Search

Sgvizler is a small JavaScript wrapper for visualization of SPARQL results sets. It integrates well with HTML web pages by letting the user specify SPARQL SELECT queries directly into designated HTML elements, which are rendered to contain the specified visualization type on page load or on function call. Sgvizler supports a vast number of visualization types, most notably all of the major charts available in the Google Chart Tools, but also by allowing users to easily modify and extend the set of rendering functions, e.g., specified using direct DOM manipulation or external JavaScript visualization tool-kits. Sgvizler is compatible with all modern web browsers.

Due to various Web authoring tools, the new web standards, and improved web accessibility, a wide variety of Web contents are being produced very quickly. In such an environment, in order to provide appropriate Web services to users’ needs it is important to quickly and accurately extract relevant information from Web documents and remove irrelevant contents such as advertisements. In this paper, we propose a method that extracts main content accurately from HTML Web documents. In the method, a decision tree is built and used to classify each block of text whether it is a part of the main content. For classification we use contextual features around text blocks including word density, link density, HTML tag distribution, and distances between text blocks. We experimented with our method using a published data set and a data set that we collected. The experiment results show that our method performs 19% better in F-measure compared to the existing best performing method.

Today’s data needs flexible visualization, and the advent of new tech-nologies defined under HTML5 has opened up new possibilities for rich web-applications. We used JavaScript and the new HTML5 ‘canvas’ element, to create a software module GraphVis for online visualization and exploration of large relationship networks. As a first test-case we applied the module to Princeton’s WordNet English thesaurus, and made the result available as the web-application WordVis. WordVis is an exceptionally agile, online synonym-dictionary explorer. It shows a customizable graph of interlinked words and meanings, which are arranged by real-time animation, and behave like repelling magnets that are also rearrangeable by the user. WordVis shows the merits of using JavaScript: the visualizer can be fully integrated inside a web page, and that enables communication with other HTML elements in the page. It demonstrates the latest developments in web technology, and its agility may serve as an inspiration for other web-applications. As an intuitive, online synonym-dictionary explorer, it helps to ‘find the right word’ while writing English texts. It offers an online, free and more powerful browsing experience than any existing word-network visualizer.

We report on the embedding of a domain specific language, (X)HTML, into Haskell and demonstrate how this web language can be represented and rendered for strong World Wide Web Consortium (W3C) compliance. Compliance of web content is important for the health of the Internet, accessibility, visibility, and reliable search. While tools exist to verify web content is compliant according to the W3C, few systems guarantee that all dynamically produced content is compliant. We present CH-(X)HTML, a library for generating compliant (X)HTML content for all dynamic content by using Haskell to encode the non-trivial syntax of (X)HTML set forth by the W3C. Any compliant document can be represented with this library, while a compilation or run-time error will occur if non-compliant markup is attempted. To demonstrate our library we present examples, performance measurements, and a discussion of library version considerations.

In this paper, a new flexible range markup technique for Web documents is presented as an alternative to XPointer technology, based on the RDFa standard. The principal objective is to define semantic annotations that support the evolution of annotated Web documents more effectively than XPointer. The term ‘Flexible Range' indicates that annotations can be defined over different ranges of text within a Web page and multimedia objects within it, independently of its HTML tags.

The paper presents a new metric and a framework to assess the effort of repairing the accessibility of a Web site. For that all the HTML elements of all the pages of a site are considered, excluding those that are duplicated. The rationale is that those elements are originated in a reusable construct, such as a template and, therefore, need to be corrected only once. The evaluation then ap-plies the accessibility evaluation techniques on those elements instead of on all the instances that are presented to the user. The reported fails and warnings are then computed in a simple sum metric. The paper also describes the validation experiment of both metric and framework, providing very important results. These may well contribute to a different perspective from managers and development team leaders about the effort to revamp the accessibility of a site.

In many application domains in science and engineering, data produced by sensors, instruments, and networks is naturally processed by software applications structured as a pipeline. Pipelines comprise a sequence of software components that progressively process discrete units of data to produce a desired outcome. For example, in a Web crawler that is extracting semantics from text on Web sites, the first stage in the pipeline might be to remove all HTML tags to leave only the raw text of the document. The second step may parse the raw text to break it down into its constituent grammatical parts, such as nouns, verbs, and so on. Subsequent steps may look for names of people or places, interesting events or times so documents can be sequenced on a time line.

