Testing the relevance of contributed or migrated content

Managing content contributions or a migration process from an existing online resource requires patience and discipline, two things of which information designers can never have enough!

In few other situations do things get trickier than when an organisation wants to publish key messages and insists in a fit of decentralisation that its business units represent themselves in such a context.

Let’s assume we have developed a hierarchical information architecture. For the moment, though, it’s entirely empty. How do we describe each node properly in content?

Guiding principles

When deciding on what needs to be communicated by new content or through testing the relevance of old content within a new structure, we can use two time-honoured principles from centuries past:

a. Occam’s Razor

William of Ockham was not actually the inventor of the Razor, but the Friar from Sussex has always been the popular source when referring to the maxim that:

Entities should not be multiplied beyond necessity

b. (A paraphrasing of) Bentham’s utilitarianism

In the context of content:

Deliver the greatest amount of satisfaction to the greatest number of users

Along with his influential theory of utilitarianism, reformer and philosopher Jeremy Bentham can also be credited with the inspiration for the modern prison through his Panopticon.

The questions

Accordingly, these principles require us to describing and explaining a node, we satisfy only the following questions, concisely and in the precise order:

  1. What?
  2. Why?
  3. How?
  4. Who? and optionally,
  5. [When? If the node is axiomatically subject to a period of time]

The content should respond to the above and nothing more, otherwise it has exceeded the remit of its purpose to describe the node.

The content should then provide users with access to:

  • The next sibling element;
  • Thematically-aligned elements;
  • The immediate parent;
  • The root element

If these criteria are satisfied, the node is properly described and explained.

The forced downgrade: going back to Visio for web prototyping

Microsoft Visio 2003

When designing prototypes you could do a lot worse than Visio. But you could also do a lot better. Axure, for example, should make this article irrelevant, as should the fact that I generally have little positive to say about Microsoft in the web realm.

So why this article? Well, I wanted to say a few words about my recent experiences working with Visio 2003 on a client site with no budget for excellent yet costly apps like Axure or the mortgage-worthy iRise. You see, apart from a brief period around 2003, I have been fortunate enough to enjoy working with software whose definition of fit-for-purpose is somewhat tighter than Redmond’s.

The other reason is because there are lots of excellent articles out there about how to replace Visio and I think there might be a bit of mileage in discussing something few professionals in this business ever talk about, namely coping with a forced downgrade.

We’ve all done it: we’ve changed employers, clients or whatever to find that our pastures new have, well, a whiff of old age about them. And your new colleagues will smile at you with a mixture of embarassment and sympathy. Turns out they only just upgraded from Windows NT a couple of weeks ago.

In the agony of that moment, you are haunted by the notion that your complaint about having to use an ancient, usually buggy, version of this or that is likely to fall on deaf ears.

Not so bad?

So here I am sitting in front of Visio, wondering exactly how I’m supposed to fly like an eagle in a coop of turkeys. Well the wings of this particular bird of prey are clipped, and it has to pick clean a corpse of a budget just to get the licence for a knackered old copy of Visio.

When I receive it, the CD-ROM is gleaming with all the innocent hopes of 2002. That’s because it has never been out of the TechNet box. And it takes another week before a miserable-looking support fellow with sufficient rights crawls out of his windowless basement to install it.

Turns out, old Visio isn’t so bad after all. Granted, it hasn’t changed that much since the old days. What it lacks in common sense and workflow it certainly gains in ease of use out of the box: in that regard, it’s classic Microsoft.

It also helps that others have trodden these well-beaten paths in the past. The pain of setting up a decent stencil containing a range of shapes needed to render prototypes has been assuaged in no small part by the sterling efforts of Henrik Olsen and his GUUUI stencil. This stencil is the perfect starting point for prototyping in Visio and it saved me a lot of hassle in the early stages.

Visio stencils for Information Architects

Here come the problems

It wasn’t all plain sailing, of course. Sometime soon, the cracks had to appear in Visio’s everyman panacea, otherwise everyone would choose Visio (me included), wouldn’t they?

1. Updating common objects

Mimicking the purpose of server-side includes, prototyping software should always enable users to save common objects for use across multiple sheets. And it follows that, when a change is made to such an object, that change updates all instances of the object.

Those who start work with a great little stencil like that of Henrik above will quickly find that, working in drag-and-drop fashion, working up sheets is a cinch until they need to update an object (or “shape” as it’s called in Visio).

For example, let’s say you want to change the Font property of all the textual elements from Verdana to Arial. Why would anyone do that? Well, maybe Verdana just bothers you. It does with me, though I read somewhere (and write in if you know, because I can no longer remember the source - it could have been an O’Reilly book) that in Windows, Verdana occupies 13% more horizontal space than Arial. But I digress…

So you update the stencil you’ve adopted and you’ll find that your changes are not reflected throughout your prototype sheets. Don’t tear out your hair though, just read on, because this is one of those special Microsoft Moments in which there is method in the madness. If you skew your thought process forty-five degrees or so, you’ll see it makes perfect sense, like the famous skull in Holbein’s portrait of two ambassadors.

Visio appears to preserve the integrity of your adopted stencil by creating a new stencil on the fly (the “document stencil”) which may or may not be immediately visible to you. My advice is to open and start utilising the document stencil (File > Shapes > Show Document Stencil in Visio 2003, for example) as early as possible in your prototyping process because otherwise it can get messy in there pretty quickly.

Like Holbein’s skull, this approach to maintaining your adopted stencil’s integrity is actually sensible enough when you look at it the right way, it just should have been better managed in the interface. I was clued in after a lot of irritation and Googling brought me to a little forum snippet which explained more or less the same thing.

