Exploring the new Manuscripts & Pictorial website - day 5

In this, the 5th and final instalment of our deeper exploration of Manuscripts & Pictorial, I'm going to dive into the technology a little, to cover off some of the things that I care about and which I (naturally) hope you might care about too, once you know they're there.

So to begin, I say:

1. A URL is a contract with the world

Others may mock my obsession with URLs persisting over time, but I say that if someone cares enough about our website to bookmark some search results or a particular page, perhaps sending a link to a friend, or attributing some content in a book or report, the least we can do is ensure that the link still works when they come back to it tomorrow, next year, or a decade later. It's a matter of respect. A simple reference that people can quote when they're discussing something, a reference that can be resolved by anyone's web browser to show them the original item, the primary source! Anywhere in the world, any time. This is what our web forefathers and mothers surely had in mind.

Others, perhaps the same people, perhaps not, may mock my obsession with short, simple, technology-independent URLs (URIs for the pedantic).

To explore this, consider a search for "maus" on Manuscripts & Pictorial:

http://mp.natlib.govt.nz/search/?q=maus

Notice that:

• if we replace the Manuscripts & Pictorial website in 5 years' time, it's not hard to map our old URLs to ones within the new system, therefore ensuring they aren't broken by the technology refresh.


• it doesn't mention the implementation technology at all (words like cgi-bin or search.asp or results.do are common artefacts) so we can change the implementation without changing the URLs


• users can probably guess what they might expect to happen when they go to that URL.


• it's about as short as it can be, which is convenient for starters, but also means that unfortunate people using email clients like Groupwise won't find the link is broken because it spans more than one line.

If your technical people say "we can't do that", ask them to look at using an Apache reverse proxy and possibly even mod-proxy-html. It will make you sound smart, if nothing else. You could even say "I think it's worth the effort, if we care about our users" if you're prepared to be really disliked by your technology provider.

It's nice to see that DigitalNZ gets it - here's the same search on their site:

• http://search.digitalnz.org/search?search_text=maus

2. The Web of Data is here

Or, a MediaRSS feed of your search results is practically a web API.

I can't believe I made it to day 5 without mentioning Cooliris nee Piclens. I say this because in any demonstration of the website, showcasing all kinds of Important and Worthy features, the audience will save their audible intake of breath for that moment when I click on the tiny cooliris icon and show them their results as a 3 dimensional wall.

"Cooliris is not something we developed, I hastily point out, it's something enabled by our custom imageserver and standards-based MediaRSS support. It's a great example of the new 'web of data' applications that complement traditional websites."

But I don't have their attention: looking around the room, everyone is transfixed by this new gap in the space-time continuum that has opened up in the middle of what used to be an ordinary website.

If you haven't tried it yet, I do suggest you head over to Cooliris, download the plugin for your (non-linux) browser, run a search like this one on Manuscripts & Pictorial and click the small cooliris icon in the top right corner of your browser toolbar.

Besides the obvious eye candy there's something interesting in the way that new features keep being added to Cooliris and users of our website benefit, without us having to do any work. This is the promise of the web of data and decoupling and standards, and it's awesome. For instance, users of Manuscripts & Pictorial can now, when using Cooliris, collect up all the pages they're interested in as they go, and then send the collected pages to themselves or someone else by email. The email contains all the images, but also links back to the appropriate page in the Manuscripts & Pictorial website. It's a powerful feature we would love to have had the time and budget to implement, but we didn't have to.

Cooliris is the only kid on the block just now, so we have to put up with their clumsy early attempts to commercialise their plugin. This is a little awkward at times, but the web of data is a level playing field and it can't be long before there's some competing apps in the space time continuum to play with, and then may the best plugin win.

Incidentally, back in March I critiqued the Hard Rock Cafe's use of Silverlight, which offers another kind of immersive environment, but one that's not web-of-data, and works only on their website, as opposed to most of the internet. Judge for yourself. Especially, look back after a year and see which one you used most often.

While we're on the web of data, some of you will have noticed microformats in our people results - the intention is to expand this so that all the documents are events and able to be pulled into a visual timeline on another website. Opensearch integration is something else at the top of the list, so watch this space.

After so long in the workshop with Manuscripts & Pictorial, I'm really looking forward to hearing from people who are using it, and what they think. So do write to me at elliott.young@natlib.govt.nz. And have a great Christmas.