Transcript Web hypermedia cost estimation: further assessment and

Sánchez Cantón, Juan José
Blázquez González, Isaac
Alonso Caballero, David
•
Massive increased use of the Web since 1989
•
Web application. Classification:
– Business to business, Business to customer
– Web hypermedia application, Web software application and Web
application
•
•
Survey “Web-based projects” (Cutter Consortium, 2000)
– 53% of surveyed
functionality
– 84% of surveyed
– 79% of surveyed
– 63% of surveyed
delivered projects did not provide the required
delivered projects did not meet business needs
projects presented schedule delays
projects exceeded their budget
“Web engineering”: Murugesan & Deshpande, 2001:
- “the use of scientific, engineering, and management principles and
systematic approaches with the aim of successfully developing, deploying
and maintaining high quality Web-based systems and applications”
•
Application characteristics and availibity
– Integration of a higher number of distinct
components in Web.
– In common, except for a few application domain,
there is no need for availability 24/7/365
•
Technology and architecture:
•
Quality
– Web: distributed architecture applications. Network
gets importance
– Web: main quality drivers are reliability, usability,
and security.
– Common: time to market takes priority over quality
•
Structuring and manteinance
– Web: use of hyperlinks to build navigational structures.
– Web: maintenance cycles are frequent without specific
releases and often take days or even hours.
•
Disciplines, and people involved in development
- Web: usually wider range of people involved than
common
•
Legal,social and ethical issues
- Web: Content can be easier copied and distributed
without permission or acknowledgement of copyright
and intellectual property rights.
•
•
•
•
There is no standard for sizing Web applications as Web
companies vary widely
Numerous Web companies are too small to provide a
stable environment from which they can develop and
maintain their applications and rarely use standards like
CCM so the can`t reuse the experience in future projects.
Except for the Tukutuku database (Mendes, Mosley, &
Counsell, 2003), there are practically no existing
benchmark databases that can be used to obtain and then
compare Web effort estimates and to benchmark Web
projects’ productivity.
Despite the proposal of numerous Web development
methodologies, there is no single, unifying methodology
that entirely captures all the intricacies of Web
development projects and project management.

2 points of view:
◦ Dhyani, NG and Bhowmick (2001)
◦ Calero et. AI (2004):classified 385 measures from 40 studies according to a
framework called the Web quality model (WQM). This framework is structured
according to three orthogonal dimensions.(Web features,Web life-cycle processes and
Web quality characteristics:

General purpose:
◦ To predict the effort our development require to carry out the
customer requirements (black box metaphor)

Expert-based: it is often based on previous
experience with developing and/or managing
similar Projects . Braingstorming session y CMM
– drawbacks: rapidly evolution of technologies, the datas
on their own don’t give enough guarantee, Developers
and project managers are known for providing optimistic
effort estimates
•
To cope with uderestimation, it is provided a
formula:
E = (o + 4r + p) / 6
o,r,p: optimistic,realistic,pessimistic
estimate
•
AlgorithmicTechniques:
The inputs are function points, number of lines of
source code, number of Web pages, etc. and the
effort is the relationship between both sizes. It is
translated into a equation (the most popular is
WebMo - WQM):
E = a·S^b·CostDrivers :
a,b: constants
S: estimated size of an application
CostDrivers: cost drivers

Artificial Intelligence:
◦ In the last decade, it has been used as a
complement to the previous two categories. Some
examples: fuzzy logic, regression trees, neural
networks and cased-based reasoning (CBR)
◦ CBR: similar problems proved similar solutions
o Classification and Regression Trees (CARTS): use
independent variables (predictors) to build binary trees.
This abstract provide us a light revision of the
special characteristics of software projects based
on Internet and then define lines of performance
that allow to obtain a model of cost estimation
obtained from the study of several authors. Those
models have main importance on the market,
although the most popular are

Algorithmic.Techniques yet.




-“Cost Estimation Techniques for Web Projects”
Emilia Mendes. University of Auckland, New
Zealand
-“Página Web de
Almatali” <www.almatali.com.ar>
-“Página Web de Bournemouth University”. Dan
Snell < http://www.ecfc.unet.com/cost/index.htm>
-“Web hypermedia cost estimation: further
assessment and comparison of cost estimation
modelling techniques : Hypermedia and the World
Wide Web = Hypermedia and the World Wide Web”
MENDES Emilia, COUNSELL Steve y MOSLEY Nile