Information Technology Reference
In-Depth Information
However, in the current version of webble technology, there is little or no help
provided to the casual user in order to (a) search a potentially huge pool of
webbles to find those needed for a specific application and (b) know whether the
functions of two webbles are compatible (and therefore whether the webbles can
be connected). As a consequence the user must try to connect webbles by trial
and error - a formidable task, as the number of slot combinations is huge even
in a pool of a few tens of webbles.
In this paper, we propose to solve this problem by introducing annotations
at two levels: at the webble level and at the slot level. Our claim is that if these
annotations are based on standard, controlled vocabularies then one can define
search mechanisms (based on those vocabularies) allowing to access webbles of
interest, and to have information about compatibility of their slots. Additionally,
the world of webbles can then communicate with other worlds using the same
vocabularies, and in particular it can communicate with the world wide web.
Today, there are several standard vocabularies, such as Dublin Core[2], IEEE's
LOM [15], CIDOC-CRM[3], ACM's CCS [16] and others. These vocabularies are
rather widely used today in several areas, including digital libraries, archives [4],
or distant learning systems, to mention just a few areas. Annotation framework
using such standard vocabularies or more sophisticated ontology representing
domain knowledge [5][6][7] is continuously very hot research topic and already in
use in various application domains. One of most successful examples is semantic
annotation of biomedical data. Nowadays various biomedical data are annotated
by using several standard vocabularies [8][9] and those annotations are widely
used for biomedical data analysis [10] and information retrieval[11].
Our proposal is also motivated by another simple observation regarding web-
ble technology. To begin with, from a programming point of view, webbles can
be regarded as a visual, dynamically typed programming language. Compared
to static typing, dynamic typing has the advantage of enabling the connection
of arbitrary pairs of slots during runtime. However, such over-flexibility entails
the risk of misconnections, as well as diculties in choosing an appropriate com-
bination of slots from a huge number of possible combinations.
One possible solution is to enhance dynamic programming with
optional type
annotation
[12][13][14]. It is the kind of annotation that programmers optionally
embed into source code for indicating types of variables or function signatures.
Optional type annotation has no effect on the run-time semantics of the program-
ming language. However, it can be used for improving productivity and runtime
performance of dynamic programming languages by static type checking before
execution, JIT (just-in-time) compiler optimization and content suggestion in
IDEs (integrated development environments). Today, optional typing is a very
popular approach for introducing more stability and strictness to dynamic pro-
gramming languages while preserving their flexibility.
The idea of optional type annotation is not limited to the programming
world. In the world wide web, we also find various
microformats
for annotat-
ing pieces of HTML documents by using predefined, controlled vocabularies. For
instance, the
hCard
standard defines a set of class names for indicating contact
