Abstract

The field of Web-Based Publishing as presented in this course emphasizes design, implementation, and delivery issues of well-designed service for providing information using Web formats and protocols. In the introductory lecture, the range of topics covered in the course is presented. Additionally, the course project is presented, which spans the whole semester and is a joint project of all students taking the course. The project covers numerous areas, and students may choose the area in which they would like to specialize and acquire practical skills.

About Me

• Computer Science at Technical University of Berlin (TUB) (88-91)
  • working on DAPHNE, an SGML-based document preparation system
• Ph.D. at ETH Zürich (92-97)
  • thesis on Group and Session Management for Collaborative Applications
• Post-Doc at International Computer Science Institute (ICSI), Berkeley (97/98)
  • book on Technical Foundations of the World Wide Web
• Various activities back in Switzerland (98-06)
  • teaching at ETH Zürich and FHNW
  • research in various XML-related areas
  • starting and leading the ShaRef project
• Visiting Assistant Professor at UC Berkeley iSchool
  • previous courses (Fall 2006): XML Foundations and Web-Based Services
  • future courses (Fall 2007): XML Foundations and Web Architecture

About this Course

• Course Web page: http://dret.net/lectures/publishing-spring07/
• Course mailing list: subscribe at majordomo@sims.berkeley.edu (or at My.SIMS)
  • no subject (leave blank)
  • body of message: subscribe wbp
• Classes will be lectures and project discussions and work
• Attendance is mandatory

About these Slides

• Generated from XSLidy XML
  • all Slidy presentations are generated from this source
  • toc.html for displaying the summary on the course's Web page
• Designed for online presentation and use (lots of links!)
  • for printing, use a (all slides), and s (smaller font) a couple of times
• A good real-world example for XML applications
  • Slidy is useful, but there is no support for structures and hyperlinking
  • XSLidy adds these features by adding an XSLT transformation
  • XSLidy is useful, but there is no interface (XML editing only)
  • additional functionality can easily by incorporated (e.g., FormuLaTeX)

Additional Resources

• My Online Glossary at http://dret.net/glossary/
  • suggestions, updates, corrections are very welcome
  • another exercise in how to use XML and XSLT for information management
• My bibliography at http://dret.net/biblio/
  • suggestions, updates, corrections are very welcome
  • produced by an XML-centric system for managing bibliography data
• The World Wide Web Consortium (W3C)
  • the organization which invented XML and XLink
  • as well as many other technologies covered in this course

About You

The Big Picture

• Web-based publishing has a wide variety of facets
  • modeling content and designing a format implementing this model
  • providing support for creating and/or modifying contents
  • storing contents in a database for structured queries and retrieval
  • making contents available over the Web
  • making contents available in various representations (formats, languages, layouts)
  • doing all the above things with as little hardcoded limitations as possible
• Content creation, management, and dissemination are the focus of the course
  • much of these activities can be done using Web-Based Services
  • whenever possible, we stick to Web Architecture
• Goal of the course is an understanding of the publishing process
  • the advantages and limitations of technologies
  • the possible alternatives for certain implementation decisions
  • the relationships and dependencies of technologies
  • the practical issues arising when implementing a publishing scenario

Practical Skills

• Project work is a major part of the course
  • lectures provide the foundation and cover topics which are not part of the project
  • active participation is necessary and a core part of the course concept
  • all classes are mandatory, even if some of them might not be lectures
• All projects depend on each other
  • if there is a problem, please let me know
  • choose what you are interested in, and then stick to it
  • coordination (progress & problems) will be a useful lesson, too
• Design issues should always be discussed in class
  • we will have a lot of time for these discussions
  • without design coordination, the project cannot succeed

Project Issues

• Web-based publishing covers a wide variety of application areas
  • for Web-oriented content, hypermedia issues should be considered
  • regrettably, much of today's Web content is rather poor hypermedia
• XLink has AJAXish concepts (declarative instead of procedural)
  • when activating a link, the client changes the document as a result
  • XLink activation does not necessarily replace the whole document
  • declarative is always better than procedural (if it is possible for a given problem)
• AJAXLink is intended to live longer than 5 months
  • focus should be on well-documented code and clean design
  • first goal (until May) is to prove that XLink can be implemented using AJAX
  • further goals depend on the lessons from this semester and possible application areas

