--------------------------------------------------------------------------
Software Development    		Fall 2005
CSU 670
--------------------------------------------------------------------------
Assignment 3       		Due date: Sep. 30, 2005

--------------------------------------------------------------------------

This assignment is stored in file $SD/hw/3

in file assign.txt
--------------------------------------------------------------------------

=========
We reuse the abbreviations already introduced in hw 1:
=========

THEMES: Evolution of Java programs,
          change design, translate to Java

READING:
Read section 12 of TPP: Domain Languages.
   What are the minilanguages that we are using?
   What is our favorite way to define a minilanguage?
Read section 14 of TPP: The power of plain text.
   What kind of knowledge do we keep in plain text?
Read chapter 3: The Basic Tools.
   Are you using one editor that works well on Unix and Windows?
  
This homework requires use of DemeterJ (only a few features) and DJ.

Browse the resource guide at:
 $D/DemeterJava/


Installing:
For this homework, you need DemeterJ. Follow the instructions in:

http://www.ccs.neu.edu/research/demeter/software/docs/install.html

Please note that homework submission is electronic
submission using Blackboard and hardcopy. This will allow the teaching assistant
to run your programs.
In ADDITION, please turn in your answers in hardcopy.
See instructions in previous assignments.


Question 1:
-----------

(This question jointly developed with Jingsong Feng.)

PART 1.A
--------

Motivation: Translation between languages / XML

The *.beh files contain Java methods between {{ and }}
that are assigned to various classes. Each *.beh file is of the form:

X{
  {{
     // method 1
     // method 2
     // etc.
  }}
}

Y{
  ...
}
...

and therefore the .beh files are basically 99.99999 % Java code.
When you write your own .beh files,
use your editor to make sure that the parentheses are balanced inside
{{ and }}.



In the lecture we translated bus route descriptions to Scheme make-* expressions.
See
/home/lieber/.www/courses/csu670/f05/examples/BusRoute

In this homework we translate bus route descriptions to XML documents.

Use the following class dictionary as the definition for the input language.

import edu.neu.ccs.demeter.dj.*;
import java.io.*;

BusRoute = <name> Ident "buses" <buses> List(Bus) "towns" <towns> List(Town).
Bus = <name> Ident <passengers> List(Person).
Town = <name> Ident "busStops" <busStops> List(BusStop).
Person = <name> Ident.
BusStop = <name> Ident <waiting> List(Person).
List(S) ~ "(" {S} ")".
Main = <s> String.

Bus route descriptions in the language are short and your task is to translate them
into longer XML docmuments.

In XML the bus routes have the following format:
(file: busroute.xml)

<?xml version="1.0"?>
<!DOCTYPE busRoute SYSTEM "busRoute.dtd">
<busRoute name="Route1">
   <bus name="Bus45">
     <person name="Dimitrios"/>
     <person name="Viera"/>
   </bus>
   <town name="Boston">
      <busStop name="MassAve">
		<person name="Bryan"/>
		<person name="Mike"/>
      </busStop>
   </town>
   <town name="Cambridge">
      <busStop name="CentralSquare">
	<person name="Alec"/>
	<person name="Matthias"/>
      </busStop>
   </town>
 </busRoute>

The files mentioned in this hw are in the hw directory (see /xml).
See also $SD/examples/sameInXML*.

XML has something similar to class dictionaries. They are called 
DYDs and the above XML document satisfies the following DTD:
(file: busRoute.dtd)

<?xml version="1.0"?>
<!ELEMENT busRoute (bus+,town+)>
<!ELEMENT bus (person+)>
<!ELEMENT town (busStop+)>
<!ELEMENT busStop (person+)>
<!ELEMENT person EMPTY>
<!ATTLIST busRoute name CDATA #REQUIRED>
<!ATTLIST bus name CDATA #REQUIRED>
<!ATTLIST town name CDATA #REQUIRED>
<!ATTLIST busStop name CDATA #REQUIRED>
<!ATTLIST person name CDATA #REQUIRED>

This is basically the same information as in our class dictionary with
the exception that the lists must be non-empty in the XML document.

In the following you must use Microsoft Internet Explorer version 5 or later.

The following HTML file, when loaded into Explorer will test whether your XML document
satisfies the DTD.
(file: checker.html)

<html>
<HEAD>
<TITLE> A checker for verifing XML file against its DTD</TITLE>
</HEAD>

<body>
<br>
<br>
<h2>This is a checker for verifing XML file against its DTD.</h2>
<br>

<b>The following is the result for verification:</b>

<br>
<script language="JavaScript">
	var filename="busroute.xml"
	var xmlDoc = new ActiveXObject("Microsoft.XMLDOM")
	xmlDoc.async="false"
	xmlDoc.validateOnParse="true"
	xmlDoc.load(filename)

	if(xmlDoc.parseError.errorCode!=0){
		document.write("<br>This xml file "+filename+" is invalid.")
		document.write("<br>Error Code: ")
		document.write(xmlDoc.parseError.errorCode)
		document.write("<br>Error Reason: ")
		document.write(xmlDoc.parseError.reason)
		document.write("<br>Error Line: ")
		document.write(xmlDoc.parseError.line)
	}
	else{
		document.write("<br>The xml file "+filename+" is valid.")
	}
</script>