The amount of information contained in databases in the Web has grown explosively in the last years. This information, known as the Deep Web, is dynamically obtained from specific queries to these databases through Web Query Interfaces (WQIs). The problem of finding and accessing databases in the Web is a great challenge due to the Web sites are very dynamic and the information existing is heterogeneous. Therefore, it is necessary to create efficient mechanisms to access, extract and integrate information contained in databases in the Web. Since WQIs are the only means to access databases in the Web, the automatic identification of WQIs plays an important role facilitating traditional search engines to increase the coverage and access interesting information not available on the indexable Web. In this paper we present a strategy for automatic identification of WQIs using supervised learning and making an adequate selection and extraction of HTML elements in the WQIs to form the training set. We present two experimental tests over a corpora of HTML forms considering positive and negative examples. Our proposed strategy achieves better accuracy than previous works reported in the literature.

We present a system that enables secure user authentication by leveraging a portable USB-based trusted device. The heart of our system runs a protocol which guarantees trusted behavior at multiple layers; from the hardware device itself, to the software executing on the hardware, and finally to the application hosted in the remote server. This combination assures end-to-end trust and makes our system resilient to physical attacks (e.g. to the device and wire tapping) as well as logical attacks (e.g. main-in-the-middle attack). Our system utilizes web-based proxy communication using standard HTML tags and JavaScript to coordinate communication amongst different components. This enables our system not having to install any extra drivers typically required for supporting communication in most existing technologies.

Information extraction (IE) has been emerged as a noveldiscipline in computer science. In IE, intelligent algorithms areemployed to extract the required data, and structure them so thatthey are appropriate for query. In most IE systems, a web-pagestructure, e.g. HTML tags are used to recognize the looked-forinformation. In this article, an algorithm is developed torecognize the main region of web-pages containing the looked-forinformation, by means of an ontology, a web-page structure andgoodness-of-fit Χ2 test. After recognizingthe main region, the existing records of the region are recognized,and then each record is put in a text file.

With more and more hypertext documents available online, hypertext classification has become one popular research topic in information retrieval. Hyperlinks, HTML tags and category labels distributed over linked documents provide rich classification information. Integrating these information and content tfidfresult as document feature vector, this paper proposes a new weighted hyper-sphere support vector machine for hypertext classification. Based on eliminating the influence of the uneven class sizes with weight factors, the new method solves multi-class classification with less computational complexity than binary support vector machines. Experiments on benchmark data set verify the efficiency and feasibility of our method.

In this paper, to improve the security, sensitivity and efficiency, we propose two novel fragile watermarking schemes for tamperproof web pages, in which a message authentication method and a sparse coding method are used to generate the watermarks for a web page, respectively. The former ensures high security and good sensitivity of the watermarking scheme. The latter attaches more importance to satisfy all the requirements of security, sensitivity and efficiency well. The generated watermarks are then embedded into the web page through the upper and lower cases of letters in HTML tags. The properties of the schemes are illustrated with several experiments, and the results show that the proposed schemes can be good practical tools for the tamperproof of web pages.

The PHP string analyzer developed by the second author approximates the string output of a program with a context-free grammar. By developing a procedure to decide inclusion between context-free and regular hedge languages, Minamide and Tozawa applied this analysis to checking the validity of dynamically generated XHTML documents. In this paper, we consider the problem of checking the validity of dynamically generated HTML documents instead of XHTML documents. HTML is not specified by an XML schema language, but by an SGML DTD, and we can omit several kinds of tags in HTML documents. We formalize a subclass of SGML DTDs and develop a translation into regular hedge grammars. Thus we can validate dynamically generated HTML documents. We have implemented this translation and incorporated it in the PHP string analyzer. The experimental results show that the validation through this translation works well in practice.

A dedicated browser application named Voice Browser for Groupware systems (VoBG) was developed. The target users are visually impaired persons who are computer novices. This browser is a client application for a specified groupware system named Garoon 2, which is one of the most popular groupware system packages in Japan. The groupware is Web-based system and basically accessible for the visually impaired users. However, the structure of each Web page generated by the groupware server is too complicated to get information exactly for beginners with visual impairment. To solve this problem, VoBG reduces information of these pages using information of HTML tags, and reconstructs each page to understandable format for the target users.

