Semantics through Action - Scenarios

Scenario 1 - academic making a list

Jane is a lecturer specialising in the history of musical instrument production. She is looking at an article and realises it will be perfect for a course she is teaching in the autumn. She finds the publisher's web page for the journal, locates the article and clicks the Aspire bookmarklet. Unfortunately, many of the articles in her discipline are published in a few specialist journals and are not recognised by the standard Aspire bookmarklet.

So, she is prompted to enter the details by hand, but, to help her out, Aspire shows the original page alongside the list of bibliographic fields to fill in and she is able to click on items on the page and they appear automatically in the form. This is faster and easier than retyping.

What she doesn't realise is that behind the scenes the new Aspire has been observing.

A few days later, Jane finds another article in the same journal that she bookmarked previously. She clicks the bookmarklet, and this time, instead of having to select each item, it has pre-populated the bibilographic-details form. She has to adjust one of the fields as this is an editorial, and has a slightly different layout, but otherwise she is simply able to confirm the details and continue.

"So much easier", she thinks as she remembers how it used to be entering the complete details every time.

Scenario 2 - the subject librarian

Jules is a subject librarian for Humanities and is aware that many of the articles are poorly listed. He knows a lot of the problem entries have DOIs or links to the publishers' sites, but do not have clean details because the relevant journals did not have bookmarklet recognisers.

Happily Jules has recently been to an Aspire training session, which included creating bookmarklet recognisers. He knows he can produce these using a specification notation, but is a little nervous of doing this entirely from scratch. However, he was also shown how to create them semi-automatically as a side effect of manually correcting entries.

He chooses one of the most popular journals, and finds several examples of articles in reading lists. He bookmarks them, then goes to the Aspire recogniser designer and selects the relevant resources to work with.

Just like Jane, Jules is able to correct the entries in the resources, and like her, after completing the first resource by hand, for the rest he is able to simply confirm most of the automatic entries. However, he also knows he can examine the patterns inferred by the machine-learning algorithms used by Aspire. These are largely in CSS, which he has seen in web design courses, and he is able to slightly modify some of the rules, for example, saying which entries are required for a rule to 'match', and also check the overall URL pattern being matched.

After he has edited the rules, Aspire finds more resources matching the URL pattern, and Jules is able to verify it is working.

Finally, Jules tells Aspire to apply the rules to all existing entries. He has the option of telling it to completely rewrite the bibligraphic data with the generated data, or, if he is more cautious, to add it as tentative information, which he can check by hand himself, or will be presented to academics next time they are viewing their own reading lists for them to confirm.

Of course, in future, when an academic uses a bookmarklet on the journal, the details are filled in using the newly created recogniser.

Scenario 3 - expanding the range of resources

Jocelyn is an acedemic using Aspire Community edition. She works in the Department of Computing and New Media, and her course materials often include references to code::art, a website that showcases snippets of Processing code. She uses the Aspire plugin to embed rich resource snippets into her course websites, which she develops in WordPress. Unfortunately, the default Facebook-like image plus title + text snippet is not ideal for some of the kinds of sites she frequently refers to.

She is able to override the default presentation using standard options, for example, choosing an alternative icon, and omiting the 'accesssed on' date. However, she would like to be able have a more site-specific embedding including the full Processing code as a preview. Of course, this is precisely what oEmbed is for, and the Aspire plugin can use this if available, but few sites have oEmbed definitions, and is great for popular websites such as Flickr, but not specialised ones.

Happily, she read about the Shared External Semantics Initiative (pronounced 'sexy' ;-)) on the last Aspire Community mailing and decides to have a go.

In fact it turns out to be very like the bibliographic entry system that she is used to (as in Scenario 1), except that instead of having a predefined set of fields, she needs to define her own. She starts with standard entries from predefined bibliographic and oEmbed schemas, but also adds a few fields of her own, noteably for the Processing code snippet. She then selects a layout template for embedding the resource (she knows she could define her own, but the standard ones look good enough). She drags elements of data she has defined for the page into areas of the template, both the on-page part and the preview pane. When finished, she views it on her page and likes the effect.

Next time she visits code::art, she is offered the choice of the schema she defined last time, she confirms this and it autofills entries, as in previous scenarios. This time her own page is slightly different and she would prefer pop-up infoboxes rather than inline list-style references, so she selects and modifies a different layout template, dragging in fields as before.

Thereafter, when she includes items from code::art, she finds she simply has to choose one of the two layout styles she has previously used and everything else is automatic.

However, she also recalls the discussion of 'Shared' in the original mailing and is reminded by the 'share semantics' tab that has appeared on her Aspire home page. She decides to have a look.

"Do you like the look of code::art, you have produced?", it says, "Why not share it with others?".

The sharing page then goes on to explain she can share the semantic annotation she has produced with others, both within the Aspire community and externally. She clicks the 'share as oEmbed' button and the relevant fields of the specification she effectively created through her interactions are transferred into the SxSi crowdsourced semantics site.

She tells her colleagues that she has shared the definition, so that they can use it in their own use of Aspire. She hopes that they might be encouraged to add details of their popular sites so she doesn't always have to do them all, but also is a little proud of the kudos when her colleagues link to code::art resources in Aspire and see the option "use infobox style by Jocelyn".