Analysis of NFP Web Sites

Brian Kelly reviews NFP web sites using a variety of automated tools, and makes some recommendations on a number aspects related to the design of web site architecture.

National Focal Points

National Focal Points (NFPs) have been established in all European Commission Member States and in other European countries to promote the Telematics For Libraries programme and to assist proposers requiring sector-specific advice and information.

The different countries have taken a variety of approaches to the services provided by NFPs. The aim of this article is to review the different technical approaches taken by NFPs in the provision of their web sites. The issues which emerge may be of use in developments under the Fifth Framework.

In this Web Technologies column a report of a number of mainly automated analyses of the web sites used by National Focal Points is given. Note that no attempt has been made to analyse the content of the web sites. The analyses were carried out on 4-5^th October 1999.

NFP Web Sites

The analysis of NFP web sites makes use of the central list of National Focal Points maintained by the Commission [1].

Of the eighteen countries listed eleven provide an NFP web site. Details of the web sites addresses is given below.

The UK web site provides access to an archive of postings to a mailing list. As it is different in character to the other NFP web sites it is not analysed further in this article.

URL Naming

As can be seen from Table 1 a variety of URL naming schemes are in use. No fewer than six of the web sites include a file name (all of which end in .htm). One web site uses the tilde (~) convention. Two sites (Finland and Norway) use short URLs with a clearly defined directory used to manage the contents.

A review of URLs for Telematics for Library web sites [2] provided advice on URL naming conventions, which are repeated below:

• Short URLs are desirable. Short URLs are desirable as they are easier to cite and to type.
• Directory names are preferable than file extensions. It is often better to cite a URL containing simply a directory name than a file name, as the former can be shorter and avoids possible confusion over the file extension (e.g. http://info.rbt.no/eu/ is shorter than http://info.rbt.no/eu/intro.htm and avoids confusion with http://info.rbt.no/eu/intro.html).
• Avoid the ~ character. Many users will regard URLs containing the ~ character as a personal home page, providing personal rather than organisational information. Also the ~ character may be difficult to find on the keyboard.

These simple guidelines should make NFP web sites more accessible. Of course these guidelines may conflict with local policies for hosting web sites, so they should be regarded as guideslines and not rules.

Server Technology

UKOLN's doc-info [3] and http-info [4] web-based document analysis services were used to analyse the entry point of the NFP web sites. A summary of the findings is given in the following table.

Seven of the web sites are (probably) hosted on a Unix server and three on an Windows NT server.

It was interesting to observe that only one instance of use of metadata was present in the NFP web site entry points.

Analysis Of NFP Web Sites

The Microsoft SiteServer software [5] was used to analyse each NFP web site. A summary of the findings is given in the following table.

Due to the lack of a directory to differentiate the contents of the NFP web site from other areas on the server the information for a number of the web sites listed in Table 3 is likely to be too large.

It should be noted that failure to use a directory structure to group resources related to the NFP not only makes auditing difficult. It also makes it difficult to provide automated harvesting of the resources.

Links To NFP Web Sites

The linkpopularity.com web site [6] was used to obtain details of the number of links to NFP web sites. The results are given in Table 4.

It should be noted that the information on the number of links is taken from the databases hosted by the AltaVista, Infoseek and HotBot search engines. It cannot be guaranteed that the information held on the databases is complete. In addition the linkpopularity.com web site states that the AltaVista results have been erratic recently.

Searching NFP Web Sites

None of the NFP web sites appeared to provide a search facility. Since the sites appear to be small, browsing may be adequate for exploring the web sites. However as the web sites grow, a search facility will become of increasing importance.

As described in an analysis of search engines on UK University web sites published recently in Ariadne [7] small organisations which possess limited technical expertise may find it useful to provide access to a search facility hosted remotely, such as a global search engine.

An example of this approach is shown below. This example uses the HotBot search engine. The interface is configured to search across all NFP web sites by default, although searches of individual NFP web sites can also be chosen.

Using HotBot To Search NFP Web Sites

Search All NFP web sites Austria Belgium Finland France Germany Ireland Norway Spain Sweden Switzerland the Web for

---

Note The following points should be noted:

Not all pages will be indexed

The search is limited to the pages which have been indexed by the HotBot indexer. HotBot may not have indexed all pages on the server, and the pages may have changed since the pages were last indexed.

All resources on server may be searched

The search is carried out across all resources on the server, and not just on the NFP pages. For example all pages on the server ireland.iol.ie will be searched and not just those under the directory ireland.iol.ie/~libcoun/.

Although web sites will ideally have their own search engine which can be configured to support local requirements (e.g. index new resources when they become available, index a range of file formats, omit certain resources from the index such as draft document, etc.) the use of a remote index may be worth considering, especially if remote search service allow searches to be restricted to areas of the web site.

Support for the Robot Exclusion Protocol

The Robot Exclusion Protocol [8] enables a web site administrator to specify directories which robots should not access. Although it does not provide a security mechanism this protocol can be used to avoid search engines indexing draft documents and personal files. It can also be used to stop search engines from wasting server capacity by attempting to index files such as images, CGI scripts, etc.

The Robot Exclusion Protocol makes use of a file with the name robots.txt which is located at the root of the web server. A typical file (taken from the Irish NFP web site) is shown below, with annotations explaining the role of the statements.