Most text categorization methods are based on the vector-space model of information retrieval. One of the important advantages of this representation model is that it can be used by both instance-based and model-based classifiers for categorization. However, this popular method of document representation does not capture important structural information, such as the order and proximity of word occurrence or the location of a word within the document. It also makes no use of the mark-up information that is available from web document HTML tags. A recently developed graph-based representation of web documents can preserve the structural information. The new document model was shown to outperform the traditional vector representation, using the k-Nearest Neighbor (k-NN) classification algorithm. The problem, however, is that the eager (model-based) classifiers cannot work with this representation directly. In this chapter, three new, hybrid approaches to web document categorization are presented, built upon both graph and vector space representations, thus preserving the benefits and overcoming the limitations of each. The hybrid methods presented here are compared to vector-based models using two model-based classifiers (C4.5 decision-tree algorithm and probabilistic Naive Bayes) and several benchmark web document collections. The results demonstrate that the hybrid methods outperform, in most cases, existing approaches in terms of classification accuracy, and in addition, achieve a significant increase in the categorization speed.

In this paper, we focus on the problem of displaying news web sites on mobile devices. Our middleware acts as an interface between the existing news web site and user. News items are extracted using spatial location of HTML elements and a DOM based page segmentation algorithm. The nature of the proposed algorithm is dynamic; it learns from the desktop user trails on news items, and correspondingly displays news item of the highest rank on mobile devices. The rank is calculated based on event and access frequency of news item. The experimental results show that the proposed algorithm saves highly on the bandwidth usage.

The Internet is the most popular and convenient mode of communication for people. An incredible amount of information and many services are provided through the Internet. However, many famous websites often suffer from tampering. Therefore, website security has become an important issue. Recently, Zhao and Lu proposed a PCA-based watermark scheme for web pages, claiming that their process would make web pages tamper-proof and more efficient. However, the PCA algorithm is time-consuming in processing the features of the web page. Therefore, we design a fast fragile watermark for web pages based on hash functions, because hash functions can generate rapidly a digest for the web page. According to our experimental results, our scheme is up to 100 times faster than the PCA algorithm, and, at the same time, our scheme improves case insensitivity and the malfunctions of HTML tags that are drawbacks of Zhao and Lu’s scheme.

Presentation level integration now becomes an important and fast growing trend in enterprise computing and portals are the mainstream to realize it. However, there is not yet a definitive mechanism to achieve interoperability between the basic components of a portal i.e. portlets, whereby HTML data flows smoothly from one portlet to a neighboring one. This paper proposes an HTML fragments based approach to achieve portlet interoperability. Fragments are a block of HTML elements, which are generated by portlets and are used to aggregate a portal page. We first construct a presentation component, which is named as ShadowComponent, for each portlet involved in a portlet interoperation using its fragments, then define a data flow process between ShadowComponents using ECA rules, and finally drive such a process by creating events to fulfill data flow between ShadowComponents. As the fragments of a portlet are synchronized with their corresponding Shadow Component, such a process enables the portlet interoperation. Experimental results show that the proposed approach is effective in achieving portlet interoperability in portals.

Semantic Text Portion (STP) is a text portion in the original page which is semantically related to the anchor pointing to the target page. STPs may include the facts and the people's opinions about the target pages. STPs can be used for various upper-level applications such as automatic summarization and document categorization. In this paper, we concentrate on extracting STPs. We conduct a survey of STP to see the positions of STPs in original pages and find out HTML tags which can divide STPs from the other text portions in original pages. We then develop a method for extracting STPs based on the result of the survey. The experimental results show that our method achieves high performance.

Data on the web is gradually changing format from HTML to XML/XSLT driven by various software and hardware requirements such as interoperability and data-sharing problems between different applications/platforms, devices with vairous capabilities like cell phones, PDAs. This gradual change introduces new challenges in web page and web site classification. HTML is used for presentation of content. XML represents content in a hierarchical manner. XSLT is used to transform XML documents into different formats such as HTML, WML. There are certain drawbacks in HTML and XML classifications for classifying a web page. In this paper we propose a new classification method based on XSLT which is able to combine the advantages of HTML and XML classifications. We also introduce a web classification framework utilizing XSLT classification. Finally we show that using Naive Bayes classifier XSLT classification outperfoms both HTML and XML classifications.

