A NEW APPROACH TO THE AUTOMATIC WEB

ACCESSIBILITY

Juan Manuel Fernández

, Vicenç Soler

1,2

Dept. Microelectrònica i Sistemes Electrònics, Universitat Autònoma de Barcelona, Cerdanyola del Vallés, Spain

Ciber-BBN: Bioengineering, Biomaterials and Nanomedicine. Campus UAB, Bellaterra, Spain

Jordi Roig

Dept. Microelectrònica i Sistemes Electrònics, Universitat Autònoma de Barccelona, Cerdanyola del Vallés, Spain

Keywords: Web Accessibility, World Wide Web Consortium, Web Standardization, HTML, CSS.

Abstract: The number of Web pages created in an inaccessible manner is very high, and to convert them to be

accessible can be impossible. We present a study of different tools that can help to obtain a total Web

Accessibility. With this study, we can see the advantages of the automatic correction and the possibilities

that this kind of tools offer. The result of the study shows how the applications adapt the Web pages to both

the World Wide Web Consortium HTML grammar and to the Web Content Accessibility Guidelines. With

this study we find an automatic way to improve Web Accessibility and to solve the problems that it can

offer to the Web pages authors. This way consists of combining two tools. This combination gives us results

that allow disabled people to use a system that can convert the entire Web in an accessible way.

1 INTRODUCTION

In this paper we present a study of different tools

that can improve the Web Accessibility. With this

analysis we try to find a tool, or a combination of

tools, that can help the e-integration. The potential

of the Web in the integration of disabled people is

very high, but in order to exploit it, it is necessary

that Web page developers use correct the grammar

of the HTML (W3C, 1999b) and the accessibility

guidelines (W3C, 1999b). Both of them are dictated

by the World Wide Web Consortium (W3C). In the

case of the accessibility, it exists a group of especial

interesting, called the World Accessibility Initiative

(WAI), inside of the W3C.

The use of the HTML grammar in the Web is

very low. The lack of knowledge of this grammar,

on the part of the Web page designers, has the origin

in the fact that they design to a specific Web

browser but these Web browsers are not standard

(The Web Standards Project, 2007). If the use of the

correct HTML grammar is very low, the use of the

norms developed for disabled people is very poor.

A great number of Web designers do not know the

special needs of the disabled people.

The purpose of this study is to arrive to a strategy

that can allow us to transform Web pages to the

W3C and WAI normative. More specifically, the

norms we are interesting in are the Web Content

Accessibility Guidelines (W3C, 1999b) (WCAG).

These guidelines mark what and how it is correct to

present information by the Web. In this paper, we

compare two tools that work in this manner.

Accessible Web is a tool that works with

WCAG (ACTAW) (Fernández et al., 2007). This

tool works with HTML, XHTML and CSS, and it

applies the WCAG guidelines to a Web document.

The other tool that we use in this study is the

HTML TIDY project (Dave Raggett, 2003). This

application tries to adapt HTML and XHTML code

of the Web page to the correct grammar.

The study compares the tools in two manners.

First of all, we concentrate the study on the HTML

grammar. Second, we analyze the results under the

perspective of the Web Accessibility. With these

studies we can try to find a possible way to improve

Web Accessibility.

393

Manuel Fernández J., Soler V. and Roig J. (2008).

A NEW APPROACH TO THE AUTOMATIC WEB ACCESSIBILITY.

In Proceedings of the Tenth International Conference on Enterprise Information Systems - HCI, pages 393-396

DOI: 10.5220/0001713303930396

 SciTePress

2 TOOLS USED IN THE TEST

2.1 ACTAW

The aim of the ACTAW is to improve Web

Accessibility by applying the WACG to a Web page.

The result of applying ACTAW to a Web page is a

document with a better HTML structure and better

accessibility. Moreover, ACTAW can modify the

styles implemented with CSS and offer styles with a

high colour contrast.

This software implements an HTML analyzer

and a standardization module. The parser analyzes

the HTML tags and attributes of a Web page, but it

is not based on its structure or the correct W3C

grammar. This fact helps to understand all kind of

Web pages.

The standardization module has the objective of

improving Web Accessibility. This module applies

the WACG to the documents as well as it is possible.

It is impossible to apply some guidelines, like the

one to use an easy language.

2.2 HTML TIDY

HTML TIDY is an Open Source solution to obtain a

correct HTML document. This tool transforms the

HTML original source from a Web page to a correct

HTML code that follows the W3C grammar, but its

effects reach the Web Accessibility as well.

