beckonhissingInternet και Εφαρμογές Web

10 Νοε 2013 (πριν από 4 χρόνια και 8 μήνες)

119 εμφανίσεις

The Web was originally designed to deliver static Web pages from a Web server
connected somewhere on the Internet to a Web browser sitting on a user’s desktop
computer.Basically,all a user could do was click on a hot spot or hypertext link to
retrieve a new page,read it,and then go on to the next page.The Web was not
designed to support EC sites,especially B2C sites.In its original state,it was not
possible to create pages that would allow consumers to easily determine what prod-
ucts were for sale,to select products as they moved from page to page (i.e.,an elec-
tronic shopping cart),to place an order,or to verify an order.Similarly,there was no
simple way to integrate a Web server with a database system containing product,
pricing,and promotional data with transactional systems for processing orders and
with payment systems for handling credit card purchases and settlements.
Over time,these limitations have been addressed.First,forms were added to
HTML.Forms provided a way to produce Web pages from which a consumer could
select,order,and pay for products.Second,special programming and scripting lan-
guages (e.g.,Java and JavaScript) were created.These newer languages allowed
application developers to produce interactive Web pages whose functionality emu-
lated the rich functionality of standard Windows-based applications.Finally,a stan-
dard application programming interface (API),called the common gateway inter-
face (CGI),was introduced.Generally speaking,an API provides a way for one
software program to communicate with another,whereas CGI provides a way for
software developers and application programmers to integrate Web servers with var-
ious back-end programs and data sources.Because of CGI’s inefficiencies,newer
APIs and special database gateway programs were also introduced.As a result of
these changes,the Web is now well suited for the dynamic world of EC.
This appendix examines issues of end-user interactivity and dynamic data
access.The first sections focus on Java and JavaScript,which are special program-
ming languages that can be used to create Web pages with rich graphical user
interfaces (GUIs).The remaining sections examine various methods—CGI pro-
gramming,specialized APIs,and server-side scripting—for integrating a Web
server with back-end programs,including relational databases.
Most of us are familiar with various applications running on Microsoft Windows.
One of the hallmarks of a Windows application is its GUI (i.e.,its look and feel)
and the on-screen interactivity it provides.To regular Windows users,scrolling,
clicking,double-clicking,dragging and dropping,entering data,and similar
actions are almost second nature.In Windows,the desktop computer on which
the application resides handles these actions.For instance,if a user of Microsoft
Excel enters a formula and hits the “Enter” key,then the desktop computer calculates
the result,which is displayed immediately in the Excel spreadsheet.Even with a
application program-
ming interface (API)
A standard that provides
a way for one software
program to communi-
cate with another.
common gateway
interface (CGI)
A standard that provides
a way for software
developers and applica-
tion programmers to
integrate Web servers
with various back-end
programs and data
C-2 Appendix C
client/server application where processing is distributed between a desktop com-
puter and a back-end server,the desktop still handles the user’s actions.
Initially,the type of interactivity found in a Windows application was missing
from the Web.The earliest Web pages were built solely with HTML (described in
detail in Appendix B).Until forms were added to HTML,the only action sup-
ported by HTML was the clicking on a hypertext link to access a new page.Forms
added data entry capabilities and a series of new elements (e.g.,radio buttons,check
boxes,and dropdown selection lists).However,even with these new features,the
only thing a user could do was click.In the same vein,there was no way for the
browser (i.e.,the user’s desktop) to process the user’s actions.All the processing had
to occur on a Web server.Although this is enough interactivity for some applica-
tions,many applications are better served by the types of interactions supported by
Windows.This is where Java and scripting languages such as JavaScript come into
play.Java and JavaScript provide the means to create applications that support the
types of user interactions found in Windows.They also make it possible for pro-
cessing to be distributed between the desktop and the server.