The XLink Disaster

• The XML Linking Language (XLink) was designed to add links to XML
  • XML elements have no semantics, so XML does not support linking by itself
  • XLink was intended to become the vocabulary people use when using links in XML
• XLink was never adopted by major tool or content providers
  • W3C's XLink implementation matrix is small and all implementations are incomplete
• The complete failure of XLink has several reasons
  • the standard heavily depends on namespaces, which is problematic
  • instead of defining a reusable model, the standard only defines a syntax (and a weakly defined implicit model)
  • user interface issues were not addressed at all, but would have been important for the Web
  • XLink does not sufficiently take care of the short lifespan of resources on the Web
  • the standard was not properly checked and backed by the organization creating it (W3C)

Web-Based Publishing

• XLink is unsuccessful but has many interesting features
  • links can be more complex than HTML's simple links
  • out-of-line links separate content from links
  • out-of-line links also allow the creation of third-party links
• XLink (like XML) was intended to become a client-side language
  • client-side languages always have the problem of legacy browsers
  • with AJAX becoming a de-facto standard, AJAXLink could change this
• XLink can also be useful as a server-side language
  • design, create, manage, and process content using XLink concepts
  • for publishing, turn everything into plain HTML (works for a subset of XLink)
  • advanced XLinks can be mapped to scripting (probably AJAX)
  • the mapping can be done server- or client-based (client-based needs JavaScript mapping)

Embedded Links

Linkbases

Third-Party Links

XLink Mockup

H. Weinreich, H. Obendorf, W. Lamersdorf, The Look of the Link — Concepts for the User Interface of Extended Hyperlinks, Proceedings of HT'01, August 2001

Client-Side XLink

• If XLinks use advanced features, client-side support is required
• Clients load XML documents and process XLinks
  • requires clients to be smart enough to understand XLinks
• If full XLink is supported, clients must also access external resources
  1. clients download complete linkbases
  2. clients are able to query XLink collections (instead of downloading them)
  3. clients can download resources for embedding them
  4. clients can interpret fragment identifiers of resources (for better embedding)
  5. fragment identifiers for different media types are be supported
• Client-side support must be built in stages
  • in addition to the XLink stages, add-on semantics can be included

Server-Side XLink

• XLinks on the server side are a format for storing links
• Before content delivery, content and links are merged
  • the basic process of querying a linkbase is required for the server-side as well
• Server-side processing can use HTTP Content Negotiation for smart decisions
  • generate different XLink representation based on the requesting browser
• The whole server-side process can be outsourced to a proxy
  • the proxy can be a regular or a reverse proxy

XLinkProxy

P. Ciancarini, F. Folli, D. Rossi, F. Vitali, XLinkProxy: External Linkbases with XLink, Proceedings of DocEng 2002, November 2002

AJAXLink Project

• XLink based on AJAX is an excellent publishing scenario
  • XLink is limited and underspecified, which also is a useful exercise
• Deliver XLinks through a REST service to requesting clients
• Request XLinks using AJAX code from Web pages using AJAXLink
• Project parts can be nicely identified and separated
  • a small number of parts are essential (XLink server, request code)
  • a bigger number of parts can be chosen based on personal preferences

AJAXLink Problems

• Where does the content come from?
  • small hand-generated test set or somebody specializing in harvesting XLinks
  • the biggest problem is the absence of a fragment identification culture on the Web
• Who is deciding what to do with XLink's fuzzy parts?
  • the author of Google's most popular XLink book
• What happens if somebody does not succeed?
  • close collaboration and frequent meetings avoid silent disasters
• How can we ensure that the code interoperates?
  • well-defined interfaces and frequent testing helps keeping the parts in sync
• How can we ensure that the code interoperates?
  • well-defined interfaces and frequent testing helps keeping the parts in sync
• But what … is it good for?
  • learning, publishing (papers), publishing (content), improving the Web, …

Options, Options, Options

• AJAXLink provides many possibilities for publishing-oriented projects
• Most projects can be implemented using (mainly) Web technologies
• AJAXLink is the framework
  • possible subjects range from UI design and usability to character transcoding
  • pick a subjects based on your interests and willingness to learn
• Decisions are due next Tuesday (1/30)
  • group sizes may vary between 1-3
  • in groups >1, responsibilities should be well-defined (e.g., interface design vs. implementation)
  • please pick a favorite and a second favorite
  • the goal is to cover as much as possible
  • preliminary discussions would useful (personally or on the mailing list – please subscribe!)
• If necessary, I will pick a part that nobody else is interested in …