The document, after the execution of HTML

TIDY, contains valid HTML tags and it corrects the

structure of the document if it is not valid. It means

that HTML TIDY can determine if the use of the

elements of that markup are well positioned and

replaces them by the correct solution.

2.3 W3C Validation Service

The first point of our interest is to measure the

correctness of the Web pages results from ACTAW

and HTML TIDY. To know if they are correct, it is

necessary the W3C Validation Service (W3C,

2007a). This online tool analyzes the document and

creates a report with the points where the HTML, or

XHTML, grammar is not accomplished.

2.4 TAW Validation System

Test de Accesibilidad Web (TAW) (Fundación

CTIC, 2007) is a tool developed by CTIC

foundation, with the support of the Spanish Ministry

of Education and Science. The Spanish W3C office

forms part of this foundation.

The analyzer validates the different guidelines of

WCAG 1.0. The result offers the number of errors

organized in the three priorities, and the way the

error can be validated. The report is very exhaustive

and the summary offers a quick view of the

problems found and if it needs the interaction of the

designer.

2.5 Test Bench

The elements of study are the real Web pages. We

use a group of sixty real Web pages selected

randomly in an automatic way. This test bench is

very heterogeneous and contains a different kind of

contents. An interesting point is the fact that it

contains Web pages designed in XHTML or HTML

and that they can be correct or not.

3 PROCESS OF THE ANALYSIS

3.1 Study of the Test Bench

First of all, we need to know what kind of errors has

every Web page to compare with the results of the

process. To solve this point it is necessary an

exhaustive study of the test bench with the focus on

HTML grammar and the WACG.

Errors from the original Web Pages

3668 8394

2000

4000

6000

8000

10000

Number of Errors

W3C Grammar

Accessibility

Figure 1: Errors from the Original Web Pages.

In Figure 1, we can see the totality of the

grammatical and the accessibility errors. On

average, it is the same those 23 grammatical errors

by Web page, and 139.9 accessibility errors. This

difference allows us to see the low knowledge of the

Web Accessibility development.

The priority 2, which has guidelines affect the

styles used, the creation of tables and scripts, is the

biggest with difference (See Figure 2). These points

are very difficult to accomplish if you do not have

the knowledge of the correct manner of doing it.

And also, in this priority we can find the correct use

ICEIS 2008 - International Conference on Enterprise Information Systems

394

of the HTML tags, in other words, the use of the

correct grammar.

Errors by Web Accessibility Priority

74,71%

11,23%

14,06%

Priority 1

Priority 2

Priority 3

Figure 2: Errors by Accessibility Priority.

3.2 Applying HTML TIDY

The second step in the process is to apply HTML

TIDY to the test bench. The improvement of the

accuracy to the W3C grammar is very high (see

Figure 3). With a 58.64% of improvement, HTML

TIDY has very good results. This is the same that to

have 11 errors by Web page on average. HTML

TIDY deletes non-standards tags and attributes, one

of its main problems, and it repairs structures which

do not accomplish the standard.

Grammatical Errors

3668 1517 2548

500

10 0 0

15 0 0

2000

2500

3000

3500

4000

Original

HTML TIDY

ACTAW

Figure 3: Grammatical Errors.

HTML TIDY does not try to solve accessibility

errors, but it solves some inaccessible situations due

to the final correct structure. It is surprising the fact

that the best improvement is in the priority 1, which

is 16.44%, and only 5.39% at the priority 2. The

Figure 4 shows the results of applying HTML TIDY.

Evolution of Accesibility Errors

6271

943

1180

879

5933

986

1000

2000

3000

4000

5000

6000

7000

Priority 1 Priority 2 Priority 3

Original

HTML TIDY

Figure 4: Evolution of Accessibility Errors.

3.3 ACTAW Test

ACTAW offers an improvement of the 30.53% of

the correctness of the document in the test of the

W3C without delete any tag (see Figure 3). We will

discus about it in section 3.4.1.

Where ACTAW has excellent results is in the

field of Web Accessibility. The tool obtains a

96.19% of improvement in the errors from priority

1. This is the same than having 0.75 errors by Web

page. If we observe the results in depth we can see

that the 71.67% of the Web pages do not have any

priority 1 error and can obtain the level A in

conforming with the WCAG. If we look the rest of

the priority, we can see a reduction of the errors. An

86.85% of improvement is obtained for the priority 3

which is very impressive.

