A NEW CONCEPT OF THE SEARCH ENGINE

FOR THE WEB API

Yuka Obu

, Minoru Sasaki

and Tatsuhiro Yonekura

Graduate School of Science and Engeneering, Ibaraki University, Ibaraki, Japan

Department of Computer and Information Sciences, Ibaraki University, Ibaraki, Japan

Keywords: Web service, Web API, Software as a Service, Standardization of the Web.

Abstract: The World Wide Web has become a platform for many kinds of Web services. Millions of Web users have

interests on the Web services, and there are still a lot of possibilities in it. We have focused on the capability

of the Web, and been exploring the new demands and services on the World Wide Web. In this paper, from

the viewpoint of the standardization of the Web API and the Web service that searches the best Web API for

user’s demands, we propose a framework of information that is required for the standardization of the Web

API, and finally propose the Web service that searches Web APIs for mash-up to create a new Web service.

1 INTRODUCTION

The World Wide Web (Web hereafter) has become a

platform for various kinds of Web services. Over the

latest years, many applications – map viewer,

calendar, e-mail, travel reservation, catalogue,

shopping, game, and so on - have been developed on

the Web. Accordingly, huge amount of Web users

have great interest on the Web service, and there are

still a lot of possibilities in it. We have focused on

the capability of the Web, and been exploring the

new demands and services on the Web.

Nowadays, users have a demand to use the Web

service as they want, and that demand is leading to

the development of new functions and new services.

Software as a Service, which is referred as SaaS, and

the Web API are examples of new technology for

the Web services. SaaS is a software that user can

use via the Internet. One of the advantages of SaaS

is that users do not need to install any software to

use the function of them and they can use only the

function they need. Web API is open to the public,

and users can create their own Web service or Web

site as they prefer by using the APIs. Especially, the

Web service which is created by utilyzing more than

one API is called mash-up site. User can also

construct various types of SaaS as mash-up sites by

using the Web API. Therefore, if there is no SaaS

that meets user’s demand, user can create personal

one. Though the Web API can meet the users’

demands flexibly, because it is constructed by using

JavaScript or XML, it is difficult for the non-

technical oriented users to operate the Web API.

Moreover, because the Web API is not standardized

in terms of the protocol, even the technical oriented

users need some time to create new Web sites by

using the Web API.

There is another problem to use the Web API.

That is, how does user know which is the best API

for creating a new Web service? There are many

APIs on the Web. The output to the input is slightly

different according to the API. The grammar of the

API is not standardized. Therefore, to select which

API to use comes to introduce costs to the user.

In this paper, from the viewpoint of the

standardization of the Web API and the Web service

that searches the best Web API, we propose a

framework of information that is required for the

standardization of the Web API. And we also

propose the concept of the Web service that searches

Web API and creates a new Web service using it.

2 RELATED RESEARCHES

There are some kinds of the Web service of creating

new Web site by using the Web API. For example,

Yahoo Pipes (Yahoo, 2007), PopFly (Microsoft,

2007), Google gears (Google, 2007) are new Web

services that user can use the Web API on them.

Google gears is an open source browser extension

that enables web applications to provide offline

343

Obu Y., Sasaki M. and Yonekura T. (2008).

A NEW CONCEPT OF THE SEARCH ENGINE FOR THE WEB API.

In Proceedings of the Fourth International Conference on Web Information Systems and Technologies, pages 343-346

DOI: 10.5220/0001531503430346

 SciTePress

functionality using the JavaScript APIs. By using

this API, user can create new applications that can

store and serve application resources locally. Yahoo

Pipes and Microsoft PopFly are composition tools to

aggregate, manipulate, and mash-up content from

around the web. Some commands can be combined

together to create output that meets users’ demands.

One of the features of them is that the user can

compose the new web site by using GUI based

operation. These examples are new technology on

the Web. But in order to use them, user must have

somewhat of knowledge about JavaScript and Web

APIs. In other words, if a non-technical user wants

to use these services, it takes a while to learn the

usage of them.