Java’s roots are in work done at the beginning of the 1990s by a developer named
James Gosling at Sun Microsystems.This original work focused on developing
software and networks for consumer devices (such as VCRs,PDAs,toasters,and so
forth).The result of this work was a programming language called Oak.Recogniz-
ing the growing importance of the Internet,the developers at Sun Microsystems
turned their attention from consumer devices to making Oak a premier program-
ming language for the Web.In 1995,Oak was renamed Java.In that same year,a
number of software and hardware vendors licensed Java from Sun,including IBM,
Netscape,Microsoft,and Symantec.Today,it is estimated that there are over 1 mil-
lion programmers using Java to develop Web and network-based applications.
Java is similar to Cand can be used to produce stand-alone applications or
applets.An applet is a Java program that is written for and runs in a Web browser.
As a programming language,Java is well suited to Web development because of its
native support for Internet communications.
If Java was simply a programming language,then it would not have generated
much interest outside the programming community.After all,how many of us are
excited or know much about C,even though it is the development language
used in most commercial software products? What makes Java interesting to indus-
try observers is that it also has a run-time environment called a virtual machine
(VM).Any computer that has a Java VM can run a Java program.All Web browsers
have built-in Java VMs,although in the future users of the Microsoft browser
(Internet Explorer 6.0) will have to download a Java VM to their desktops.
Java differs from other programming language in one important respect.Other
programming languages require developers to write and then compile their pro-
grams on the machine on which the program is to be run.Compilation refers to the
process of converting the human readable form of a program into binary or
machine code that can be run by a specific computer.For example,if a developer
wants a program to run on a PC and a Unix computer,then the developer must
virtual machine (VM)
A run-time environment
in Java that enables any
computer to run a Java
program; all Web
browsers have built-in
Java VMs.
Appendix C C-3
write and compile the program on both machines.Even with the same type of
computer (e.g.,a PC),the program must be written for and compiled on each gen-
eration of the machine or the machine’s operating system (e.g.,Windows 98 versus
Windows 2000).Of course,you can see that standard program languages are ill
suited for the Web because of the large variety of computers that are attached to it.
In contrast,Java is a “write once,run anywhere” (WORA) programming language.
That is,a programmer writes a Java program and compiles it.Unlike other language
compilers that turn a program into machine-specific code,the Java compiler pro-
duces an intermediate form of code (called byte code) that is not specific to a particu-
lar machine.When Java byte code is sent to a Web browser,the Java VM executes
it on the fly.Where does the Java VM reside? It comes with the Web browser.
Special <APPLET> tags are used to incorporate applets within an HTML
page.Exhibit C.1 shows a Java-based spreadsheet and chart that might be used to
analyze sales data for online beverage,food,and nonconsumable purchases.This
particular page has two Java applets—a grid and a chart.Both of these applets
allow the end user to navigate through the data,selecting different slices of data,
and to drill down into underlying detail.Although the data might be displayed in
a static HTML table and image (a .jpg or .gif file),a static table or image could not
provide end users with the same interactive capabilities.The code behind part of
this page is shown in Exhibit C.2.The applets are defined by the <APPLET>...
Java Spreadsheet from Comshare Decision
Source:Comshare Decision, 2002.
C-4 Appendix C
</APPLET> tags.With each of these applets,the “CODE” parameter indicates
the Java code to be downloaded from the Web server.The WIDTH and
HEIGHT parameters of the <APPLET> tags define rectangular regions for dis-
playing the applets on the Web page.
When a Web browser accesses a page with an <APPLET> tag,it begins by render-
ing the page.During this process,it sees the <APPLET> tag and knows to (1) reserve
a display space within the page and (2) request the Web server to send the Java applet
code.Once the applet arrives,the browser’s Java VM begins executing the applet code.
The Java VM handles any interaction that takes place between the user and an
applet.It is as if there is a separate program running within the page.One advan-
tage of having the Java VM in control is that it can enforce security,providing a
“sandbox” that limits the applet’s access to system resources.For the most part,
applets do not have access to any system resources.For instance,applets cannot
read from or write to files on the end user’s computer.This is why it is difficult for
a hacker to write an applet that plays havoc with a user’s system.However,there
are drawbacks to this restriction.The fact that an applet cannot read or write to
disk means that an end user cannot store interim work or the results of any analy-
sis.This sort of functionality is crucial for those users (e.g.,mobile executives) who
do not have constant access to the Web server.