So if you update the objects in your document stencil, the updates will be reflected throughout your prototype sheets.

2. Managing multiple prototype sheets and hierarchies

Navigating larger prototype documents is difficult in Visio because the software does not enable users to create hierarchical structures. In short, you can’t organise your sheets into folders to make them easier to manage or to better express their relationships to one other.

So whilst a few prototype sheets can be mocked up within a short space of time, when you try to work with twenty or more in a site structure that matches what will eventually become your intended site map, life becomes that much tougher.

Would it be that much of a stretch to have the context menu in the Drawing Explorer provide an option for “Create New Folder” or similar? After all, the developers already generated folders in which to collect the objects on every sheet.

To express an information architecture hierarchy in my Visio document, I found myself resorting to creating pages with names like o Organisation or ooo Policymaking just to create a sort of indent with which to visualise the structure at authortime. Should you ever find yourself having to do this in Drawing Explorer, by the way, ensure you use an alphanumeric as the indenting character (i.e. not something like — Federal support), otherwise hyperlinks between pages will not work.

3. Checking for broken links

If you alter the name of or delete a prototype sheet in a large document, how can you be sure that you’re not breaking hyperlinks? In Visio 2003 you can’t be, since there appears to be no link checking facility. I’d be happy to hear from you if you’ve found this essential feature. Had I the choice, this would probably have been the dealbreaker.

Conclusion

In Visio, you can brainstorm ideas for a new software product. In Visio, you can model your new software product using the conventions of UML. When you release the software, Visio will help you graph the success of sales. On the proceeds, your company can move to new offices and Visio will help you visualise the new desk layout and the organisation chart for your expanding workforce. If the company continues to grow and you need to consolidate, you can map more efficient processes in Visio.

So in common with most Microsoft products, Visio tries hard to be a lot of things to a lot of people and in a general sort of way, it succeeds admirably.

Indeed the three issues I identified above are not so serious that you’ll want to kidnap a minor royal to pay for the alternative. It’s just that other products do the job better and more efficiently and if you have the means, try them.

And I can’t comment on Visio 2007. I haven’t tried it and it’s not part of this article, though Microsoft doesn’t seem to have started a revolution with it. It might well have improved significantly however.

So I won’t say that this forced downgrade, this rollback, this blast from the past, turned out badly. That would be wrong - it was a question merely of making do. I’m not the prototype of a prototyping snob.

CmapTools for concept mapping and OWL authoring

I’d be among the first to admit that, despite being a stickler for standards, sometimes I like to do things my own way.

For use cases, I should be using UML but I take one look at the diagrams and I run for my life.

For preparing ontologies, I should be writing OWL, but I end up writing my own XML variant with (shhh!) no schema. This is (almost) inexcusable, so I’ve decided to try to jump on this particular train and start using a bit of software to get me started.

Introducing Cmap Tools COE OWL

I wanted to have my cake and eat it. I wanted the graphical modelling capabilities of Visio plus the OWL export and minus the Office 2007 interface. Enter Cmap Tools COE OWL: a JAVA-based concept mapping tool with OWL export.

And what better way to introduce an concept mapping product than to display a concept map in place of a homepage? Well, a normal web page would probably have been better in terms of usability but it was a good idea anyway!

Cmap Tools Homepage

Note that Cmap Tools COE OWL is a distinct variation of the standard CmapTools.

According Cmap Tools’ custodian, the Institute for Human and Machine Cognition, the software “empowers users to construct, navigate, share and criticize knowledge models represented as concept maps”.

It’s available in great-tasting flavours Windows, Mac OS-X, Linux, Solaris and the One Laptop Per Child machine XO. It’s also free for individual, non-commercial use.

Working with the software

At such a keen price, I decided to try it out! I installed version 4.11.01

I don’t personally know any JAVA developers, but my experience of JAVA software has often been that the user interface was an afterthought - it seems to be a feature of the platform!

Lest I protest too much, Cmap Tools turns out to be a pleasant exception to the rule.

Cmap Tools screenshot

The UI is relatively straightforward, echoing the product’s overall simplicity. A separate initial window entitled Views is for file management and collaboration. From there, we can start work on a new project.

Most of the donkey work gets done via standard drag-and-drop functionality, enabling rapid drawing from a standing start with no prior experience of the software.

Growing the concept map is dead easy. Users need simply select the parent of the intended new node and drag from the top of the label. Dragging from the top does feel slightly counter-intuitive if one intends to branch below the parent, but that’s a tiny (and probably entirely personal) niggle.

Cmap Tools Nodes

As the ontology spreads, there’s an absolutely essential Autolayout feature that helps the user to bring under control any concept map that starts to look like Phil Spector on a bad day in court.

If, like me, the user likes to see a simplified view of the strict hierarchical order of their taxonomies and ontologies, then to the right under Cmap Outline, he/she will find a more traditional tree layout. What I think is a minor shame here is that the list cannot be edited from Cmap Outline. The same can be achieved in the forms-based adjacent tabs.

Export

Now for the all-important test: how’s the OWL output?

From my very first attempt, I find that the XML result isn’t exactly as expected. I need to go back to the software and try to be more explicit in the relationships between objects before eventually I achieve more predictable output.

Conclusion

Protégé-Owl

Cmap Tools is a great bit of software for simple rendering of taxonomies and ontologies in the form of concept maps.

It’s not a serious development tool but as an all-rounder starter kit that’s safe to put in front of clients (with a commercial licence of course), it’s usable enough to invite the uninitiated to participate.

I’ve also recently started to look at Protégé, maintained by Stanford University’s National Center for Biomedical Ontology. This is the serious tool which, along with the Protégé-Owl extension, takes advanced OWL-based ontology editing and visualisation to another level.