Inleiding webprogrammeren

1
Inleiding webprogrammeren

• voor
• OGO 2.2
• door
• Reinier Post
• HG 7.71
• e-mail r.d.j.post_at_tue.nl (niet r.post_at_tue.nl)

2
Web technology overview

• What is the Web?
• What do we need the Web for?
• What is the Web made of?
• How does a user interact with the Web?
• What does a webserver do?
• What is a webserver made of?
• Some technologies for webserver programming
• Web-specific development issues

3
What is the Web?

• The World Wide Web a set of software techniques
  to support interlinked documents and applications
  on the Internet, standardized by the W3
  Consortium.
• It is client/server technology
• documents and applications reside on the server
• the user has a client program that connects to
  the server
• Clients include web browsers, robots, proxies.
• Servers include webservers, FTP servers, other
  servers.

4
What do we need the Web for?

• to publish collections of documents, e.g.
  documentation/support
• as a platform for user communities, e.g.
  discussion forums
• as an interface to databases and other
  applications
• ()

5
The Web as a universal user interface

• Webservers can be used to interface with
• local collections of files
• relational databases (by speaking SQL to them)
• mail servers, mailing lists
• other webservers (proxies, rewriting proxies)
• ()
• any application !3
• The webbrowser and webserver implement the user
  interface.

6
What is the Web made of?

• the World Wide Web a specific set of software
  techniques, including
• document addresses (URLs)
• document formats that support URLs (HTML, XML)
• the HTTP client/server communication protocol

7
How does a user interact with the Web? (1)

• The client program receives the order to retrieve
  a URL e.g. http//cgi.tue.nl/cgi-bin/WebPh?namep
  ost
• It determines from the URL
• the communication protocol to use (here HTTP)
• the server host to communicate with (here
  cgi.tue.nl)
• The client starts communication with that host,
  using the remainder as extra information. (Here
  it uses HTTP to ask cgi.tue.nl to return the
  document Web.ph?namepost .)

8
How does a user interact with the Web? (2)

• The server host responds
• it interprets the request from the client
• it prepares a document in response
• it sends the document back to the client, with
  some extra info
• The client finishes
• it receives and interprets the response
• it decides what to do with the received document,
  based on the extra info
• it does it (e.g. a browser usually displays the
  document)

9
How does a user interact with the Web? (3)

• This describes a stateless protocol (e.g. HTTP).
• There is a request with a response.
• By contrast, in a stateful protocol (e.g. FTP,
  SSH, telnet, ) we have
• client connects to server
• server accepts a session is established
• client and server exchange many messages
• the session can remain open indefinitely / it can
  be closed at some point

10
The HTTP protocol

• HTTP is a stateless client/server protocol for
  the World Wide Web
• The word webserver means HTTP server.
• 3 versions HTTP/0.9, 1.0, 1.1 (1.1 has
  extensions to support sessions)
• HTTP defines
• the format of a request
• the format of a response

11
A HTTP request

• a HTTP request consists of
• a path (e.g. /Web.ph?namepost)
• the method to use (usually, GET)
• input data (if the method is POST or PUT)
• headers, e.g.
• the name of the client software
• the types of documents the client is prepared to
  handle
• the Internet address of the clients computer
• extra information to link this request to a
  previous one (a cookie)

12
A HTTP response

• a HTTP response consists of
• headers, e.g.
• MIME document type (e.g. plain text, HTML , PNG
  image, Excel spreadsheet)
• document length
• last modification time
• reusable? (if the client should need this
  document again, can it reuse this copy?)
• a cookie
• content (the document itself)

13
What does a webserver do?

• receive a HTTP request
• decide how to handle it
• if it refers to a plain file on the file system
• find the file
• open it, read it
• if it refers to a built-in document generator
• call it with the right arguments
• read the output
• if it refers to an external program that
  generates documents
• find it, open it, call it with the right
  arguments
• read the output
• generate (extra) response headers
• pass result back to client as a HTTP response

14
What does a webserver do? (example)

• URL http//cgi.tue.nl/cgi-bin/WebPh?namepost
• webserver the HTTP server on cgi.tue.nl
• receive the request, parse it method GET,
  argument string /WebPh?namepost
• decide how to handle it
• it refers to an external program WebPh that
  generates documents
• call it with the argument string namepost, read
  the output
• generate (extra) response headers
• e.g. last modified now, reusable no, document
  type HTML
• pass output back to client as a HTTP response

15
What is a webserver made of?

• core functionality
• e.g. serving files from the filesystem
• simple built-in document generators
• e.g. error messages, status messages, directory
  indexes
• interface with external document generator
  programs
• CGI
• special modules for document generation
• ASP, PHP, Java servlets, etc.

