Behind the Exploit Interactive Web Site

Brian Kelly describes developments to the Exploit Interactive web site.

Background

The Exploit Interactive web magazine aims to provide information on Telematics for Libraries project work, and more general information of interest to the library and information professional. As well as its information dissemination role, Exploit Interactive enables UKOLN (the publishers of Exploit Interactive) to gain experience in the provision of a high quality web service and to use it as a proof of concept for its applied research interests.

In this article we summarise the technical architecture for Exploit Interactive, describes some recent innovations, and outline ideas for future work.

The Technical Architecture

Exploit Interactive is hosted on a Windows NT server. The Microsoft SiteServer software is used, which provides a high-level web management system which sites on top of the Microsoft IIS web server software.

Articles are normally received using email. The articles are in either HTML (based on the Exploit Interactive article format) or Microsoft Word format. MS Word articles are converted to HTML using the import feature of an HTML authoring package - normally HoTMetaL.

Articles are stored as HTML fragments. The server-side include (SSI) technology is used to merge the article fragments with the navigational elements and other items in the header and footer. The server-side include technology also makes use of Active Server Pages (ASP) which enable resources to be processed into an appropriate format.

Recent Developments

Searching

Searching Exploit Interactive has been enhanced through use of Dublin Core metadata. Each article has the author's name included as a DC.Creator Dublin Core attribute. The SiteServer's indexing component supports fielded searching, which enables searches on the author's name to be carried out, as illustrated below.

Figure 1: Search Using the Author's Name

As well as the author's name, Dublin Core metadata is also provided for a description of the article, which is normally taken from the initial lead-in paragraph of the article.

Notification

Issue 4 of Exploit Interactive sees the launch of an automated notification service. Readers who wish to be receive an email message when a new issue is available simply have to provide their email address on the notification page - http://www.exploit-lib.org/notify/.

The notification service, illustrated below, makes use of Netmind's Mindit service [1].

Figure 2: Exploit Interactive's Notification Service

Promotion of the Exploit Interactive Web Site

As described in the article on Promoting Your project Web Site elsewhere in this issue [2] a number of tools are available which can be used to ensure that web resources are indexed by search engines or included in directory services. Use of a number of such packages was made recently. The results are summarised below.

Although such results must be treated with caution, we intend to keep a record of the results, in order to detect trends.

We also investigated the number of pages from the Exploit Interactive web site which were indexed by three of the major search engines. The results are given below.

Error Messages

Issue 3 of Exploit Interactive featured an analysis of NFP web site [3]. The article provides some suggestions for improving the dreaded 404 error message. The Exploit Interactive web site has implemented some of the suggestions: the 404 error message makes use of the Exploit Interactive style, contains a search interface and provides tailored messages, depending on whether the incorrect URL was entered directly or a link was followed, as illustrated below.

Figure 3: A 404 Error Message

For example, you should get a different message if you follow this invalid link or enter the invalid URL http://www.exploit-lib.org/foo directly.

Use of the Robot Exclusion Protocol

The analysis of NFP web site mentioned above also suggested that web sites should make use of the Robot Exclusion Protocol by adding to the /robots.txt file the location of areas of the web site which should not be indexed by robot software. The /robots.txt file on the Exploit Interactive has been updated to prevent robots from indexing areas other than the main issues' directories.

Improved Style Sheet Support

A number of our readers reported difficulties in printing some articles. This is due to poor support for style sheets in certain versions of the Netscape browser. In order to overcome this problem we have introduced a simple test for the browser type. Netscape versions 3 and 4 will have a simple style sheet file embedded, which should overcome these difficulties. We have also made a number of minor changes to the default style sheets and provide a different style sheet for displaying on the screen and for printing.

These changes have been made for issue 4. If they appear to provide an improved service, we will update the back issues to link to the new style sheets.

Additional Information

We now provide a link to further information about each article. Currently this enables the articles to be checked for accessibility and conformance to HTML standards. Following experiments in the use of an automated translation service in issue 3 [4], we received a number of favourable comments ("Good idea to have a link to Systran. Have tried it with French and German, it's not bad as a starter! and Cool!). A link to the Babelfish translator is now included in the further information section. We are also testing a facility for providing statistics on accesses to articles.

Statistics

We have introduced an online statistical analysis service, using analysis services provided by Nedstat [5] and SiteMinder [6]. A review of the service, together with other approaches to analysing web site statistics, will be featured in the next issues of Exploit Interactive.

In The Future

Although this issue has seen a number of new developments releases we do not intend to stop development work. Ideas we are currently considering include use of a database for managing parts of the web site (e.g. article metadata, author details, etc.) and access to a range of statistics about the web site. Read the next issue of Exploit Interactive to find out how developments have progressed.

References

1. Mindit, Netmind,
   <URL:http://mindit.netmind.com/mindit.shtml>
2. Promoting Your Project Web Site, Exploit Interactive, issue 4,
   <URL:http://www.exploit-lib.org/issue4/promotion/>
3. Analysis of NFP Web Sites, Exploit Interactive, issue 3,
   <URL:http://www.exploit-lib.org/issue3/nfp-websites/>
4. Extending Your Browser With An Automated Page Translation Feature, Exploit Interactive, issue 3,
   <URL:http://www.exploit-lib.org/issue3/translation/>
5. Welcome | Nedstat Basic Website, United Kingdom, Nedstat,
   <URL:http://uk.nedstat.net/>
6. Sitemeter, ,
   <URL:http://www.sitemeter.com/>

Author Details

Brian Kelly
UK Web Focus
UKOLN
University of Bath
Bath
England
BA2 7AY

URL: <http://www.ukoln.ac.uk>
Email: b.kelly@ukoln.ac.uk

Brian Kelly is UK Web Focus. He works for UKOLN, which is based at the University of Bath

For citation purposes:
Brian Kelly, "Behind the Exploit Interactive Web Site", Exploit Interactive, issue 4, January 2000
URL: <http://www.exploit-lib.org/issue4/exploit-interactive/>

[HTML Validation] - [Accessibility check]