For non-technical oriented users, visual

programming tools that allow user to create

programming code easily have been developed. Islay,

an interactive animation-authoring tool based on the

state-transition diagram, was first proposed in 2005

(Okamoto, Kamada, Nakao, 2005) (Okamoto,

Shimomura, Kamada, Yonekura, 2006). Islay uses

the state-transition diagram to make authoring

interactive animation intuitively comprehensible. By

using the authoring tool, non-programmers can

define the animation of the characters. As a side

effect, the user may learn how to define dynamic

objects while having fun with animation. The

GUEST (Graphical User interface Editor by State

diagram) (Obu, Yamamoto, Yonekura, 2007) is a

Firefox (Mozilla org. 2007) extension that enables

users to define the behavior of the browser using a

state-transition diagram. Using the GUEST, the user

can define behaviors easily even if he or she has no

programming experience.

Web API Search (Web API Search, 2007) and

API Compare-and-Matching Service (Metadata,

2007) is a search engine for the Web API. On the

Web API Search, user inputs some keywords, and

then retrieval results from the database are displayed.

API Compare-and-Matching Service provides

functions of the Web API search related to the key

word and the Web API. It is possible to review the

Web APIs by leaving history of inspected API. By

using this search engine, user can create an outline

of the planning document of the mash-up Web

service. These services are helpful to find a suitable

Web API. But the result is a kind of catalogue.

Therefore the user has to understand the functions of

the Web API and how to use it in order to create

new Web site or new mash-up site. Moreover, if the

user needs some Web APIs, user must consider

about the relationship between the APIs.

3 NEW CONCEPT OF SEARCH

ENGINE FOR WEB API

3.1 New Search Engine for the Web

API

The current search engines for the Web API

mentioned above are a kind of searching the

catalogue of the Web APIs. The users input the

keyword according to input, output, or function that

they need. And they needs to have technical

knowledge and comprehend the usage and functions

of the API. Accordingly, we propose a new concept

of a search engine that has more high performance.

Our aim is that user can search the Web API

without considering about the details regarding

functions and codings. Our new concept takes only

one step to find the best Web API, while the current

system needs user to compare the APIs.

In our new concept, the search engine has to be

more semantically. Hence, some data for the system

are defined as below.

(1) Input Data: Data that user input to the Web

API. For example, Japanese, English, and so

on.

(2) Inputdata Type: Data type of input. For

example, sentence, a word, and so on.

(3) Process: The function of the Web API. For

example, translating sentences, looking up

the words from a dictinary, and so on.

(4) Output Data: Data that the user needs.For

example, English, German, and so on.

(5) Output Data Type: Data type of output. For

example, sentence, words, and so on.

By using these data above, the system can find

the best Web API. Let’s assume that we create the

translation service from Japanese sentence to

German sentence. Input data, input data type,

process, output, and output data type are Japanese,

sentence, translation, Germany, and sentence,

respectively. Then the system show results the

translation Web API. Moreover, if there is no Web

API to translate Japanese sentence to German

sentence, we might employ the two possible APIs:

the one that translates sentence from Japanese to

English, and the other that translates English to

German. Then the system searches these two steps

from input data listed above (Figure 1).

WEBIST 2008 - International Conference on Web Information Systems and Technologies

344

Figure 1: Outline of the search engine for the Web API.

Figure 1 is an outline of the search engine for the

Web API. Semantics analyzer and indexer create the

index database (Index DB). Indexer extracts

keywords from the Web API, and input data, API,

and keyword are registered into the index DB as an

inex. In the index DB, some attribute is also

registred. This attribute indicates what kind of input

data is used for what kind of Web API.

3.2 Standardization of the Web API

In order to realize the system which is mentioned

above, the standardization of the Web API is

necessary in terms of their protocol. Semantics

analyzer in Figure 1 may handle to decide types of

data that the API uses, and data format to input or

output to be defined. And it needs the format that

represents how the API processes from the input to

the output.

The standardization may lead the Web APIs to

mash-up each other more flexibly, and it becomes

benefit for users and developper of the Web API.

And it will be the great forward push of the new

World Wide Web.

3.3 An Example of the Web Service to

Create Mash-up Site with Web API

We are designing a new Web service that searches

the Web APIs and creates a new mash-up site.