16
Webserver software examples

• Apache
• very popular
• free, open source
• written for Unix/Linux, also available on MS
  Windows
• server-side programming support for Perl, PHP,
  Java, etc.
• IIS
• very popular, but limited to MS Windows platform
• closed source, Microsoft only (comes with MS
  Windows Server)
• server-side programming with ASP, ASP.NET (in
  VB, Perl, C, etc.)
• many others (see webserver surveys 1, 2)

17
CGI Common Gateway Interface

• CGI a standard way to call external programs
  from a webserver
• Such a program is usually called a CGI script,
  but it can be any executable program.
• The program receives the HTTP request
  information.
• It returns the document content and (optionally)
  HTTP response headers.

18
CGI examples

• CGI scripts can be written in any programming
  language.
• CGI script examples (use View Source to see the
  exact output)
• with source code (if MS IE doesnt display the
  source code as text, use View Source to see it)

19
Web programming which language to use?

• In general, instead of Perl, you can use any
  language that
• has more or less decent syntax and semantics
• provides a good range of libraries for the things
  you need to do
• installs and runs smoothly on the platform you
  choose to use
• provides good support forums and sites to answer
  your questions
• Examples VBA, Java, PHP, Python, C.
• Non-examples C, C, Bourne shell, LISP,
  assembler.

20
example webprogramming language Perl

• general-purpose scripting language
• popular on Unix, also runs on MS WIndows
• compare Unix shells, Tcl, Python, Ruby, MS VB,
  PHP, Java, C/C
• very flexible syntax (write-only)
• very many libraries (CPAN)
• details www.perl.org

21
Web programming in Perl interfacing with the
webserver

• possible ways to connect a Perl program with a
  webserver
• using CGI
• via Apaches mod_perl
• faster
• better integration with server
• via IISs ASP
• (same advantages)
• other (e.g. a webserver in Perl)
• (not popular these days)

22
Web programming in Perl manually

• Decode request into data structure
• Read input data, interpret arguments
• Do the processing (e.g. talk to a database
  server)
• Construct the output document content (usually
  HTML)
• Print output headers and content
• How?
• do it from scratch with string parsing and print
  statements? only a good idea for very simple
  programs
• find libraries that do all the dirty work for
  you!
• use output templates, components, or website
  management systems

23
Web programming in Perl libraries

• the CGI.pl CGI library (argument processing,
  output construction)
• the URI (URL) library (processing document
  addresses)
• HTML processing libraries (output construction)
• Apache interfacing libraries (to control the
  server from the program)
• LWP web client library (to talk to web servers
  from your program)
• DBI database processing (to talk with SQL
  database servers)
• Template systems, document component systems (for
  more advanced webserver programming)
• XML libraries (for standardized data exchange
  between programs)

24
Output template technology (ASP, PHP, JSP, SSI, )

• template output document with holes that
  contain executable code
• Examples of template languages
• ASP (executable code in VB, Perl, etc.), ASP.NET
• PHP (executable code in the PHP language)
• SSI (for Apache very primitive language for
  executable code
• Cold Fusion (special purpose webserver language)
• TemplateToolkit, HTMLMason (for Perl)
• JSP (for Java)

25
Server Side Includes

• SSI (http//httpd.apache.org/docs/mod/mod_include.
  html)
• PHP (http//www.php.net/)
• ASP (http//stardeveloper.com8080/articles/041600
  -1.shtml)
• JSP (http//java.sun.com/products/jsp/index.html)
• XSP (http//xml.apache.org/cocoon/xsp.html)
• ColdFusion (http//www.macromedia.com/software/col
  dfusion/)

26
example webprogramming language PHP

• Special web programming language.
• popular on Unix/Apache
• comparable to Perl/mod_perl, ASP/VB,
  JSP/servlets, etc.
• Perl-like language, but simpler (PHP vs. Perl)
• ASP-like, with special language (PHP vs. ASP)

27
PHP has built-in functions for

• SQL DB access
• XML
• Web client functions
• URLs treated as files
• Cookies
• File compression
• everything Perl has
• multi-language, translation (gettext)
• etc.

