But this is not just a standard tag based system either. When a user first selects the rate it button they are presented with a list of adjectives to describe the talk they have just seen. They are also told they have three votes to allocate to these listed adjectives, either one vote each to three different ones or divided between two or less. This lets the user feel a great deal of control over how they are rating a talk. Especially on the first time rating where the user has not yet seen the ratings screen.

Once the user has entered their votes and submitted them they are then presented with a tag cloud of terms showing how they and everyone else has voted.

So some people would think that calling this tagging is a mistake in terminology, but I think the name fits. A user is still tagging the media with their choice of terms, they just have a more limited vocabulary to work from. Is this a useful way of rating talks, well I think yes and it could also lead to some very useful ways to navigate the site based upon these ratings.

But how could it be better? Well apart from some simple UI points concerning small hit areas for the ratings boxes, required explanatory text and the ratings method being un-obvious, I think the greatest area for improvement is around the controlled vocabulary itself. How about a system where users were allowed to suggest more terms if they thought they had somthing more appropriate. The system could remain a controlled vocabualry system in that various methods could be used to control the flow of new terms. firstly the mere fact that a user would have to enter the new terms themselves would lead to them being less common, but if certain ones keep appearing then the offered vocabulary could be extended, or altered as some terms removed if little used. Editing of tag submissions could also eradicate synonyms and the appearance of extra tags on the ratings cloud could prove interesting for users. For users of course most of this would be invisible, they would just have an option at the bottom of the form to add their own tag.

Plus sage (feed reader extension for firefox)… Why don’t I have all my feeds in netvibes? How is my use of sage different from the way I use feeds in netvibes? The feeds that I have in netvibes are mostly aggregating feeds ( like reddit ), or really frequently updated ones ( like boing boing ). I put them all together so that I can tell at a glance when they have
updated, and also see the headlines without opening each feed one by one. The sage feeds on the other hand tend more to be more often single authors or less often updated magazines ( like ALA ). These feeds more often have single articles which I might read at my leisure, rather than a list of new posts to scan and only read a few of.

I have though thought about bringing all my feeds into netvibes but haven’t gotten very far. The such new feeds tab contains a few new feeds, bit also a couple of tools. So really that tab is more of a mish mash than a real experiment. Having said that though, I did have a couple of those feeds in my sage list at first which didn’t work out so well

The whole upshot of all this is that for me at least, I consume different ways. Ways which depend upon many factors, from update frequency to feed content. The way that I use my rss feeds mean that just one way, such as netvibes doesn’t work for me. But could it? Working on several different computers means that my unread feeds on sage are always out of sync. Perhaps what I need is a full page feed reader widget for netvibes. So there is a project I could do…

Meanwhile the tools tab is really underused by me at the moment. There are times when I use it, mostly job related times. Like if everyone at work is using google calendar or RMTM todo lists then I use the tools tab quite a lot. Mostly though it is neglected.

As for the delicious tab.Well all I use this for is when I have too many tabs open and I decide to send the bookmarks to delicious so that I can remember that I haven’t read the articles. Not tremendously efficient and not really a great way to use delicious but it beats having 150 tabs open.

So on that note, tonight I am upgrading my netvibes to the new version, ‘ginger’ we’ll see if it changes me…

Overall I was actually a little disappointed with the conference. I felt that the topics of the speakers and the quality of the presentations wasn’t really as good as I expected. While that is my opinion as an inexperienced presenter, I am not alone in thinking it. On the following Monday in my office we had a discussion on how we think the conference could be improved for next year. One of the first things mentioned was more community involvement in selecting presentation topics. Maybe not on the scale of SXSW, but well just more.

One thing which I thought would be a great step forward next year might be a mentoring program for less experienced speakers. We also talked about the idea of panels and workshops. I didn’t go last year due to my exciting New England roadtrip but I have heard both good and bad things about this event. I think it is fantastic that Eric Shied has put in all the work to create a really good event, and it was very well organised and fantastic to be able to see and meet so many peers. But it could be better and it is up to the community to make it so.

Technorati Tags: ozia2007, tagging, conferences

Meanwhile here is a link to the presentation Exploring multidimensional tagging (PDF 1.5MB).

Quick summary of presentation

I briefly talked about what tagging is and where it has gone in the last few years. I discussed some of the failings of tagging but was enthusiastic about its potential. I then went on to outline a few methods to add utility to a tagging framework including multi level tagging, tag groups and not-tags. I then outlined why I thought these might be good ideas, and opened the floor for questions

Technorati Tags: ozia2007, tags, tagging, speech, presentation

Anatomy of a redesign at fairfax digital

Liz Pek, and Andy Coffey described the workflow behind the recent redesign of smh.com.au and theage.com.au. Using a solidly UCD based process mixed with strong collaborative processes. Liz described going through many design iterations exploring different ways of displaying the information on the news page while fulfilling the business and user’s needs. She also described FD’s use of Personas and otherUCD tropes. Andy talked about his teams use of conceptual design in the early stages and a strong grid later in the process to visualise the information on the site.

