Dear respondents,

        Thank you for your reply and now I send a copy of my results to you.

Chapter 7       Conclusion

7.1     The strengths and weaknesses of the five configurations

7.1.1 The strengths and weaknesses of 'pwWebSpeak'
'pwWebSpeak' is the cheapest of all the configurations.  This may encourage
VIP to buy it if their screen readers do not work well with a standard Web
browser.  Moreover, it can work with a speech synthesizer or a sound
blaster card and is compatible with Windows 3.x and Windows 95.  This fits
most VIP computer systems.  In addition, it supports an audio application
so that VI users can listen audio programmes on the Web.  Regarding
functionality, it has the ability to notify VI users when a hyperlink is
crossed, handle image map links, give a link list of frames (valid for
pwWebSpeak 1.43) and provide information about the page layout.  Given its
ease of operation, users can easily access hyperlinks and fill in forms.
For its responsiveness, signals are given when downloading Web pages and
files.  However, it cannot access newsgroup, FTP and Gopher links and Java
applets.  Also, it cannot provide sufficient information to access tables,
lists and most text formats.
7.1.2 The strengths and weaknesses of 'VIPinfoNet'
        As with 'pwWebSpeak', 'VIPinfoNet' has similar functions to access
hyperlinks, image maps and frames.  In addition, it has the ability to read
each cell of a table along with both the row and column title of that cell.
 Given its ease of use, it is very easy to access hyperlinks.  Regarding
its responsiveness, it also gives responses when downloading Web pages and
files.  However, it cannot handle all desirable links (e-mail, newsgroup,
FTP and gopher) and Java applets.  Moreover, it cannot provide enough
information to access page layouts, lists and most text formats.
Furthermore, it cannot work with a speech synthesizer or an audio program
(the speech is through a soundblaster card).
7.1.3 The strengths and weaknesses of 'Hal 95'
        Regarding technical support and documentation format, 'Hal 95' is much
better than other configurations.  With functionality, it can access
desirable links and Java applets which contains text and sound.  In
addition, it can provide more information about text formats and lists.  In
contrast with talking browsers, it lacks the ability to handle image maps
and frames.  Moreover, it is difficult to access hyperlinks and forms and
cannot provide enough information to access page layouts.  In addition, it
cannot give responses when downloading Web pages.
7.1.4 The strengths and weaknesses of 'Jaws'
        Similarly with 'Hal 95', the manufacturers of 'Jaws' provide a variety of
documentation formats about the product.  Also, it can access desirable
links and Java applets which contain text and sound, and can provide more
information about text formats and lists.  In addition, its compatibility
with hardware and software is as good as 'pwWebSpeak'.  However, apart from
accessing forms with slight difficulty, it has all the drawbacks of 'Hal
95'.
7.1.5 The strengths and weaknesses of 'outSPOKEN'
        'outSPOKEN' is much more reliable since no crashes have occurred and no
bugs  have been found while testing.  Regarding other criteria, its
strengths and weaknesses are similar to 'Jaws'.
7.1.6 Recommendation
         Both 'pwWebSpeak' and 'VIPinfoNet' have advanced features to access
essential hyperlinks, especially image map links, and frames.  In addition,
they are easy to use and have good responses.  Therefore I conclude that
the talking Web browsers are superior to the tested screen readers working
with a standard browser for accessing the Web.  Although they cannot handle
most desirable hyperlinks and Java applets, and cannot provide more
information about text formats, these elements may not be as important as
essential hyperlinks and frames since accessing most desirable hyperlinks
is required to link them to other applications, most Java applets are not
accessible and  information about text formats may sometimes hinder VI
users from reviewing the content of  a text.  Considering  cost and
compatibility, 'pwWebSpeak' is more preferable to 'VIPinfoNet'.
7.2 Findings from questionnaires
It seems that most respondents prefer to use products from their own
country, possibly because of the ease of getting information about local
products and the advantage of local technical support; however; it is worth
considering whether they can access support and information about foreign
country's products as easily as their local companies.  Most respondents
preferred to use a textual screen reader because it is stable, logical,
controllable and effective; however, these respondents had a tendency to
shift to GUI screen readers because of the popularity of GUI software
programs.  Most respondents considered that functionality and compatibility
were very important in a screen reader (i.e. the functions the product
provided and how compatible it was with software and hardware).  In this
group of respondents, activities like browsing the web and communication
through e-mail are very popular.  However, quite a lot of respondents had
problems in accessing the Web.  Their problems come from their
inexperience, the design of screen readers, Web browsers, Web pages, and
Java applications.  Using a talking Web browser can overcome the
limitations of a screen reader working with a Web browser; however, the
design of Web pages and the design of Java applications are also crucial.
Reading news was not very popular in this group of users.  The reasons may
be that messages in newsgroups may not be as informative as that on the Web
and also users normally employ a newsreader to communicate with newsgroup
users only.  Chatting on the Internet was also not popular among these
respondents.  The cause may be related to the performance of a screen
reader with an IRC program.  Other Internet tools such as telnet and FTP
were not often used among these respondents.  This may be due to the fact
that most tasks can be done through the Web.  In short, not many
respondents use newsreaders, IRC, FTP or telnet programs relative to Web
browsers or e-mail programs.  Particular difficulties with these programs
may not be shown.
Their strategies for enhancing access to the Internet are mainly to join
mailing lists for absorbing other users' experience, familiarise themselves
with their tools and standardise the design of Web pages.  However, they
appeared not to be aware of other accessible systems that may help them to
access the Web (e.g. WAB).