Automatic and Manual Processes, and Both Combined
Automatic
processing of information content is widespread, and for good reasons.
It is cost-effective and works satisfactorily most of the time.
However, there are cases where it is not good enough and where the
input of experts is invaluable. The situation isn't black and white. We
can do both at the same time. Infoloom uses a unique methodology,
combining automatic processing with input from experts and skilled
professionals.
Preserving the Workflow
Organizations have to adjust to changing
conditions, but do so most efficiently by leveraging accumulated
lore. The biggest asset of an organization is the information it
creates and is able to use. Creators of information usually want to
concentrate on the information content itself rather than on the
technologies that they need to use in order to insert the content into
the system. Because our approach privileges post-processing of
information, it allows the possibility of letting people work the
way they are used to, the way they like, the way they are familiarized
with, while processing and insertion into the system is delegated to
a later, and independent, stage.
Flexible Linking
|
We help our customers
create "views from above" over their information
assets. This corresponds to what is often called
"metadata", and makes it possible to separate the
management of knowledge from the management of content. This
idea is at the core of RDF, Topic Maps, and the
Data Projection Model, and is essential for integrating vast
corpora of knowledge.
|
|
|
Accessing Legacy Data
Do we want to see just the
content of information, do we want to preserve the
original presentation, or both? In the former case, we
can use a mapping technology to attach meaning to
portions of content, at significantly reduced
costs compared to the cost of converting
existing content to a common format. In the latter two cases, a combination of mapping and XSLT-based technologies can be used to preserve original presentations. We focus on ways to address relevant information for efficient reuse.
Multiple Perspectives from the Bottom Up
Many companies struggle with two contradictory
requirements: to unify the architecture of the information within one
organization or one community of practice, and to leave enough room for
accommodating variants and/or future points of view that were not known
when the system was designed. Top-down is generally a sound strategy,
but it will work even better when complemented with bottom-up
approaches. Such design will enable integration with diverse projects and
information systems that are created independently of the main system.
The belief that semantic
interoperability can be achieved by everybody adopting
a common way of thinking only works within a small
community and for a limited period of time. When
somebody comes up with something new and unexpected, the
system crashes.
Integrating various sources of
information can prove quite challenging, because of their diversity.
When converting everything into a common format is too costly, we
propose alternative, softer solutions to handle the
various contributing perspectives concurrently.
|
