![](bg5.png)
When the information is already located, we only
need to serialize it and submit it through a TCP
socket. With quite little additional upgrade on this
code and the one responsible for data serialization,
we could provide more options to recover different
parts of the document. Nevertheless, no more fea-
tures will be developed so far in order to get feedback
from users to improve further developments. So far,
just support for recovering a certain node in the re-
quested document is provided. As this project gets
support and on-line experience provides feedback, we
will develop support for more advanced queries.
5 AN EXAMPLE
In order to test the already developed project, we de-
ployed several experiences. As our server is so far the
only one with these capabilities, the first step in this
testing process was to clone some web pages from the
Internet to our own server. The first obvious test is to
request for some particular element from the DOM
tree. This test was successfully accomplished with no
special incidences.
The first serious attempt to test our system consists
of a simple web page, as explained in the introduction,
where contents were address using our new schema;
we, therefore, developed a web page made, mainly, of
frames.
By using this concept we can compose any web
content as they can be accessed by mean of its DOM
tree. It is important to notice that the DOM tree cor-
responding to this source code, at first glance, may
seem to be quite simple, but we may be dealing with
a huge structure.
One of the main constrains for web design using
this technique is due to limitations to use of exter-
nal addresses. To access external resources, the only
suitable tag included in the HTML 4.01 specification
is the tag f rame so we are forced to use it. In this
situation, it would be very suitable to be able to use
other tags, such as div to reference external resources.
As this tag provides a certain position to place certain
contents and it is widely supported for laying out with
CSS, it would quite useful to be able to use it as con-
tainer for external resources.
But the most interesting test was about its func-
tion as a real interoperability booster. As previously
stated, the aim in this project is to allow the content in-
terchange in the World Wide Web. The expected way
this tool is going to be use is to include contents from
several web resources in just one single new web re-
source, mainly, a web page. To achieve this goal, we
designed a web page with contents both, from its own
and externally located. In the insofar web pages, it is
a quite usual practice to include as contents for frames
someone else’s full web page; an application for this
project is to allow web designer to include just a sin-
gle piece of an existing web page in their own web
sites.
Several concerns may arise on content presenta-
tion. DOM tree nodes do not just include informa-
tion about data but also about presentation so prob-
lems may arise if no measures are taken. During the
design of the test page some resources had to be de-
voted to fix presentation bugs. These problems may
get worse when dealing with contents poorly format-
ted or even wrong. So we must bear in mind problems
on placing external HTML code on new pages.
The key concept to achieve a higher level of con-
tent reusability involves merging different DOM trees
from several sources into a new single tree where all
the information is stored. Besides, these transforma-
tions are not in the scope of the final client, which
just will ask for a simple web resource, neither in the
scope of the HTML designer that just ask for con-
tents in a simple way regardless of where they are al-
located. The only needed tool to build up contents is
the use of frames as previously presented.
It is important to notice that this goal could be
accomplished just by using other already developed
technologies, but by using this technique we can pro-
vide these contents with no need of dynamic contents
or client side processing.
6 CONCLUSIONS
The main idea of this paper is to propose an alter-
native way to reuse contents in a cost-effective way.
As the only needed changes are on to the software in-
stalled in web serves, not in the information itself, and
no changes in client side, this technique can spread
quickly with nearly no efforts in the short term and
support legacy information systems. This proposal
deals with is related to partial recover, server side pro-
cessing of static contents. In this particular situation
we could not find any suitable solution in terms of
simplicity with short time-to-market.
The main goal in our proposal is to provide web
developers with the chance to reuse contents from al-
ready existing web contents. This is possible with
no changes on the contents themselves. To be able
to perform this operation we mix different DOM tree
in a single new tree. The undertaken steps to fulfill
our proposal are related with information recovery by
sending pieces of files on HTTP channels. It, there-
fore, is possible to build up contents just by using
WEBIST 2008 - International Conference on Web Information Systems and Technologies
288