In spite of its power,few EC applications use Java applets.Most applications
still rely on Web forms and client-side scripting (see discussion later in appendix)
to provide end-user interactivity.However,Java is used quite extensively for
<APPLET CODE”CDGrid.class” CODEBASE”/sales/classes”
<PARAM nameViewName value”SalesData”>
<APPLET CODE”CDChart.class” CODEBASE”/sales/classes”
<PARAM nameViewName value”SalesData”>
Java Applet Tags for Spreadsheet
in Exhibit C.1
Appendix C C-5
server-side processing in the form of Java beans and Java servlets.Over the past
year,Microsoft has introduced a Java competitor called C# (i.e.,“C sharp”).Unlike
Java,C# runs only on computers with a Microsoft operating system (i.e.,Win-
dows XP,Windows 2000,etc.).
Java applets are not the only way to add interactivity to Web pages.Interactivity can
also be provided through client-side scripting using a language such as JavaScript.
JavaScript,often confused with Java,is not a lightweight version of Java.The two
languages have different capabilities and are used for different purposes.JavaScript
is used to control the objects,content,and interactions within a Web page.Unlike
Java,JavaScript cannot be used to produce stand-alone applications.
To program in JavaScript,you simply write the JavaScript programming state-
ments and functions directly in the Web page,interspersed with the HTML state-
ments.When a Web browser receives a Web page containing JavaScript,it auto-
matically knows to execute the program.Exhibit C.3 shows how JavaScript is
embedded among the HTML statements on a Web page.The page shown in
Exhibit C.4 is what results when the code is executed.
In Exhibit C.3,the JavaScript program is contained between the
<SCRIPT>...</SCRIPT> tags.Without going into the details,when the user
clicks the SUBMIT button,the function “onClick” is executed.The function
checks to see if the value of the “Name” input field in this particular form is empty.
A language, distinct
from Java, that is used
to control the objects,
content, and interactions
within a Web page;
unlike Java, it cannot
be used to produce
stand-alone applications.
<SCRIPT LANGUAGE  “javascript”>
FUNCTION valcheck(){
if (document.forms[0].elements[0].value  ““){
alert(“Please enter a name”);
<H2>Javascript Validation</H2>
<P>Name: <INPUT TYPE  “text” NAME  “T1” SIZE  “32”>
<P><INPUT TYPE  “button” NAME  “B1” VALUE  “Submit” onClick  “valcheck()”>
JavaScript Validating User Input
C-6 Appendix C
Java Form with Validated Input
If it is empty,then an alert window pops up the message,“Please enter a user
name.” Otherwise,the form is submitted.
When JavaScript was first introduced,one of its major purposes was to vali-
date user input (just like the example in Exhibit C.4).Today,it is used extensively
at EC sites to enhance the dynamic character of Web pages and to accomplish the
sorts of tasks listed in Exhibit C.5 (on p.C-7).
Like everything else on the Web,JavaScript has gone through a number of
changes.In the past couple of years,JavaScript has been intimately tied to the set
of elements that can be contained in a Web page.These elements include the text
and all the other components of a page,such as HTML tables,images,buttons,
forms,and so forth.The collection of possible Web page elements is called the
document object model (DOM).For each of these elements in the DOM,there
is a well-defined set of user actions that the element can recognize—in program-
ming vernacular,these actions are called events.Events include things such as
mouse actions,keyboard entry,and changes in the state of a document (e.g.,load-
ing a document is a state).The collection of events is called an event model.Taken
together,the DOM and event model determine how a scripting language such as
JavaScript can be used to manipulate a Web page.Today,the combination
of JavaScript and the DOM provide a rich environment for creating Web appli-
cations that closely mirror the types of interactivity provided by stand-alone
Windows applications.For a good introduction to JavaScript,see Harris (2001) or
Negrino and Smith (2001).
document object
model (DOM)
In JavaScript, the col-
lection of possible Web
page elements, for
each of which there is
a well-defined set of
user actions (events)
that the element can
Appendix C C-7
Functions of JavaScript
Function Example
Client-side form validation Ensuring that there are no blank input fields, that numbers
are numbers and text fields are text, that the length of a
field (like a ticker) is the correct length, and that numeric
input values fall within a given range
Client-side calculations As users input or select specific values, new calculations
can be computed and immediately displayed on the page.
