Professional Documents
Culture Documents
ABSTRACT by external sources. Today, the way that people build soft-
Web mashups are becoming the main approach to build ware has radically changed, from implementing from scratch
Web applications. Current approaches to enable compo- to searching for already packed functionality that alone or
nent selection include description-based techniques and so- combined, fully or partially satisfies each requirement.
cially generated metadata. The explosive growth of APIs API Web catalogs provide, besides documentation, social
makes increasingly harder selecting appropriate components information about the real use of APIs on the registered
for each mashup. Unfortunately, description-based tech- mashups. In our previous work [11], we have argued the
niques rely heavily on the quality of authors’ information, need to combine these two sources of information, where
and social-based approaches suffer problems like “cold-start” description-based techniques can be leveraged by social in-
and “preferential attachment”. This article proposes (1) formation. The results of this combination, allows to dis-
two new measures of socially ranked fitness of candidate cover candidates that would have passed unnoticed because
components, (2) an API functional taxonomy using For- of their poor quality descriptions or their low popularity. It
mal Concept Analysis based on descriptions, and (3) a com- is important to mention that we cannot use only the social
bined approach that improves description-based techniques information, because as has been showed in previous work
with these social ranking measures. We use social rank- [3] this leads to the cold start problem for new APIs, and
ings based on past (co-)utilization of APIs: WAR (Web API makes the discovery process to exhibit a preferential attach-
Rank) measures API utilization over time, and CAR (Co- ment trend.
utilization API Rank) measures its co-utilization with other Typically, mashups are built with more than one API.
APIs. The measures and the combined approach are illus- These APIs are iteratively selected and previous selections
trated with a case study using the well-known Web APIs influence current ones. Then, when composers are discover-
catalog ProgrammableWeb 1 . A prototype tool allows it- ing APIs, they must consider which other APIs they have
erative discovery of APIs and assists the mashup creation already selected. Developers community can help composers
process. to discover the most appropriate APIs based on their past
co-utilization. This information is obtained from the regis-
tered mashups and we can exploit it to support the discovery
Keywords process.
mashup, formal concept analysis, recommendation system, Currently, most techniques have tackled the discovery pro-
social network cess using separately social and semantic networks. In this
work, we present an approach to use both networks to enrich
1. INTRODUCTION and improve the discovery process results.
Our main contributions presented in this paper are
Mashups are becoming the de facto approach to build
customer-oriented Web applications, by combining opera- • a novel iterative approach to discover APIs to build
tions on several Web APIs (Application Programming In- mashups driven by social utilization and co-utilization
terfaces) into a single lightweight, rich, customized Web over semantic techniques,
front-end. They allow to construct complete applications by
searching, composing and executing functionality provided • a new model to represent APIs and their collaboration
1
to create mashups,
http://www.programmableweb.com
• a web tool that support this approach and