we can see any 
successful deployment of the low-cost high return technology we are 
developing.

We will be moving to the web in the "near" future now our
generic 
software has been suitably tiered – it will be .net in the first 
instance but the object-oriented design should lend itself to 
conversion to Java as and when.

In the meantime I welcome comments and questions and offers of 
collaboration and partnerships.

Well done to all those that got to the end of this posting.




------------------------ Yahoo! Groups Sponsor --------------------~--> 
Make a clean sweep of pop-up ads. Yahoo! Companion Toolbar.
Now with Pop-Up Blocker. Get it for free!
http://us.click.yahoo.com/L5YrjA/eSIIAA/yQLSAA/0bmwlB/TM
--------------------------------------------------------------------~-> 

----------
Thanks for playing. 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/facetedclassification/

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/