Examples include mortgage calculations, currency
conversions, and calculating the price of a product based
on the features selected by a user.
Client-side lookup databases JavaScript provides associative data arrays that enable
values to be looked up on the basis of keys. For instance,
given a company name, an associate array could provide
the stock ticker (e.g., StockTicker(“Comshare”)”CSRE”).
Simple databases can be embedded in JavaScript,
providing lookup tables that can assist end users with
data input.
Client-side image maps Image maps are large graphics that are divided into areas
where each area provides a different response when
clicked by an end user. For instance, an image map might
represent the U.S. states. When the user clicks on a
particular state, a specified Web page is retrieved. In a
client-side image map, each of the areas can be associated
with particular JavaScript functions, making it possible
to carry out a variety of client-side actions besides
simply retrieving a page.
Client-side interaction The combination of JavaScript and the browser DOM
provides the means to create user interfaces that mirror
standard Microsoft Windows interfaces that respond to
events such as drag and drop, etc.
Personalizing a document If the JavaScript code is contained in the <HEAD> of a
before it is displayed document, the code is executed before the document is
displayed. In this way personalized documents can be
created based on the date, data contained in cookies, etc.
Manipulating Java applets JavaScript can be used along with various input objects
or ActiveX objects (e.g., buttons) to control applets or ActiveX objects
within a page contained on the same page. For instance, there might be
an applet displaying a graph and a series of buttons linked
to script that enable a user to select the type of graph,
background color, etc.
C-8 Appendix C
When Java and JavaScript first appeared,Microsoft saw them as a direct threat to its
control of the desktop marketplace.With these languages,software vendors could cre-
ate desktop and front-office applications (word processing,spreadsheets,and the like)
that could be stored on any Web server (not just Microsoft’s Internet Information
Server,IIS),downloaded to any Web browser (not just Microsoft’s Internet Explorer),
and run on any computer (not just a PC with Microsoft Windows or NT).Clearly,
widespread adoption of Java and JavaScript applications could have reduced Microsoft’s
market share (especially on the desktop),although this has not come to pass.One of the
things that Microsoft did in response to this potential threat was to introduce its own
proprietary components and scripting language called ActiveX components and
VBScript (Visual Basic Script),respectively.Although both are designed for the Web,
they work only with Microsoft’s Internet Explorer and Microsoft’s Web server.
Like Java applets,ActiveX components are stored on a Web server.When a
user of Internet Explorer downloads a Web page referencing an ActiveX compo-
nent,the browser retrieves the component and executes it.Unlike Java,ActiveX
does not run in a “sandbox.” ActiveX components can access system resources
(e.g.,reading from and writing to files on the desktop).
All of Microsoft’s office applications—Word,Excel,Access,PowerPoint,and
others—come with a scripting language called Visual Basic for Applications
(VBA),which can be used to write everything from simple macros for automating
various manual processes to full-blown custom applications built on top of one of
the desktop applications.You can think of VBScript as VBA for Internet Explorer.
VBScript works the same way as JavaScript and has the same capabilities.This
means that VBScript programs,intertwined with the HTML on a Web page,can
control the elements of the DOM and their associated events.In fact,both
VBScript and JavaScript can be included on the same Web page.Of course,it
requires Internet Explorer to run the page.If the page were accessed with
Netscape’s Navigator,the VBScript would be ignored and the JavaScript would
run,even though the page would not make much sense.
Most EC sites on the Web provide some level of personalization,interactivity,and
data access.In this case,static pages that are constructed ahead of time and stored
on a server just will not work.With a static page every user sees the same content.