The explosive growth of World Wide Web has already made it the biggest image repository. Despite some image search engines provide con-venient access to web images, they frequently yield unwanted results. Locating needed and relevant images remains a challenging task. This paper proposes a novel ranking model named EagleRank for web image search engine. In EagleRank, multiple sources of evidence related to the images are considered, including image surrounding text passages, terms in special HTML tags, website types of the images, the hyper-textual structure of the web pages and even the user feedbacks. Meanwhile, the flexibility of EagleRank allows it to combine other potential factors as well. Based on inference network model, EagleRank also gives sufficient support to Boolean AND and OR operators. Our experimental results indicate that EagleRank has better performance than traditional approaches considering only the text from web pages.

The Semantic Web needs ontologies as an integral component. Current methods for learning and enhancing ontologies, need to be further improved to overcome the knowledge acquisition bottleneck. The identification of concepts and relations with only minimal user interaction is still a challenging objective. Current approaches performed to extract semantics often use association rules or clustering upon regular flat text. In this paper we describe an approach on extracting semantics from Web Document collections which takes advantage of the semi structured content within XHTML (an XML dialect which can be obtained from traditional HTML documents) Web Documents. The XTREEM (Xhtml TREE Mining) method uses structural information, the mark-up in Web content, as indicators of term boundaries and for co-hyponymy relations.

Facing the retrieval problem according to the overwhelming set of documents online the adaptation of text categorization to web units has recently been pushed. The aim is to utilize categories of web sites and pages as an additional retrieval criterion. In this context, the bag-of-words model has been utilized just as HTML tags and link structures. In spite of promising results this adaptation stays in the framework of IR specific models since it neglects the content-based structuring inherent to hypertext units. This paper approaches hypertext modelling from the perspective of graph-theory. It presents an XML-based format for representing websites as hypergraphs. These hypergraphs are used to shed light on the relation of hypertext structure types and their web-based instances. We place emphasis on two characteristics of this relation: In terms of realizational ambiguity we speak of functional equivalents to the manifestation of the same structure type. In terms of polymorphism we speak of a single web unit which manifests different structure types. It is shown that polymorphism is a prevalent characteristic of web-based units. This is done by means of a categorization experiment which analyses a corpus of hypergraphs representing the structure and content of pages of conference websites. On this background we plead for a revision of text representation models by means of hypergraphs which are sensitive to the manifold structuring of web documents.

A genre or a style is another view of documents different from a subject or a topic. The genre is also a criterion to classify the documents. There have been several studies on detecting a genre of textual documents. However, only a few of them dealt with web documents. In this paper we suggest sets of features to detect genres of web documents. Web documents are different from textual documents in that they contain URL and HTML tags within the pages. We introduce the features specific to web documents, which are extracted from URL and HTML tags. Experimental results enable us to evaluate their characteristics and performances.

We propose a method of acquiring attribute words for a wide range of objects from Japanese Web documents. The method is a simple unsupervised method that utilizes the statistics of words, lexico-syntactic patterns, and HTML tags. To evaluate the attribute words, we also establish criteria and a procedure based on question-answerability about the candidate word.

Tools that allow effective information organisation, access and navigation are becoming increasingly important on the Web. Similarity between web pages is a concept that is central to such tools. In this paper, we examine the effect that content and layout-related aspects of web pages have on web page similarity. We consider the textual content contained within common HTML tags, the structural layout of pages, and the query terms contained within pages. Our study shows that combinations of factors can yield more promising results than individual factors, and that different aspects of web pages affect similarities between pages in a different manner. We found a number of factors that, when taken into account, can result in effective measures of similarity between web pages. Query information in particular, proved to be important for the effective organisation of web pages.

We propose a new web information extraction system, PIES, to convert web information into XML documents. PIES uses a user-specified ontology and HTML tag pattern descriptions. The ontology validates the web information the pattern descriptions extract. We designed a new language to describe HTML tag patterns and extraction rules. We implemented PIES and applied it to the US patent web site for evaluation.

Regional web search engines provide users with spatially restricted web contents. We propose a time-specific regional web search engine, where users can retrieve both spatially and temporally restricted information from the Web, i.e. a description of a shop available at certain time in a certain region. Because there are many web pages containing descriptions of more than one geographic entity, a matching mechanism of spatial and temporal attributes is required. We used HTML tag structures to make matches between addresses and temporal interval expressions. Based on the extracted result, map-based visual interface was implemented.

HTML tables are information rich and are used frequently in HTML documents, but they are mainly presentation-oriented and are not really suited for database applications. To wrap HTML tables, in this paper, we introduce a conceptual model for HTML tables, and based on it, we present a new approach to wrap HTML tables into XML documents. It can automatically convert basic HTML tables, nested tables, composite HTML table and the tables without marked headings, into XML documents.

