Quick Technical Wins

AGIMO - WPG Review Team
The Department of Finance Archive

The content on this page and other Finance archive pages is provided to assist research and may contain references to activities or policies that have no current application. See the full archive disclaimer.


While a major site redevelopment and renewal is underway, business as usual still needs to continue.  Some of the minor technical changes which don't require high-level sign-off are easy to roll-out in parallel.

So, this week, without any major changes to our underlying technology, governance arrangements or content scope, we've been able to quickly perform the following:

1. Customised Agency Search results

Previous implementations of our search results page relied on an XML wrapper to process the results from Funnelback's Agency Search service - with our move to some new infrastructure a few months ago, this process 'broke', and the quick fix was to move back to the hosted variation of the Agency Search Service (using the default look and feel).

Only recently have we been able to secure some development time to align the search results screen with the rest of the site's look and feel (and meet WCAG2 checkpoints regarding form input). Previously, we had a script-only solution to prevent empty searches being run - now we've been able to adapt the Agency Search template tags to provide a server-side check (the belt and suspenders plan).  We've included the code snippet for re-use by anyone else using the Agency Search service:

<s:Compare <s:cgi>query</s:cgi> == > <p id="noSearch">No search term was entered. Please enter one or more search terms.</p> </s:Compare>

Thankfully, the web-based administration and customisation process provided by Agency Search was very straightforward, and we were up and running and fully 'skinned' in only a couple of days.  That said, let us know if you spot any problems.

No business case or high-level sign-off required for this one - it was regarded as restoration of an existing service.  Design, test and development time:  10hrs.

2. Fixing broken links

This used to be a painful process given our internal infrastructure and the proliferation of clunky 'free' web-based tools, until the W3C link checker really took off.  Revising redirects, while not 'broken' links, is also included in this process.

No high-level sign-off required, as it's part of our monthly 'weed-pulling'. Design, test and development time:  0.5hrs.

3. Fix up some typos in diagrams

These are a little harder to spot via automated spellchecking methods, and highlight the need to keep editable versions of web-friendly images on hand (and, ideally, ensure that they're editable by almost anyone on the stock-standard SOE).  We began to ponder what additional formats (if any) it may be appropriate to provide these diagrams in for agency re-use.  SVG, PNG and PDF (with Visio also considered) seem to be leading the pack for now....

No high-level sign-off required. Design, test and development time:  0.25hrs.

4. Add some site-wide metadata

The site has been exposing only a bare minimum of metadata on its homepage for some time now, and we're hoping to flesh it out over the next couple of months across [nearly] every page on the site.  Given the site is presently only around 90 pages (and we expect to grow a little with some new content), it appears to be an achievable goal.

Site-wide metadata that was very straightforward to add included:

  • content-type (constant)
  • robots (constant)
  • Content-Language / DC.Language (constant)
  • DC.Format (constant)
  • DC.Coverage.jurisdiction (constant)
  • DC.Rights (constant)
  • DC.Publisher (constant)
  • DC.Identifier (system generated)
  • AGLS.Function (constant)
  • DC.Title (generated from CMS)
  • DC.Description / Description (generated from CMS)

No business case or high-level sign-off required (internal re-use of existing content).  Design, test and development time:  9hrs.

5. ...and prepare for more metadata

We'll be aiming to flesh out some of the page-specific metadata fields (DC.Contributor, AGLS.Audience, DC.Date.*, DC.Subject, AGLS.Mandate, etc.) gradually over time, particularly as content author participation is expected to be required.

National Archives Records Management Agency Support service also came in handy here (thanks David) when figuring out how to expose (and use linked data principles) AGLS.Mandate values (coming soon).

<link rel="AGLSTERMS.mandate" href="http://www.comlaw.gov.au/examplepath" title="Example Act 2009 (Cth)" />

We'll also be pursuing a few controlled vocabularies for populating:

Comments (6)

Sub-headings 1, 2 & 5 should be in lower case, i.e.
1. Customised Agency Search results
2. Fixing broken links
5. ...and prepare for more metadata

No mention of RDFa and AGLS?

Hi Daniel -

There is mention of AGLS-specific metadata in the post (Mandate, Function), along with the Dublin Core components of AGLS.

The syntax we'll be using for referencing these elements is now "DCTERMS.*" and "AGLSTERMS.*" instead of "DC.*" and "AGLS.*", to reflect changes in both of those standards.

What isn't discussed at this point is how that metadata will be exposed publicly. Several mechanisms are currently under discussion, including:

- html head meta
- RDF/XML alternative versions referred to using link rel="alternate"
- Embedded RDFa
- Metasearch-driven user navigation
- Any combination of the above

Ideally, we'd like to maximise the return on any metadata investment using any and all available means.

Unless its been fixed, the AGLS RDF endpoint is broken (it was the last time I checked, but that was september 2009). This makes RDF representations of AGLS metadata effectively useless until it starts working.

Thanks, xtfer -

I suspect that this is the endpoint you're referring to?




If you're observing shortcomings with this file, you'd do well to get in touch with:

David Bromage, National Archives of Australia

Comments on this post are now closed. Please let us know if you would like to discuss this post.

Last updated: 27 July 2016