|
Q: What is aspectweb.org? Is it an
on-line data repository, is it a search engine, is aspectweb.org
trying to build a community of developers, to attract user groups or
is simply a playground for experimentation? |
|
A: Technically speaking all of that and
a little bit more. However I would not set the emphasis on the
technological aspect which inevitably reflects the global state of art
adapted to our simplicity, scalability and flexibility needs
(interested readers can get an idea on the current status from aqoo
software manifesto). In few words, we propose to locate/relocate free
and shareable meta-data from the exclusive possession of specific
application to a globally accessible context, simplify to extreme
publishing, searching and integration of structured information. The
philosophical metaphor speaks about creating a generative field, about
potentizing static, apparently complete information by means of
active, dynamic knowledge, about leaving the control behind and let
the unfolding happen, about transparency and emergence.We invite
people to play, explore, have fun and innovate. |
|
Q: Pretty abstract. Do you have any
concrete examples? |
A: The collaborative applications are
pretty en vogue these days, but imagine a scenario where sharing of
information is less an application feature but more a global status
quo.
Imagine that while you are browsing *any* web document you are able to
enhance the content, supplement contextual navigation possibilities,
see the contribution other people, filter and group relevant aspects
to you based on logical/ semantic criteria, etc. And all of these
without any contribution, agreement or control from the entity
publishing the core information.
Or imagine virtual documents with thousands of users where
readers/contributors can filter, extract or print aspects related to a
particular interest. Is not the density and the relevance of the
returned information which counts only, but the multitude of
simultaneous perspectives on a given topic, integration of time as
qualitative and expandable (compactable) dimension. This sort of
document tailors its objective materiality to the subjective focus
each user has.
Or think about a small box in the right corner of your browser ( a
browser feature, a bookmarklet or a greasemonkey script for instance)
informing visually with color codes or ideograms, about the topic,
source reliability or contact persons (retrieved from a more generic
directory) for that specific knowledge domain. Aspecting may become
a highly efficient, independent and non-biased information channel to
describe resources. |
|
Q: Is this yet another attempt to
accumulate semantically structured (personal) information to feed the
advertising industry? |
A: This is more about enhancing content
or adding contextual navigation helpers and traffic signs to the
already existent data roads, providing ad-hoc, light-weight
connectivity between information fragments.
What we call aspects are small collections of facts logically
coupled together. Aspects can be aggregated further in higher level
artifacts such articles, documents, etc. Interesting enough (or
counterintuitive for most of us) the 'data query' is also the
compilation/aggregation agent. As example, accessing an on-line CV (
regarded as a composition of several aspects such contact information,
interests, prior work experience, etc.) is simply a matter of query
formulation like person:openID and ( tag:vcard or tag:interests or
tag:work_experience). The expected output is a higher transparency, a
better integration and a faster access to relevant or related content.
|
|
Q: This sounds very interesting. I
understand now that aspectweb.org enables us to publish and retrieve
to/from own repositories atomic information tokens (aspects are basic
pieces of semantically structured data) and eventually associate them
with regular web resources. Although your approach looks like a
pragmatic alternative to mainstream semantic-web, text mining, natural
language processing or fact extraction technologies, etc. the
challenge is still high. Are you able to face success, to scale
accordingly? |
|
A: Your assessment is correct. Actually
the core part, the data cluster, was written from the very first line
with this challenge in mind. It is designed to run distributed on
cheap, commodity computers. The experimental deployment on one or few
machines can grow or shrink dynamically according to needs and
potentially without any off-line time. Autonomy, self-diagnosis,
optimization and recovery during operation, auto-partitioning based on
data density or load are very exciting topics on aqoo data cluster
agenda. The software itself, now alpha, is made available to the open
source community. Current pilot projects started by aspectweb.org are
simple catalysts to foster interest and innovation. |