Figure 2 shows an outline of the service. The flow of

processing of the service is as follows.

(1) User inputs data that is needed for the new

mash-up site.

(2) The search engine searches the Web API.

(3) The result is shown as a state-transition

diagram.

Figure 2: Concept of the service.

This service shows the result by using a state-

transition diagram. By using a state-transition

diagram, user can comprehend the Web API

intuitively, and it is easy to understand the flow of

processing and change of the data. When the user

arranges the process, he re-edits the state-transition

diagram instead of the source code.

Figure 3: Screen shot of the current search engine.

Figure 3 is a screen shot of the current search

engine of the Web API. In this search engine, there

are two text-input edit boxes; these are for input data,

and output data of the Web API. When these two

data are input and the button is clicked, the Web API

search is executed. As a result of the search, some

Web APIs that can be suited for the input and output

data are listed.

As an example, shown in figure 3, input data is

PC, and output data is price. Then the Web API

names as results are Rakuten Ichiba (Rakuten, 2007)

input

process

output

APIs

User’s

inpu

Process 1

Process 2

output

(1)

(3)

Site with

Open API

Clorer Control

Semantics

analyzer and

Indexe

New keywords

Cache DB

Index DB

Keywords for Input

data and Output data

results

Searching

robots

(2)

World Wide Web

A NEW CONCEPT OF THE SEARCH ENGINE FOR THE WEB API

345

and Wikipedia (Wikipedia, 2007). This means if the

user wants to search price by using PC as an input

data, the user can use the Web API of the Rakuten

Ichiba, or that of Wikipedia.

4 CONCLUSIONS

We proposed a new concept of the search engine and

the standardization of the Web API. And an example

of the Web service to create mash-up site with the

Web API is proposed. By using our new concept,

more semantic search engine may be realized, and

many users may be able to find the best Web API

they want and create a new mash-up site easily. This

will lead the Web APIs to mash-up each other more

flexibly, and it becomes benefit for users and

developper of the Web API and will be the great

forward push of the new World Wide Web.

ACKNOWLEDGEMENTS

We would like to appreciate all of our colleagues

participating in this project, namely, Professor

Masaru Kamada of Ibaraki Univ. and Assoc.

Professor Shusuke Okamoto of Seikei Univ. and

Mizuaki Yamamoto of Hitachi Co.Ltd. We give our

thanks to Mr. Osamu Miyamoto for his Funds. This

study was partially supported by the JSPS Grants-in-

Aids No. 18300027.

REFERENCES

Yahoo, 2007. Yahoo Pipes. In http://pipes.yahoo.com/

pipes/

Microsoft, 2007. PopFly. In http://www.popfly.ms/

Google, 2007. Google gears. In http://gears.google.com/

Okamoto, S., Kamada, M., Nakao, T., 2005. Proposal of

an Interactive Animation Authoring Tool Based on

State-transition diagram. Vo.46, No.SIG 1(PRO24),

pp19-27. Information Processing Society of Japan. in

Japanese.

Okamoto, S., Shimomura, T., Kamada, M., Yonekura, T.,

May 2006. Programming with Islay, an Interactive

Animation Authoring Tool. Vo.47, No.SIG 6(PRO29).

Information Processing Society of Japan. in Japanese.

Obu, Y., Yamamoto, M., Yonekura, T., Kamada, M.,

Okamoto, S., 2007. Exploring State-transition

diagram-based Web Browser Programming. In

CYBERWORLDS 2007, International Conference on

Cyberworlds. Cyberwrolds Press.

Mozilla org., 2007. Mozilla Firefox. In

http://www.mozilla.or.jo/products/firefox/

Web API Search, 2007. Web API Search. In

http://kommy.s254.xrea.com/WebAPI_Search.php

Metadata Incorporated, 2007. API Compare-and-Matching

Service. In http://www.api-match.com/

Rakuten Ichiba, 2007, In http://www.rakuten.co.jp/

Wikipedia, 2007, In http://ja.wikipedia.org/

WEBIST 2008 - International Conference on Web Information Systems and Technologies

346