[[page created automatically from word-processed document; for original see: Postscript version]]

RESEARCH LESSONS AND PROJECTS

Peter Brown

Dept of Computer Science

Univ. of Exeter, UK

Aims

• To describe current research in order to:
  • draw some lessons.
  • give people an idea of what I am/was interested in.
• To describe current projects.

Early research

• Systems software:
  • Operating systems
  • Just-in-time compiling: ideas that have recently found their time again.
  • Macro processors:
    • lightweight structural pattern replacement in documents, e.g.

      IF a THEN b ELSE c;

      may need systematically mapping into:

      (a) ? b : c

    • not easily translatable into graphical interfaces.

Theme: regular need for new research directions

• "It is hard to escape from success".
• Trick question: what research will you be doing in 5 year's time?
• Good catalysts:
  • Moving to a new environment.
  • Sabbaticals: especially a do-something-new sabbatical.
  • Incoming visitors.
  • Contacts with outside departments.

1980-1: sabbatical at Stanford

• Chip design:
  • the famous Stanford VLSI design course.
  • Ours was the `misfits' project.
  • Value of cross-fertilisation.
• Being at the centre of the graphics workstation revolution:
  • Xerox PARC.
    • ideas -> projects?
  • At Stanford: Sun, Silicon Graphics, ... .

Research in hypertext

• Exploiting the new medium (but not looking exactly like a book).
• Hypertext concepts date from 1945; in a mature form by 1968; late eighties before their time came.
• Current work: is a hypertext link a goto?
  • `Goto considered harmful' by Dijkstra.
  • `Links considered harmful'.
  • Questions of state.
  • Is a link actually a procedure call, and does it therefore suddenly become a good thing?
  • Although some analogies are superficial, there are huge potential gains in bringing software engineering and programming disciplines to hypertext.
• In 1991, the first WWW presentation:
  • "nothing special about it".
  • hypertext features were (and still are) weak.

My favourite quote

• "People do not want hypertext; they want solutions. Hypertext is only likely to be part of any solution."

• Perhaps the reason for the WWW's success.
• You can substitute almost any technology for "hypertext" in the above quote.
• Certainly it applies to context-awareness.

A further hobby-horse

• Collaborative work is likely to be more successful if initiated outside government schemes.
• MD's comment: government pays 25% of project costs; however all the form-filling, delays and hassle count for minus 30%.

1992: high time for a change of direction

• Sabbatical at Xerox EuroPARC in Cambridge, working mainly with Mik Lamming's group.
  • Work on a memory prosthesis and its realisation in the Forget-me-not project.
  • A project well ahead of its time; unfortunately not all work is open.
  • Themes:
    • an aid to human memory.
    • use of small mobile networked devices.
    • automatic capture of context.
    • use of context as a filing/retrieval system.
    • involved work with psychologists.
    • issues of personal privacy: problems with active badges; `very bright' correlates with `prima donna'.
    • is leading to a Xerox `satchel' product.

My main current project

• Context-aware mobile systems; building new software, new data representations and new storage/retrieval algorithms to exploit context.
• Current work is concentrating on retrieval.

Categorising context-aware applications

• A database or GIS application.
• A type of information retrieval.
• What we need to do is solve the HCI issues.
• An area in need of a sound theoretical base.
• For real applications you need to solve hard AI issues.
• It is a systems/software engineering problem, e.g. provision of toolkits.
• A problem requiring the design of new data structures, notations (the equivalent of HTML) and algorithms.
• The area is waiting for the right hardware at the right price.
• A particular sort of distributed system.
• You must solve the social problems first, e.g. privacy.

• All the above are true. The real key, is bringing them together.