Built for Discovery and Reuse

Persistence

It is usual for European Commission-funded projects, such as Share-PSI 2.0, to be set up on their own domain, however, by creating a sub-site on w3.org, the requirement of persistence is readily met. The W3C's website is one of the oldest sites on the Web and is subject to a strict control policy that is designed to ensure persistence. The use of 'dated space' is a key aspect of this. Even if a future project were to be called Share-PSI, the inclusion of /2013/ (the project's start year) avoids any clash.

Furthermore, w3.org generally enjoys particular prominence in search results due to the nature of the organisation behind it. Therefore, the requirement of discoverability is also addressed. It is, however, important to note that no part of the Share-PSI website is endorsed by the W3C Membership and it is not part of any formal standards. The closely associated Data on the Web Best Practices document, to which many references are made by Share-PSI as part of the advice offered for implementing the (Revised) PSI Directive, is part of that formal standards process.

Discoverability

Discoverability is addressed in various ways. Pages are created using HTML5's semantic markup so that pages are correctly divided into sections, asides etc. Generated content, such as lists of Best Practices and implementations, are generated server-side so they are easily parsed by search engines. Structured schema.org metadata is included in several pages to semantically link the website's advice to the European Commission's (Revised) PSI Directive and to increase the likelihood of search results providing easy to find summaries of the site's content. A sitemap provides search engines with an easy route to the most important pages on the project's website, and this has been submitted to both Google and Bing. Further SEO steps taken include the declaration of a canonical URL on versioned pages and care to ensure that all links are extant (i.e. there should be no 404s).

Reusable and Citable

The Best Practices evolved throughout the Share-PSI project and to manage this, the same version control system was used as is used by W3C for its formal publications. That is, each version of each Best Practice has its own URI that points to a stable, immutable version of the document. A second, shorter URI, always points to the latest version. This is in line with one of the Data on the Web Best Practices: Assign URIs to dataset versions and series. Since w3.org uses Apache servers, this is achieved by installing a .htaccess file in each of the short URI locations. For example, the .htaccess file for Develop and Implement a Cross Agency Strategy is as follows:

RewriteEngine On
RewriteBase  /2013/share-psi/bp/cas/
RewriteCond  %{REQUEST_URI}  !/2013/share-psi/bp/cas-[0-9]{8}/
RewriteRule  ^(.*)$  /2013/share-psi/bp/cas-20160725/$1 [P]

This file needs to be updated each time a new version is published as it hard-codes the redirection to what is currently the latest version of this BP, published on 25 July 2016. Maintaining version control in this way encourages reuse and citation of the Best Practices, either via the latest version link, http://www.w3.org/2013/share-psi/bp/cas and its analogues, or to the immutable, dated version such as http://www.w3.org/2013/share-psi/bp/cas-20160725.

Care has also been taken throughout the Share-PSI website to identify the language in which text is written. The default for the site as a whole is en-gb, i.e. British English, but the local guides are, of course, written in a variety of languages and these have been encoded using the HTML attributes lang and hreflang where appropriate.

Much of the project website is static but there are two datasets that lie behind the dynamic aspects. These are both encoded in JSON and cover the Best Practices and the Implementations (the local guides). These files can be used by any application anywhere if so desired. As noted, for search engine optimisation, these data files are processed server side (using PHP). However, the choice of JSON makes it easy to use same data source in client side scripting, for example, to offer a choice in how the BPs are listed.

Updates and Additions

The downside of hosting the project website on w3.org is that write access is restricted, more so than would be the case if a separate site had been established. To minimise this effect, the JSON files are hosted on the project's GitHub repository and so are more readily edited and updated than if they were on W3C's servers. A copy of the PHP code that drives the dynamic parts of the site is also on GitHub for reuse by others but, unlike the JSON files, for security reasons, this is not the live code used on the site.

GitHub's Issue Tracker is also employed as a means of collecting feedback on the Best Practices. Each BP links to an open issue about that BP on GitHub allowing comments to be made and experiences shared without the need to update the BPs themselves. Future updates to the BPs are possible, but this will require a W3C Staff member to be involved which is a bottle neck that the partners are keen to avoid if possible.