We present a new approach to automatically convert HTML documents into XML documents. It first captures the inter-blocks nested structure, then the intra-blocks nested structure, which consists of blocks including headings, lists, paragraphs and tables in HTML documents, by exploiting both formatting information and structural information implied by HTML tags.

Increasing proliferation of hand-held devices as well as the need for delivering information to visually impaired persons have caused the need for transcoding web information into documents that can be delivered as audio. Web information is typically represented as HTML (Hyper Text Mark-up Language) documents. Audio delivery of web documents is done using VoiceXML. Due to this difference in mark-up notation, much of the web is inaccessible via audio. One way to solve this accessibility problem is to automatically transcode HTML documents to VoiceXML. In this paper, we describe such an automatic transcoder that converts HTML into VoiceXML. The transcoder is compositional and is realized in two phases: The parsing phase where the input HTML file is converted to HTML node tree, and the semantic mapping phase where each node in the HTML tree is compositionally mapped to its equivalent VoiceXML node. Our transcoder is extensible in the sense that: (i) it can be upgraded easily by users to accommodate modifications to and extensions of HTML; (ii) it provides means for the user to modify the translation logic while dealing with certain HTML tags. The translator is being publicly distributed.

XML is gaining popularity as an industrial standard for presenting and exchanging structured information on the web. Meanwhile, the majority of documents on-line are still marked up with HTML, which are designed specifically for display purposes rather than for applications to automatically access. In order to make web information accessible to applications so as to afford automation, inter-operation and intelligent services, some Information Extraction programs, called ”wrappers”, have been developed to extract the structured data from HTML pages. In this paper, we present a layout-based approach to separate the data layer from its aspect of presentation in HTML and extract the pure data as well as its hierarchical structure into XML. This approach aims to offer a general purpose methodology that can automatically convert HTML to XML without any tuning for a particular domain.

This paper presents MetaNews, an information gathering agent for news articles on the Web. MetaNews reads HTML documents from online news sites and extracts article information from them. In order to recognize and extract articles from an HTML document, MetaNews removes redundant HTML tags, builds a reference string which is a sequence of semantic components of a noise-removed document, and performs pattern matching between the reference string and each of pre-defined information patterns for articles. With a few training inputs from the operator with intermediate-level skills, MetaNews is capable of adding new sites easily and extracting articles in real time. By reducing the complexity of designing and creating wrapper interfaces, the proposed techniques in MetaNews are useful for many information-mediator applications such as meta-search engines, information-push solutions, and comparison-shopping agents.

Providing information on the Internet requires the knowledge of HTML[1], and we have been teaching it mainly to the students who belong to Informatics departments by means of teaching HTML tags. As the Internet has become more popular, many people provide information on the Internet. It has become necessary to teach HTML to students whose major is not Computer Science and who have fewer chances to write HTML but have more chances to provide information on the Internet. In spite of the different goals of these two types of students, we are still teaching them in the same way. Here, taking the difference into consideration, we propose to develop our method of teaching web authoring as to fit current needs. We also consider and try to explain what are the good pages. We also propose a method to teach how to create good pages with simple and restricted conditions. Although they are not perfect, we refer this topic in order to start discussion.

The recognition and extraction of semantic/logical structures in HTML documents are substantially important and difficult tasks for intelligent document processing. In this paper, we show that alignment is appropriate for recognizing characteristic semantic/logical structures of a series of HTML documents, within a framework of case-based reasoning. That is, given a series of HTML documents and a sample transformation from an HTML document into an XML format, then the alignment can identify semantic/logical structures in the remaining HTML documents of the series, by matching the text-block sequence of the remaining document with the one of the sample transformation. Several important properties of texts, such as continuity and sequentiality, can naturally be utilized by the alignment. The alignment technology can significantly improve the ability of the case-based transformation method which transforms a spatial/temporal series of HTML documents into machine-readable XML formats. Throughout experimental evaluations, we show that the case-based method with alignment achieved a highly accurate transformation of HTML documents into XML.

