Voice XML: Chapter 10 Exercises

VoiceXML: Introduction to Developing
Speech Applications

10-1 Write two versions of a VoiceXML form that solicits first name and age from a job applicant:

Test exercise 10.1

The VoiceXML form should be application driven.

<?xml version="1.0"?> <!DOCTYPE vxml PUBLIC "-//BeVocal Inc//VoiceXML 2.0//EN" "http://cafe.bevocal.com/libraries/dtd/vxml2-0-bevocal.dtd"> <vxml version="2.0" xmlns = "http://www.w3.org/2001/vxml"> <!-- Exercise 10.1.A System directed dialog --> <form id = "job_application_form"> <block> <prompt> Welcome to the job application form </prompt> </block> <field name = "name"> <prompt> What is your name? </prompt> <grammar type = "application/grammar+xml" root = "applicant_name"> <rule id = "applicant_name"> <one-of> <item> sam </item> <item> fred </item> </one-of> </rule> </grammar> </field> <field name = "age"> <prompt> How old are you? </prompt> <grammar type = "application/grammar+xml" root = "applicant_age"> <rule id = "applicant_age"> <one-of> <item> twenty </item> <item> thirty </item> </one-of> </rule> </grammar> </field> <filled> <prompt> Your name is <value expr = "name"/> Your age is <value expr = "age"/> Good Bye </prompt> </filled> </form> </vxml>

B. The VoiceXML form should be mixed initiative.

<?xml version="1.0"?> <!DOCTYPE vxml PUBLIC "-//BeVocal Inc//VoiceXML 2.0//EN" "http://cafe.bevocal.com/libraries/dtd/vxml2-0-bevocal.dtd"> <vxml version="2.0" xmlns = "http://www.w3.org/2001/vxml"> <!-- exercise 10.1.B. Mixed-initiative dialog --> <form id = "job_application_form"> <grammar type = "application/grammar+xml" src="job_form.grxml"/> <initial name = "start"> <prompt> What are you name and age? </prompt> </initial> <field name = "name"> <prompt> What is your Name? </prompt> <grammar type = "application/grammar+xml" src = "job_form.grxml#applicant_name"/> </field> <field name = "age"> <prompt> What is your age? </prompt> <grammar type = "application/grammar+xml" src = "job_form.grxml#applicant_age"/> </field> <filled> <prompt> Your name is <value expr = "name"/> Your age is <value expr = "age"/> Good Bye </prompt> </filled> </form> </vxml>

The following grammar is stored as job_form.grxml

<grammar type = "application/grammar+xml" root="job_form" xml:lang = "en"> <rule id = "job_form" scope = "public"> <item repeat = "1-"> <one-of> <item> <ruleref uri= "#applicant_name" /> </item> <item> <ruleref uri= "#applicant_age" /> </item> </one-of> </item> </rule> <rule id = "applicant_name" scope = "public"> <one-of> <item> sam <tag>name="sam"</tag> </item> <item> fred <tag>name="fred"</tag> </item> </one-of> </rule> <rule id = "applicant_age" scope = "public"> <one-of> <item> twenty<tag>age="twenty"</tag> </item> <item> thirty <tag>age="thrity"</tag> </item> </one-of> </rule> </grammar>

C. Perform a comparison test of (a) and (b) with novice callers. Which version do you recommend for novice callers?

Most novice users prefer (a) because it leads them to do the right thing. They don't have to guess what the right words are.

D. Perform a comparison test of (a) and (b) with experienced callers. Which version do you recommend for experienced callers?

Several experienced users (who understand mixed initiative dialogs) find (b) faster. However, they can always drop back into the application-directed mode when they need to.

10.2 Repeat exercise 8-2 for the pizza restaurant, this time developing a mixed initiative dialog.

<?xml version="1.0"?> <!DOCTYPE vxml PUBLIC "-//BeVocal Inc//VoiceXML 2.0//EN" "http://cafe.bevocal.com/libraries/dtd/vxml2-0-bevocal.dtd"> <vxml version="2.0" xmlns = "http://www.w3.org/2001/vxml"> <!-- exercise 10.2 mixed initiative --> <!-- get drinks (get pizza toppings not implemented; --> <!-- they are similar to get drink) --> <!-- get drink --> <script> <![CDATA[ var drink_size = new Array(); var drink_content = new Array(); // set listindex var drink_list_index = 0; ]]> </script> <form id = "get_drink"> <grammar type = "application/grammar+xml" src="drink.grxml"/> <initial name = "start"> <prompt> What would you like to drink? </prompt> </initial> <field name = "current_drink_content"> <prompt> Which drink? <break/> coke, pepsi, dietpepsi, lemonade, water? </prompt> <grammar type = "application/grammar+xml" src="drink.grxml.#drink_kind" /> </field> <field name = "current_drink_size"> <prompt> Drink size <break/> small, medium, or large? </prompt> <grammar type = "application/grammar+xml" src = "drink.grxml#drink_glass"/> </field> <filled> <script> <![CDATA[ drink_size[drink_list_index] = current_drink_size; drink_content[drink_list_index] = current_drink_content; drink_list_index = drink_list_index + 1; ]]> </script> <goto next="#another_drink"/> </filled> </form> <form id="another_drink"> <field name="more"> <prompt>do you want another drink? </prompt> <grammar type = "application/grammar+xml" root = "yes_no"> <rule id = "yes_no" scope = "public"> <one-of> <item> yes </item> <item> no </item> </one-of> </rule> </grammar> <filled> <if cond = "more == 'yes'"> <goto next="#get_drink"/> </if> <submit namelist= "drink_list_index drink_size drink_content"/> </filled> </field> </form> </vxml> With the following grammar, drink.grxml: <grammar type = "application/grammar+xml" root="get_drink" xml:lang = "en"> <rule id = "get_drink" scope = "public"> <item repeat = "1-"> <one-of> <item> <ruleref uri= "#drink_kind" /> </item> <item> <ruleref uri= "#drink_glass" /> </item> </one-of> </item> </rule> <rule id = "drink_glass" scope = "public"> <one-of> <item> small <tag>current_drink_size ="small"</tag> </item> <item> medium <tag>curreint_drink_size="medium"</tag> </item> <item> large <tag>current_drink_size="large"</tag> </item> </one-of> </rule> <rule id = "drink_kind" scope = "public"> <one-of> <item> coke<tag>current_drink_content="coke"</tag> </item> <item> pepsi<tag>current_drink_content="pepsi"</tag> </item> <item> dietpepsi<tag>current_drink_content="dietpepsi"</tag> </item> <item> lemonade<tag>current_drink_content="lemonade"</tag> </item> <item> water<tag>current_drink_content="water"</tag> </item> </one-of> </rule> </grammar>