Imagine,for instance,a Web site such as Yahoo! that provides stock quotes that are
updated every 15 minutes.There is no practical way for a site like this to “grab” the
price/volume data for every stock and then generate new static pages (including
tables and charts) every 15 minutes.Obviously,other external programs running
on the Web server (or some other server) are used to generate “dynamic” Web
pages on the fly in response to individual user requests.In this case,when an indi-
vidual user requests an update on a particular stock (say,Microsoft,whose ticker is
MSFT),the external program receives the request,grabs the new stock quote (in
real time),and then generates a Web page with the updated information.How this
is actually accomplished is detailed in the following sections.
Appendix C C-9
Originally,Web pages with dynamic content were generated with the aid of a CGI
program.How CGI works is diagrammed in Exhibit C.6.From a technical stand-
point,CGI is a set of standard methods and routines used to write a stand-alone
software program that knows how to receive requests from a Web server and
return data to the server.In other words,CGI is a standard protocol used to con-
verse with a Web server.A program that has this capability is called a CGI pro-
gram.These programs can be written in a wide variety of programming languages.
On Unix,the Perl programming language is often used.On Microsoft Windows,
Visual Basic or C/Care frequently used.
To illustrate the basics of CGI programming,consider the simple example
shown in Exhibit C.7.In this example,the user can obtain the current stock price
(within the last 15 minutes) for either Microsoft or Oracle by clicking on the com-
pany name.Behind each name is a URL (see the code in Exhibit C.8).In this case,
the URL refers to a program rather than a static document.For instance,the URL
associated with the Microsoft selection is “www.stocknews.com/cgi-bin/ find-
stockprice.exe?tickerMSFT.” Based on this URL,the Web server that receives
the request is “www.stocknews.com.”
Upon receipt,the server knows that it needs to invoke a CGI program called
“stock.exe” rather than retrieve a particular Web document.When it invokes the
program,it passes the program the information after the “?” mark.The infor-
mation after the “?” mark is called a query string.The query string specifies vari-
ous parameters and values that the CGI program will use during execution.In
this case,the query string specifies the stock ticker whose current price is sought.
Once the program determines the associated price,it will return the results to
the Web server in the form of an HTML page.In turn,the server will return the
page to the user’s browser in exactly the same fashion that it would return any
other page.
In most real-world applications,HTML forms are used instead of URLs with
query strings.The details of how this is done are well beyond the scope of this
appendix.Readers who are interested in the details of CGI programming should
see Hamilton (2000).
Static HTML Pages
How CGI Works
query string
The information that
specifies various para-
meters and values that a
CGI program will use
during execution.
C-10 Appendix C
<H2>Retrieve Current Stock Price</H2>
<A HREF  www.stocknews.com/cgi-bin/stock.exe?tickermsft>Microsoft</A>
<A HREF  www.stocknews.com/cgi-bin/stock.exe?tickerorcl>Oracle</A>
HTML Code for Accessing
Stock Price
Accessing Stock Prices
Through URLs
Appendix C C-11
Although CGI programs were widely used,they suffered from some significant
technical limitations,especially at Web sites with a lot of traffic.CGI programs are
either separate executable programs or interpreted scripts.When a Web server
invokes a CGI program,an individual copy of the program is brought into mem-
ory.After the program has run,the executable program is unloaded from memory.
If the server receives another request while the first program is still running,then
it will invoke a second instance.The number of copies in memory will depend on
the number of requests the server is trying to service.At a large site,the number of
copies could run in the thousands.Obviously,the frequent loading and unloading
of a large number of copies is very inefficient,not to mention the fact that the per-
formance of the server will degrade substantially as the number of requests
Other gateway architectures and APIs have been developed to overcome the
limitations of CGI programming.Two alternatives are FastCGI and Microsoft’s
ISAPI (Internet Server API).The difference between these APIs and CGI is that
the server invokes only a single instance of the program,and a technique known as
multithreading is used to handle individual requests (each request is now a thread
instead of the entire program).The overall result is increased efficiency and sub-
stantially improved performance.