Evolution of Accessibility Errors

45 124

9431180

6271

5280

500

10 0 0

15 0 0

2000

2500

3000

3500

4000

4500

5000

5500

6000

6500

7000

Priority 1 Priority 2 Priority 3

Original

ACTAW

Figure 5: Evolution of Accessibility Errors.

Figure 5 shows the evolution of accessibility

errors between the original Web pages and the

results of ACTAW, and we can see that priority 2

has a low improvement comparing it with the rest of

priorities. The reason of it is the complexity of the

problems in this priority. We can emphasize two

points: the design of the scripts and the use of the

tables.

A NEW APPROACH TO THE AUTOMATIC WEB ACCESSIBILITY

395

3.4 Comparative of the Results

3.4.1 HTML Grammar

The differences between the two tools are

considerable, but it is necessary to comment one of

them. ACTAW does not delete the tags which do not

form part of the W3C grammar. In opposition to

this, HTML TIDY deletes them. Both actions are

correct but it is difficult to know which one is better.

The best action is to replace the incorrect tags by

the equivalent ones. The existence of equivalent tags

is not possible in all the cases and the number of

non-standard tags is very high. The option to delete

the tags and attributes can make the Web page not

working properly. It is important if the Web page is

a graphical interface of an online application.

The best solution for problems with incorrect

tags and attributes is HTML TIDY. The process of

validation shows a reduction of 2151 errors. In this

field, ACTAW offers an improvement of the 30.53%

(see Figure 2).

3.4.2 Web Accessibility

The results offered by the tools in this field are very

good; in all the priorities they reduce the amount of

errors. We have commented the good results of the

tools, but now we are interested in priority 2. In this

priority both tools cannot offer excellent results (see

Figures 3 and 4), but ACTAW improve in an

11.01% the results of HTML TIDY. This

improvement can be low but it is compensated by

the results on the other priorities.

4 ACTAW & HTML TIDY

After comparing the different tools we are going to

study the combination of both. The amount of

grammatical errors shows us that the best

combination is to correct grammar errors and after

the accessibility errors. We reduce 2512 errors if we

apply ACTAW and after it HTML TIDY. This is an

improvement of the 73.34%. In comparison with the

application of HTML TIDY alone, we have

improved a 14.70%.

If we take the Web Accessibility errors we obtain

the same results. It is not important the order of the

application of the tools. The improvement of the

errors is 32.95%, which is a very good result, but

lower than the ACTAW results.

5 CONCLUSIONS

We have seen a comparative between two tools that

work transforming automatically Web pages. Both

of them try to apply the standards of the W3C.

HTML TIDY works on the correction of the HTML

grammar and ACTAW transforms Web pages to be

accessible to disabled people.

The comparative shows us that the tools work

correctly outside of its field and have great results on

its field. After studying the tools, we have studied

the result of their combinations. The results show

that the combination of ACTAW and HTML TIDY

can offer an improvement of the 73.34% in the

correction of the HTML source. If we think in Web

Accessibility, the combination can improve a

32.95%. This combination offers a very good

improvement of the Web.

In conclusion, we can say that we have found a

possible automatic solution to improve Web

Accessibility. The combination of different tools can

allow us to obtain a way to correct the low index of

accessible Web pages.

REFERENCES

Fundación CTIC. “Test de Accessibilidad Web 3.0”.

http://www.tawdis.net/ (Retrieved on September 2007)

Dave Raggett, 2003. “TIDY Project”.

http://www.w3.org/People/Raggett/tidy/ (Retrieved on

March 2007)

Juan Manuel Fernández, Vicenç Soler and Jordi Roig,

2007. Automatic Conversion Tool for Accessible

Web. In Proceedings of the 3rd International

Conference on Web Information Systems and

Technologies. (Webist’07)

The Web Standards Project. http://www.webstandards.org/

(Retrieved on May 2007)

World Wide Web Consortium (W3C), 1999a. HTML 4.01

Specification http://www.w3.org/TR/html401/

(Retrieved on June 2007)

World Wide Web Consortium, The W3C Validator Team,

2007a. “The W3C Markup Validation Service”.

http://validator.w3.org/ (Retrieved on June 2007)

World Wide Web Consortium (W3C), 1999b. Web

Content Accessibility Guidelines 1.0,

http://www.w3.org/TR/WCAG10/ (Retrieved on June

2007)

ICEIS 2008 - International Conference on Enterprise Information Systems

396