An analysis of the robots.txt files on the servers which host NFP web sites is given in Table 5. UKOLN's web-based /robots.txt checker [9] was used to analyse these files.

404 Error Pages

The 404 error page can be an important navigational feature for web site, especially for web sites which have long URLs which may be difficult to type correctly. As described in an Ariadne article which analysed 404 error pages provided on UK University web sites [10] there are a range of features which can be provided on a well-designed 404 page.

A brief summary of the 404 error pages for NFP web sites is given in Table 6.

As can be seen from Table 4 only the Irish NFP web site provides a tailored 404 error page - the remainder use the default server message.

Accessibility Of NFP Web Sites

The Bobby tool [11] was used to analyse the accessibility of the main entry point for NFP web sites. The results are summarised in Table 7.

Conclusions

The analysis of the NFP web sites has shown that a variety of approaches have been taken. With the ever-increasing importance of dissemination of the activities funded by the European Commission the role of the National Focal Points (and their successor under the Fifth Framework) will become even more important. As the web becomes more sophisticated it is important that web sites are designed to facilitate automated processes, and not just for viewing by humans.

Based on this survey a number of recommendations can be made.

Short URLs

Short URLs are desirable as they are easier to remember, cite and type than long URLs.

File extensions may be confusing

There may be ambiguities in the use of file extensions (such as .htm and .html). In addition the use of server default file names (such as intro.html or index.html) can reduce the length of the entry point (e.g. http://renki.helsinki.fi/eu/ instead of http://renki.helsinki.fi/eu/index.html).

Store the content of NFP web sites in their own directories

The content of NFP web sites should be stored in their own directories, and the NFP web site entry point should be located within this directory.

Use metadata

Metadata such as <META NAME="keywords" VALUE="..."> or <META NAME="content" VALUE="..."%gt; can improve the ranking of the web site in search engines.

Beware of frames

Many search engines cannot access framed web sites. If frames are used there should be an alternative mechanism for search engines (and users with browsers which don't support frames).

Analyse link popularity

Use services such as linkpopularity.com to analyse the number of links to your web site. Consider submitting details of your web site to search engines.

Use the Robot Exclusion Protocol

Use the Robot Exclusion Protocol to stop search engines from indexing files which should not be widely disseminated.

Investigate options for implementing a search facility

Investigate the options for implementing a search facility, including use of a remote service, if appropriate.

Consider enhancements to the site 404 page

Think about enhancements to the web site 404 error page.

Accessibility

Think about the needs for people with disabilities who are accessing your web site. Consider use of an accessibility checker such as Bobby.

Consistency across NFP web sites

Think about the advantages in providing a consistent interface, structure and methodology across NFP web sites.

Develop good working links with the web server administrator

The web server administrator may be responsible for implementing a number of the recommendations given here (for example updating the robots.txt and the 404 error message files and providing and configuring a search facility).

It is hoped that these recommendations may prove useful to new Commission-funded and other project web sites which are about to be set up.

Summary of the Tools Used

The following tools were used to carry out the analyses:

doc-info, http-info and /robots.txt checker

Freely-available web-based tools for analysing web resources and HTTP headers. Developed by UKOLN as part of the WebWatch project.

Microsoft SiteServer

Web server software which includes an analysis component.

linkpopularity

A web-based service for analysing the number of links to a web site.

Bobby

A web-based service for analysing the accessibility of a web site.

References

1. National Focal Points, European Commission DGXIII - Telematics for Libraries
   URL: <http://www.echo.lu/libraries/en/nfp-list.html>
2. Web Technologies: URLs for Telematics for Libraries Project Pages, Exploit Interactive issue 1
   URL: <http://www.exploit-lib.org/issue1/urls/>
3. doc-info, UKOLN
   URL: <http://www.ukoln.ac.uk/web-focus/webwatch/services/doc-info/>
4. http-info, UKOLN
   URL: <http://www.ukoln.ac.uk/web-focus/webwatch/services/http-info/>
5. SiteServer, Microsoft
   URL: <http://www.microsoft.com/siteserver/site/>
6. linkpopularity.com, linkpopularity.com
   URL: <http://www.linkpopularity.com/>
7. WebWatch: UK University Search Engines, Ariadne issue 21
   URL: <http://www.ariadne.ac.uk/issue21/webwatch/>
8. Standard for Robot Exclusion, Koster, M
   URL: <http://info.webcrawler.com/mak/projects/robots/norobots.html>
9. http-info, UKOLN
   URL: <http://www.ukoln.ac.uk/web-focus/webwatch/services/robots-txt/>
10. Web Watch: 404s - What's Missing?, Ariadne issue 20
    URL: <http://www.ariadne.ac.uk/issue20/404/>
11. Bobby, CAST
    URL: <http://www.cast.org/bobby/>

Author Details

Brian Kelly
UK Web Focus
UKOLN
University of Bath
Bath
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, "Analysis of NFP Web Sites", Exploit Interactive, issue 3, October 1999
URL: <http://www.exploit-lib.org/issue3/nfp-websites/>

Issue Home | Editorial | Features | Regular Columns | News and Events | Et cetera
Go to Top	A UKOLN Service. Contact Us. Copyright © 1999-2006	Last Updated: 25 October 1999