Forget page length, think information scents

Iain Barker told us about a discussion taking place online discussing the myth of users disdain for scrolling. He showed some interesting statistics about users scrolling and the numbers of users who are happy to use links far down the page. He then went on to discuss the concept of “information scents,” and how the composition and “flavour” of a piece of information is probably more important than how deep down a page it is. All in all Iain’s focus was on context, and how length of a page and the efficiency of that design are completely dependent upon that context.

Optimising landing pages with numbers

Hurol Inan, talked about various method to optimise information on a page. He talked about understanding metrics, gaining control of the page / site. He talked about an approach based upon testing (technical and user testing), collaborative design and several best practice concepts.

Statistics with Steve Baty

OMG.. too much math. Actually an interesting presentation if not really about a topic which relates closely to me. Mostly by choice, we can’t know everything. Steve talked about understanding research measurements, selecting research subjects and statistical sampling, describing the data and a bunch of formulae. He also showed us some example of bad statistical researchand highlighted some dangers with statistics.

more later…

Technorati Tags: ozia2007, ia, design, conference

When I talk about ambient personalization though I am thinking about the construction of intelligent systems which learn from how the user uses the website and then go on to improve the experience for that user. Simple versions of this are already quite common. Amazon.com, for example, uses the items that a user views to help build recommendations for the user. The power of even this relatively simple example can be clearly seen in the pages of my credit card statement.

My presentation linked below is an attempt to explore some of the possibilities and some of the pitfalls inherent in asking a system to try to learn to understand users.

ambient personalisation (PDF 2.4Mb)

But there are of course increasingly times when some kind of Turing test is required, so what are the alternatives?

1. Logical gatekeeper

   The form comes with a question similar to the CAPTCHA, but instead of a hard to decipher image a simple logical question is asked such as “Is ice hot or cold?” or “How do you spell BLUE?” which requires the user to type an answer. Or a multiple choice question using a select box for the user to pick the right answer. This method delivers most or all of the benefits of the CAPTCHA while easing the burden on the user and being more accessible. The image below offers an example from zeldman.com
   

2. Session variable / GET request detection

   This is an unobtrusive alternative to CAPTCHA use, but it can be used to filter out spam-bots. The idea is that you put something in session when a GET request is made and when a form is submitted you check the session for that variable.

   This can filter out stupid bots that submit request directly to POST without getting a page with the form. Of course this system can be fooled by creating a bot that acts like a web browser so would be best used with other safeguards active as well.

3. False form elements

   Again this is an unobtrusive alternative to the CAPTCHA, but a method to reduce both bot attacks and user difficulty. The idea here is to include a dummy form element on the page that any bot will fill in but a human will leave (see example and details). The dummy element should have a real enough name to attract a bot but should be hidden from a user. Thus when the form is submitted you check to see if the hidden fields have been filled in and if so assume a bot is at work.

4. Counting key presses

   This method is also unobtrusive and relies on the fact that users actually enter information into the form while a bot will just provide data. Essentially the form will count the users key presses and compare them to the data being entered (see example). Thereby figuring out if the user is a human entering data, or a bot.

Other considerations

There is quite a lot being written on the web regarding Turing tests including a paper by the W3C on the inaccessibility of CAPTCHA and even some truly inventive if not particularly useful variations.

Personally I think the best solution is a mix. Use an unobtrusive solution (such as the false form field method) and if that solution suggests a robot, then use a CAPTCHA, or the logic question variation above.

Technorati Tags: captcha, turing test, best practices

Well I suggested delaying sign in as long as possible. Actually let the user use the new feature and play around with it while they were on the site, and then add the value of memory only once they had decided that they liked the useful new feature™. This would require using cookies to store a bunch of information about how the user was using the useful new feature™, and working out a way to make sure they understood that without signing in their info would be lost at the end of their session. But if you consider the example of shopping carts we have been doing this kind of thing since the earliest days of e-commerce. I can vaguely remember one e-commerce site that required you to sign in to use the shopping cart, and I’m pretty sure they are no longer around.

I don’t think this kind of behaviours is always going to be appropriate, but whenever there is not a real security or other reason to require sign in, well maybe it is worth considering letting the user do as much as they can.

Of course the next step beyond this is the use of things like dojo storage wherein you could let a user do a whole bunch of work without having to login, and then synchronising once they finally do. Or alternatively a user could be using a tool, and signing in could be a signal to make a backup of their work (already saved locally) to the web applications web server. hmmm… just a thought anyway.