28
PHP examples

• my own examples (source code)
• IMP (mail client)
• phpMyAdmin (MySQL db administration) (see main
  site)
• phpGroupWare (content management system)
• PostNuke (content management system)
• SourceForge (example of home-grown website based
  on popular open source software also a great
  software resource)

29
PHP versus CGI/Perl

• PHP
• GPL, www.php.net
• IIS / Apache / as CGI
• interpreted / JIT is supported
• no components GTK for GUI
• HTML (or anything) embedded PHP
• lt?php tag ?gt

• CGI/Perl
• free, www.activestate.com
• runs everywhere
• CGI has huge overhead Perl is JIT compiled
• no components Tk / GTK for GUI
• no standard for embedding code
• no tags

30
ASP

• Microsoft product
• Extends HTML pages
• Tags lt and gt
• Uses VBScript, PerlScript (Perl) or Jscript
• Built-in objects Application, ASPError, Request,
  Response, Server, Session

31
JSP

• SUN standard
• For HTML or XML
• Tags lt, lt!, lt, lt_at_ page, etc.
• XML extensions
• predefined ltjspsetProperty .. /gt, etc.
• custom
• Uses JAVA
• Gives access to Servlet packages

32
ASP versus JSP

• ASP
• Microsoft license
• IIS or PWS
• Interpreted
• ActiveX components
• HTML
• Fixed tags

• JSP
• SUN open standard
• Apache, etc.
• Compiled (JIT)
• JavaBeans
• HTML, XML
• Extensible tags

33
XSP

• For XML documents
• Uses translation to byte code (JAVA)
• Tags xsppage, xsplogic, xspexpr,
  xspcontents, etc.
• Implementation Cocoon servlet (Apache)
• N-level processing
• Mixing of producers

34
JSP versus XSP

• JSP
• SUN
• HTML XML
• Generates servlets
• No mix allowed

• XSP
• Apache
• XML
• Generates producers
• Cooperates with XSL

35
component-based web programming

• Used in ready-made webprogramming systems.
• The website is programmed by extending existing
  components or plugins.
• A component occupies a position on the screen and
  implements specific functionality.
• Conceptually much superior to template-based
  programming.
• Many systems exist, not always very mature.

36
content management systems

• Ready-to-install software packages.
• A content management system allows the website
  content to be maintained without any programming.
• This is still an illusion, so any decent system
  is extensible, template/component based, and
  comes with source code.
• Many different systems exist no clear winners
  yet.

37
Web-specific development issues (1)

• Access control on server
• - separation of authors
• source of authentication info
• OS, general database, special database?
• - abuse / intrusion
• - use access permissions of OS
• - run programs in jail to restrict
  resources
• - manual checking / following

38
Web-specific development issues (2)

• Overflow of resources
• - limits on resources
• - automatic cleanup
• - manual checking / following

39
Web-specific development issues (3)

• Access control for client
• simple password or domain based user
  authentication
• truly secure transactions SSL

40
Web-specific development issues (4)

• Session management
• Via URL
• Via cookies

41
Web-specific development issues (5)

• How to test the software you write
• Run program outside webserver environment
• Use server facilities for debugging

42
Web-specific development issues (6)

• What to use for programming the user interface?
  Options
• plain HTML (which version?)
• HTML CSS
• XML XSL
• client-side scripting (Javascript, Java applets,
  Flash, plugins, ActiveX)

43
Web-specific development issues (8)

• Server-side programming
• pick a language with good web programming support
• how to interface with the server? (see earlier
  slides)
• low performance, simple interaction CGI
• more demanding sites integrated programming
  support (Apache modules, ASP, etc.)