A. Perform a comparison test of both versions with novice callers. Which version do you recommend for novice callers?

Most novice users prefer (a) because it leads them to do the right thing. They don't have to guess what the right words are.

B. Perform a comparison test of both versions with experienced callers. Which version do you recommend for experienced callers?

Several experienced users (who understand mixed initiative dialogs) find (b) faster. However, they can always drop back into the application-directed mode when they need to.

10-3 Discuss the similarities and differences between the navigation bar of a visual dialog and the universal commands of a verbal dialog.

A. What are the principle characteristics of the navigation bar of a visual dialog?

It always visible and always available so users can jump to other pages

B. What are the principle characteristics of the universal commands of a verbal dialog?

They are always available so callers can jump to other pages.

C. What are the significant similarities between (a) and (b)? Hint: while the options on the navigation bar are always visible, the universal commands of a verbal dialog can never be seen. How does the caller discover them and review them?

By using the command "universal" or similar command to hear the universal commands

10.4 Construct and demonstrate a mixed-initiative version of the dialog in Figure 5.4

See Figure 10.5

10.5 Construct and demonstrate a mixed-initiative version of a dialog in which the caller can order between one and five cans of either Pepsi or Coke.

<?xml version="1.0"?> <!DOCTYPE vxml PUBLIC "-//BeVocal Inc//VoiceXML 2.0//EN" "http://cafe.bevocal.com/libraries/dtd/vxml2-0-bevocal.dtd"> <vxml version="2.0" xmlns = "http://www.w3.org/2001/vxml"> <!-- exercise 10.5 mixed initiative --> <!-- get cans of pepsi or coke--> <!-- get drink --> <script> <![CDATA[ var drink_size = new Array(); var drink_content = new Array(); // set listindex var drink_list_index = 0; ]]> </script> <form id = "get_drink"> <grammar type = "application/grammar+xml" src="cans.grxml"/> <initial name = "start"> <prompt> How many cans of coke or pepsi do you want? </prompt> </initial> <field name = "current_drink_content"> <prompt> Coke or pepsi? </prompt> <grammar type = "application/grammar+xml" src="cans.grxml#drink_kind" /> </field> <field name = "current_drink_size"> <prompt> Number of cans? </prompt> <grammar type = "application/grammar+xml" src = "cans.grxml#number_cans"/> </field> <filled> <script> <![CDATA[ drink_size[drink_list_index] = current_drink_size; drink_content[drink_list_index] = current_drink_content; drink_list_index = drink_list_index + 1; ]]> </script> <goto next="#another_drink"/> </filled> </form> <form id="another_drink"> <field name="more"> <prompt>do you want more? </prompt> <grammar type = "application/grammar+xml" root = "yes_no"> <rule id = "yes_no" scope = "public"> <one-of> <item> yes </item> <item> no </item> </one-of> </rule> </grammar> <filled> <if cond = "more == 'yes'"> <goto next="#get_drink"/> </if> <submit namelist= "drink_list_index drink_size drink_content"/> </filled> </field> </form> </vxml> Grammar saved as cans.grxml: <grammar type = "application/grammar+xml" root="get_drink" xml:lang = "en"> <rule id = "get_drink" scope = "public"> <item repeat = "1-"> <one-of> <item> <ruleref uri= "#drink_kind" /> </item> <item> <ruleref uri= "#number_cans" /> </item> </one-of> </item> </rule> <rule id = "number_cans" scope = "public"> <one-of> <item> one<tag>current_drink_size ="1"</tag> </item> <item> two<tag>curreint_drink_size="2"</tag> </item> <item> three <tag>current_drink_size="3"</tag> </item> <item> four<tag>curreint_drink_size="4"</tag> </item> <item> five <tag>current_drink_size="5"</tag> </item> </one-of> </rule> <rule id = "drink_kind" scope = "public"> <one-of> <item> coke<tag>current_drink_content="coke"</tag> </item> <item> pepsi<tag>current_drink_content="pepsi"</tag> </item></one-of> </rule> </grammar>

a