Information Technology Reference
In-Depth Information
design Aspects
domain specific information
After discussing the abovementioned issues, we
can define ontologies at the lowest possible level
and generalize more complex ontologies (see
Figure 3). By doing so, it is important to think
about two important points: Firstly, the ontology
has to be implementable to work in an informa-
tion system and secondly, the ontology definition
has to be done by domain experts who define all
the concepts that constitute such an ontology.
The implementation needs to have positive and
negative validation tests in order to prove the
correctness of valid targeting range. Where the
positive validation is occurred within the range
of the targeting values and the negative validation
occurs within the input values that are invalid and
outside the range of targeting values.
The reader now might ask this: why should
we do that? Typically, in an information system
there are many layers where data are validated.
The first layer is the user interface, where the user
input is validated. However, even the user acts as
a validator, because he knows the domain and can
filter evidently wrong data. The next layer is the
business logic layer that deals with the data and
check the plausibility that might be repeated at
the lower layers consequently.
This approach is correct, if we are working
within an isolated system. But in SOA solutions,
there is a good chance that different applications
are using the same services, if these applications
share data then the following problem might ap-
pear: an application assumes that the stored data
are correct, and if the other application validates
the data at a lower quality level, there is a risk to
load invalid data. The only way to prevent this is to
validate the data at the target service. This means,
it is necessary that an object does not have only
properties, attributes and methods to be existed
rather it must be able to validate its own status!
The explanation of the term domain specific
information requires defining the word informa-
tion. Related to the topic of semantics, we define
information as data enriched by annotations. This
means that the phrase <>Peter</> represents data
while the phrase <firstname>Peter</firstname>
represents information. Though there are still
higher levels to take into account like knowledge
or wisdom, but at this stage, we will only deal
with information.
Semantics at the top level of the semantic
SOA-based model are used to describe information
while at the lower levels the difference between
data and information can't be recognized easily.
If the annotation of information is inaccurate
then it represents only data to the user and not
the user's desired information and this happens
when the annotation does not represent the data
in a sufficient way.
Example: If we consider this phrase:
<fn>Peter</fn>, the consumer of this information
can at first glance guess that the intended meaning
behind this phrase is “first name”, but he can't be
sure. And for a computer system the interpretation
is much harder, because the machine will not be
able to use knowledge to process and understand
the meaning of the annotation corresponding to
the sample data. A solution for this problem is
the involvement of semantic annotations and the
ultimate goal behind its usage is to allow machines
to interpret the date.
split the semantic Annotation
from the object
The semantic SOA-based model's validation layer
provides the semantic annotation activity. Typi-
cally, a Web Service itself has the responsibility of
adding the annotation. This means, the semantic
information is added like the syntactic information
in the Web Service response (Horrock, 2008). As
Search WWH ::
Custom Search