</body>
</html>

The following HTML file, when loaded into IE, will process your XML document and compute
the names of the people waiting at the bus stops.
(file: xpath.html)

<html>

<body>

<h2>This page is to demonstrate use XPath to query data in xml file:</h2> <br>

<script language="javascript">
	var xmlDoc=new ActiveXObject("Microsoft.XMLDOM")
	xmlDoc.async="false"
	xmlDoc.load("busroute.xml")

	var nodes=xmlDoc.selectNodes(".//busStop//person")
	var length=nodes.length
	if(length==0){
		document.write("There is no person waiting in the busstops in this busroute")
	}
	else{
		document.write("There are "+length +" persons waiting in the busStops, their names are:<br><br><ui>")

		for (var i=0; i < length; i++) {
		  document.write("<li>"+nodes(i).attributes.item(0).text+"</li>")
		}
		document.write("</ui>")
	}
</script>

</body>
</html>

Use a visitor to do the translation.
Use the visitor with the traversal
that visits the entire BusRoute-object.

Turn in the modified directory  
that contains your new visitor.

The directories must contain your test inputs and a README file that contains
appropriate documentation how to use your directory.

Remember:

Before you zip up a directory generated by demeterj, run "demeterj clean".



=================== Some technicalities
DemeterJ has a project file facility
which works pretty much the same on any platform (the file paths need to be
adjusted from platform to platform). Typically, the project file is
called

program.prj

A default program.prj file is generated by "demeterj new". 

If you use DemeterJ on a PC with Windows and you copy my
program.prj, make sure you change the / to \ in the file paths.
Also, you need to replace javacc by javacc.bat.

PART 1.B
--------

Translate from XPath to traversal strategies. Assume that the context
node is of type Start.

.//B//C
.//Book | .//Cd

Part 1.C
--------

Translate from traversal strategies to XPath.

from A to B
from A to {B,C}




Question 2:
-----------

The following files and directories refer to $SD/hw/3

Motivation: Learn about how an adaptive program defines a family of
object-oriented programs. Learn how to modify one adaptive
program instead of two object-oriented programs.

Consider the Java programs *.beh in

onlyDJ-average-default-list
onlyDJ-average-other-list

They are identical and you can convince yourself by running check-diff which is provided 
in the hw 3 directory.

Basically, I want changes to the *.beh files that
are the same to both directories that would give the "correct" result in both
directories. The differences between the 2 directories is the class
dictionary. One is much simpler than the other. 

The Java program (in *.beh)
is supposed to print out the running average as Weight-objects
are encountered during the traversal.
The last number printed by the program should be the average weight.
Unfortunately, the program is broken and you are asked to repair
it with minimal changes to the files.

Your task is to correct one or more of the *.beh files
so that they work correctly in BOTH

onlyDJ-average-default-list
onlyDJ-average-other-list


demeterj test

allows you to test the corrected *.beh files after modification.
You should find the MINIMAL modification necessary to make the
program correct.

Turn in for each behavior file that you changed the
corrections that you made.
Clearly indicate what you corrected in each *.beh file.

Question 3:


Write a class dictionary for Chinese food.
Your class dictionary should accept the sentence shown below:

Food Menu:
  Meal 1:
    <Dishes:> gongbao chicken 
    <Soup:> egg soup 
    <Dessert:> bun, dumpling
  Meal 2:
    <Dishes:> five flavored beef, tiger skin egg 
    <Soup:> three delight soup 
    <Dessert:> noodle
  Meal 3:
    <Dishes:> gongbao chicken, five flavored beef, tiger skin egg
    <Soup:> egg soup
    <Dessert:> bun, dumpling, noodle
  Meal 4:
    <Dishes:> tiger skin egg, gongbao chicken, five flavored beef 
    <Soup:> egg soup

Turn in a class dictionary that defines a menu language
implied by the above example.

It should start with:

ChineseFoodMenu = "Food" "Menu:" List(ChineseMeal).
...

Use DemeterJ to test your class dictionary for syntactic and semantic
correctness:

1. Install DemeterJ
2. demeterj new
3. put your class dictionary into program.cd
4. put above example into program.input
5. In method Main.main in program.beh, remove "Main m = parse(System.in);"
   then add "ChineseFoodMenu m = ChineseFoodMenu.parse(System.in);"
6. demeterj test
     should not give syntax or semantic errors on the class dictionary
     or compilation errors from the Java compiler.

======================
System help:

When you think DemeterJ behaves unexpectedly, use

demeterj clean

which will remove all the generated code.