Storage Management

• XLinks are XML structures and need to be stored and managed
  • XML document(s), XML database, or CMS?
• Storage solutions depend on the application requirements
  • the amount of data being stored
  • the access to the data (various access right associated with user roles)
  • the ability to quickly repurpose the data by writing new queries
  • HTTP Content Negotiation based on the request sent by the client
• AJAXLink needs a server storing the XLink data
  • the cheapest solution is a read-only server using XSLT or XQuery
  • write access would enable clients to modify and add XLinks through the service
  • pluggable queries would enable clients to request customized data

Link Semantics

• XLink supports typed links
  • XLink supports built-in semantics which can be extended
  • XLink applications can act on these extended semantics
• XLink implementations should implement the built-in semantics
  • actuate defines when to activate a link: onLoad, onRequest
  • show defines when to activate a link: new, replace, embed
• Additional link semantics can be implemented
  • actuate and show also can be specified as other
  • XLink has various other places for application-specific semantics
• Ultimate goal: full XLink implementation with useful extensions

Protocol Design

• XLink management as a Web-Based Service
• Access should be designed as a RESTful service
  • based on plain HTTP and easily usable by an AJAX implementation
  • service features can be implemented by using HTTP Content Negotiation
• AJAX is based on the idea of sending an HTTP request and receiving XML
  • for a generally usable service, this is a very useful model
  • AJAX always uses JavaScript on the client side, making JSON a better choice
  • if the client side is always the same, the result may even be more customized than JSON
• Depending on the effort, the protocol can be simple or powerful
  • minimal read-only implementation with XML (and JSON) result formats
  • well-designed REST service with CRUD methods and pluggable result formats

Interface Design

• XLinks support hypermedia contents beyond today's Web
• Interface design for XLink has to take into account XLink's model
  • links can be multi-ended (an anchor links to multiple destinations)
  • links can be typed (the destinations are of a certain class)
  • links can be retrieved from various sources (inline, out-of-line, third-party)
• Interface design could address additional interesting applications
  • XLink is open for extensions (defining alternative or supplemental semantics)
  • some of today's popular concepts may be expressible as XLink semantics
• Depending on the effort, the interface design can be simple or advanced
  • basic interface design based on simple JavaScript link code
  • configurable link design using advanced JavaScript, maybe in conjunction with pluggable result formats

Content Creation

• Mom, where do XLinks come from?
• Creating complex hypertexts is a complex tasks
  • the big and dirty secret of hypermedia has always been: it is too expensive
• Content creation in the form of an editor is not an option
• Various lightweight forms of content creation for this course:
  1. hand-crafted examples with a small number of documents and links
  2. use an existing body of hypertext such as the Web and XML glossary
  3. harvest XLinks from well-known markup structures on existing Web contents
  4. implement XLink creation as a simple annotation service (highlight text, push button)
• Depending on the scenario, content may require a customized Interface Design

Content Management

• How to manage and understand a large set of content & links
• In the simplest case, a small set of content and links will be used
  • for this limited scenario, no special management is required
• For more complex scenarios, content management becomes an issue
  • what are the dependencies between the exiting artifacts?
  • does changing an artifact affect any links using this as a resource?
  • for third-party linkbases the question of link rot is a very important one
• Data warehousing and data mining are applicable for linkbases
  • data warehousing is the process of accumulating all data relevant for a certain task
  • data mining is the task of performing complex queries on a large body of data
  • for large data warehouses, data mining is a good method for finding interesting correlations

Browser Integration

• Documents must specify the linkbases they are using
  • a reasonable and realistic assumption for documents and links from the same source
  • a severe limitation for scenarios involving third-party links
• Bookmarklets (JavaScript bookmarks) can be used to access a linkbase
  • access a linkbase when the user clicks on a bookmark
  • depending on the level of sophistication, linkbase access could be made configurable
• Questions of ownership and responsibility become important
  • users assume that a Web page is provided by one source (remember the frame problem?)
  • contents and links from different sources are hard to understand for most users
• Third-party links and browser-configurable linkbases are a challenge
  • concepts of responsibility and credibility become very blurred
  • while there are very interesting scenarios, there are also many questions
  • could content providers endorse link providers? disallow them? who controls this?

Technologies & Project

• Course goal is an overview of relevant technologies
• Features and limitations of the technologies will be discussed
• Implementation alternatives and technology dependencies
• AJAXLink covers the full spectrum of publishing issues
• Depending on what is covered, some parts will be left out
• Project goal is gaining practical experiences with a real-world project