This paper presents the i-Cube environment, a tool-set that allows for Internet data and content originally available as HTML Web pages and programmatic scripts to be denoted, modeled, and represented in the form of XML documents. These XML documents conform to specific Document Type Definitions and other structural constraints that are fully customizable by the end-user or the service provider. The approach is based on representing HTML document data content in the form of annotated trees. Specific areas of interest and data content in the original HTML document that need to be encoded in the form of an XML representation, are represented as a collection of annotated sub-trees in the tree that corresponds to a large HTML document. A service integration module allows for different categories of analysis and presentation rules to be invoked according to script based user-defined logic.

In this paper a Web mining tool for content-based classification of Web pages is presented. The tool, named WebClass, can be used for resource discovery purposes. Information considered by the system is both the textual contents of Web pages and the layout structure defined by HTML tags. The representation language adopted for Web pages is the bag-of-words, where words are selected from training documents by means of a novel scoring measure. Three different classification models are empirically compared on a classification task: Decision trees, centroids, and k-nearest-neighbor. Experimental results are reported and conclusions are drawn on the relevance of the HTML layout structure for classification purposes, on the significance of words selected by the scoring measure, as well as on the performance of the different classifiers.

XML is rapidly emerging, and yet there still exist numerous HTML documents on the Web. In this paper, we present a heuristic approach for converting HTML documents to XML documents. During the conversion process, we eliminate all the HTML elements in an HTML document from the resulting XML document since these elements are designed for the display of data exclusively, but retain the character data of each element along with the implicit hierarchy among the data. The proposed conversion approach extracts the data hierarchy of HTML documents as closely as possible with no human intervention. The approach can be adopted to construct the data hierarchy of an HTML document and to collect data in HTML documents into an XML repository.

Recently, a huge quantity of HTML documents have been created in Internet, which really constitute a treasury of information. HTML, however, is designed mainly for reading with browsers, and not suitable for machine processing, whereas XML was proposed as a solution for this problem. In this paper, we give a case-based transformation method from HTML documents to XML ones. There are many series of HTML pages in actual Web sites, and each page of a series usually has a quite similar structure with each other. Therefore a case-based transformation must be a promising method in practice for a semi-automatic transformation from HTML to XML. Throughout experimental evaluations, we show this case-based method achieved a highly accurate transformation, i.e., 85% of actual 80 pages can be transformed in a correct way, with this case-based method.

QUEST is a system for Querying Semantically Tagged documents on the World-Wide Web. The advent of new markup languages, such as XML, facilitates authoring of Web documents that contain not just html tags for instructing a browser how to view a document, but also contain objects that represent the semantic structure of the document. When such documents become widely available, more powerful methods to access and query information on the Web will be possible. The QUEST system was designed and implemented for querying and manipulating documents written in the markup language ohtml. ohtml combines html and objects of the oem data model. QUEST has several new features. First, QUEST can be used to query a combination of hypertext and object structures. Second, The results of queries are ohtml pages and thus of the same type as the data being queried. Third, QUEST implements a new approach for querying semistructured data that produces meaningful answers even when the input data is incomplete, i.e., when some variables of the query cannot be bound to database values. Finally, the experience of developing and using QUEST for querying semantic documents on the Web can be useful for the design and implementation of query languages for xml. This paper provides an overview of the QUEST system and its components.

We define HTML as an embedded domain specific language in Haskell. To this end, we have designed and implemented a combinator library which provides the means to create and modify HTML elements. Haskell's type classes are employed to enforce well-formed HTML to a large degree. Haskell can then be used as a meta language to map structured documents to HTML, to define conditional content, to extract information from the documents, or to define entire web sites. We introduce container-passing style as a programming convention for the library. We also pinpoint some shortcomings of Haskell's type system that make the task of this library's implementor tedious.

As the WWW becomes a major source of information, a lot of interest has arisen, not only for searching for information, but for reusing this information in new pages, or directly within applications. Unfortunately HTML tags do not provide a significant level of structure for identifying and extracting information, since they are mostly used for presentation issues. Moreover the simple link mechanism of the Web does not support the controlled traversal of links to related pages. Particularly promising is the proposal for a new standard, XML, which could bring the power of SGML to the Web while keeping the simplicity of HTML. In this paper we present a system and a language that allow reusing of information from various sources, including databases and SGML-like documents, by combining it dynamically to produce a virtual document. The language uses a treelike structure for the representation of the information objects as well as link objects. The paper focuses on the selection and the traversal of XML links to extract information from linked pages. The strength of our approach is to be an SGML-compliant solution, which makes it ready to take full advantage of XML for reusing information from the Web as soon as it is widely used.