With the exception of the Perl programming language,it takes a skilled program-
mer to write a Web-based application using one of the Web gateway APIs.Fortu-
nately,simpler scripting languages can be used to accomplish many of the same
tasks.These scripts are run on the server side,not the client side.Included among
these specialized server-side scripting languages are Microsoft’s Active Server
Pages (ASP),Apache’s Hypertext Preprocessor (PHP),Java Server Pages ( JSP),
and ColdFusion’s Markup Language (CFML).All of these languages are used
widely throughout the Web.For example,in April 2002,the Netcraft monthly
Web server survey (netcraft.com/survey) estimated that approximately 48 percent
of the 37 million sites it surveyed were using either ASP or PHP scripting (Berkes
2002).Buser (2000) provides a good introduction to ASP scripting,and White-
head and Desamero (2001) offers a good introduction to PHP scripting.
All of these server-side scripting languages work in the same manner.Basically,
a server-side script is nothing more than a Web page that contains script statements
interspersed with various HTML tags.The script statements are denoted by spe-
cialized tags.Exhibit C.9 displays a very simple ASP page.The server-side script
statements are denoted by the <% ...%> tags.The first script statement
(“request.form”) captures the username entered on a form.The second script state-
ment (“response.write”) simply writes the user’s name on the Web page.The rest of
the statements in this script are nothing more than standard HTML tags.
When this page is accessed,the Web server first hands it off to the appropri-
ate server-side script processor.How does it know which processor to use? The
appropriate processor is selected by the file extension of the page being accessed.
For instance,an ASP page has an “.asp” file extension,whereas a PHP page has a
C-12 Appendix C
“.php” extension.The selected processor then executes the script code within the
page and sends the resulting page back to the browser via the Web server.When
the browser receives the page,it renders the rest of the HTML code.In this way
the content of the page is dynamically generated in much the same way that the
special API programs generate dynamic content.The resulting page is shown in
Exhibit C.10 (assuming that the user entered the name John Smith).
ASP Output
<% username  Request.Form(“username”) %>
<% Response.Write username %>
Thank you for joining us. Would you like to see
your orders?
<INPUT TYPE  “submit” VALUE  “Yes”>
ASP Code
Appendix C C-13
Berkes,D.“Survey Says:PHP Passes Microsoft Active Server
Pages.” newsforge.com/newsforge/02/06/11/011243.
shtml?tid5 ( June 11,2002).
Buser,D.,et al.Beginning Active Server Pages 3.0.Birmingham,
UK:Wrox Press,2000.
Buyens,J.Web Database Development:Step By Step.Redmond,
WA:Microsoft Press,2000.
Hamilton,J.“CGI Programming 101.” cgi101.com(2000).
Harris,A.JavaScript for the Absolute Beginner.Indianapolis,IN:
Premier Press,2001.
Negrino,T.,and D.Smith.JavaScript for the World Wide Web,
4th Ed.Berkley,CA:Peachpit Press,2001.
Whitehead,P.,and J.Desamero.PHP:Your Visual Blueprint for
Creating Open Source,Server-Side Content.New York:
Hungry Minds,Inc.,2001.
application programming
interface (API),p.C-1
common gateway
interface (CGI),p.C-1
document object
model (DOM),p.C-6
query string,p.C-10
virtual machine
Relational databases are an integral part of virtually every EC site.Most of the
pages that appear on the Web are dynamically generated from database content.
This is one of the major roles of server-side scripting.Scripts are written to process
the input from an HTML form,to use the input to construct a relational database
query behind the scenes (i.e.,to produce an SQL statement that the relational
database can execute),and then to dynamically produce and return an HTML
page from the results of the query.
For example,you may have used the FedEx or UPS Web sites to track a ship-
ment.The status of any shipment for a given customer is contained in a database.
To determine the status of a package,the customer enters the shipment number in
a Web form.When the customer clicks the submit button,the shipment number
is passed to a server-side script.The script takes the shipment number and con-
structs a relational query (e.g.,SELECT shipment_no,shipment_status FROM
shipments) that is passed to the shipment database.The database produces a result
containing the shipment_status,which is passed back to the script.The script
takes the results and generates an HTML page on the fly,which is then returned
to the customer.
Again,the specific techniques that are used to dynamically access the informa-
tion in a relational database are beyond this discussion.Readers who are interested
in this topic should refer to Buyens (2000).