LarryRoth.net

Just my thoughts

Tagging vs. Categorization

April 26, 2006

Henning von Vogelsang has written what I felt is a great article about The difference between categories and tags. Besides being a very concise comparison of tags and categories and when to use each, he has a fantastic real world example.

From the article:

Take a simple walk to a clothing store. Maybe you are entering the store with the idea to buy a pair of jeans. Chances are, you will leave the store with a pair of jeans, and a jacket, some socks, briefs, or two plain t-shirts. Why? Because you were surrounded with offerings that were not sorted by category but by association. You were allured by visual suggestions. That is the same like tags can work on a web site. They can draw you into something. Not because you were looking for it, but maybe out of curiosity.

What I like about this section of the article is that it helps me to overcome the idea that tags are simply a flat organizational structure for classifying information. They are indeed that, but when used appropriately can be much more. I must admit, I am often hung up on the fact that tagging can only be as good as the tagger, but isn't the same true about an information architecture? I really like the notion of being inspired by a tag. Can you say the same thing about a categorized list? I can see a lot of situations where tags can not only help to create new information relationships, but also tie in an emotional attraction that quite frankly would be hard to create with an organized, categorized list.

The trick is to find the right tags, which are most universally understood and evoke the correct emotional response. This really makes me wonder where they can fit in the corporate world. For instance, should IBM tag their product page for Websphere application server community edition? Maybe so. It could be a great way to expose people to a lot more of what they have to offer. Having the tag community edition would be one-click to all of their free product offerings. Or java, leading to their expansive knowledge and offerings in that arena. What a great way to make a sticky site!


Tagging vs. Categorization

When do you test?

April 21, 2006

When planning out a project schedule for a new or redesigned Web site, one common question is when do we run usability tests? I don't believe there is an one easy answer to that question.

My preferred answer is: If there is budget for it, test early and often! Setting up a test plan early on will allow you to not only incrementally test and tweak your designs, but if you are engaged in a site redesign you can also use the same repeatable testing procedures to get a baseline from the existing site. Not only will you be able to the new design is solid, but also you can show improvement over the existing site.

Testing could occur as follows:

  • Baseline test of the existing site
  • Low fidelity test of approved wireframes
  • Test of graphic comps
  • Test of completed site on staging server
In addition, the same test plan should be used after the site is live, to keep an eye on the site as it progresses.

Of course, the biggest concern is the cost—in time and money—of running so many test. Rather than give the overused response of how testing up front saves more money down the road, I prefer to point out that setting up the test and creating the report is the lion's share of the time required, not recruitment and subsequent execution. It's amazing how efficiently you can write up a report for an existing test the second time around.

Another issue is that clients will often be concerned that poor test results will affect the project schedule. Instead of jumping into a discussion about what is most important, I have found it helpful to let them know that the results of testing are usually minor tweaks that can have major impacts. It is usually the case that they feel poor results mean we will mean starting all over.

Despite the fact that I believe in testing early and often, at BrandLogic we have only every had one client that would allow us to engage in this level of testing. And it's interesting, they were ranked by an independent third party to be number 1 in their vertical market with regards to usability. hmmm.....


When do you test?

Why good design is so important

February 2, 2006

An interesting study by researchers at the Carleton University in Ontario provides evidence that Web users form an opinion about a Web site in some cases before they even discern the contents of the page. Researchers flashed pictures of Web sites for as little as 50 milliseconds, and test subjects were still able to form a judgment about what they saw. [1]

From the article:

"...the strong impact of the visual appeal of the site seemed to draw attention away from usability problems. This suggests that aesthetics, or visual appeal, factors may be detected first and that these could influence how users judge subsequent experience.... Hence, even if a website is highly usable and provides very useful information presented in a logical arrangement, this may fail to impress a user whose first impression of the site was negative." - (Lindgaard 2006)

What this study re-enforces is that even though usability is an important aspect of any Web design, you can not overlook the emotional response to good design. Your users are going to make decisions about your site, based on their visual impressions—and this study seems to show that this occurs a lot quicker that one might have imagined.

Make sure you do a usability study, add some useful and cool functionality, and provide top notch content, but just don't forget skimp on the design.

[1] First Impressions Count in Website Design, http://www.websiteoptimization.com/speed/tweak/blink/ , posted January 17, 2006, viewed February 2, 2006
Why good design is so important

Does font size and color affect usability?

February 1, 2006

If I were asked what the single most argued point of Web design is at BrandLogic, it would unequivocally be font size and color. This is an argument that occurs both internally between graphics designers and Web strategists, and externally when we present our concepts to clients. Everyone seems to have a thought about what the ideal font size/color is, and it is always the one item about which they are not too shy to comment. Inevitably, reference is made to some unknown study that claims 12pt black type (or some other combination) is ideal. And what is the number one rational for their point of view? Legibility/usability.

Please allow me to clear up one misconception. Legibility and usability do share a relationship, specifically for a user to find a block of text usable it must be legible. However, this relationship does not mean that legibility is usability. Usability—with regard to the Web—emcompasses the Web site as a whole. There are many things to consider. Audience: who are your users? There is also content: Is it appropriate for the audience? is it clear, on brand, on point? There are user interface elements such as labeling, focal points, and action feedback. And of course there is graphic design, which at a minimal encompasses layout, color, graphics, line length, page size, and yes, typography.

Typography itself is an art form that is much more than font size and color. Thanks to CSS we Web designers are finally learning this. I remember an early lesson when I was a seasoned Web developer, but naive to design (and I probably still am). A very brilliant, and thankfully patient designer was rebuking my challenges that his chosen font size and font color made for illegible content. To my astonishment, he did not disagree with me, but simply pointed out that we had only applied two attribute of this typography specs. He gave us a quick lesson on typography, probably one-one thousandth of what he knew on the subject, and asked us to adjust some other attributes, specifically leading and kerning (letter-spacing and line-height). To my astonishment, those simple changes drastically affected not only the legibility, but also the feel of the page as a whole. The page felt lighter, my eyes could track the text much easier, it was a more enjoyable experience.

Still, even with more granular, and properly applied control over fonts, concerns about size and color still come up. Why is font size such a contentious issue? First off, it's an easy target because size and color preference are just that—preferences, and that makes them by nature very subjective. I contend that the major problem is there are no hard and fast rules for what color or size a font should be. It is very audience and design dependant. To some degree, the design will drive the decision of font size and color, but in my humble opinion, a good designer will let the needs of the intended audience drive the design.

Industry experts have weighed in and given us their opinions. Leading Web usability expert Jacob Nielsen once wrote: "Tiny text tyrannizes users by dramatically reducing task throughput." [1] A bold statement, and not without truth! And in the same article, he goes on to make his recommendations—both to browser makers and Web designers. But the rules are not concrete, or at least I do not view them that way.

It boils down to this: font size and color alone can affect usability, but you must examine the Web site as a whole. Conduct a usability test on the site, and consider these guidelines with regard to font size and color:

  • Ask the users to read a page.
  • Refrain from asking them subjective questions like, what do you think of the font size?
  • Instead, ask them questions about the content and while you are noting their comprehension, see if they offer opinions about the legibility.
  • At the end of the test, ask what they think of the design of the page as a whole.
  • At the end of the test, if they haven't already offered the information, ask what they thought about the legibility.

Don't let a good design be altered by a subjective opinion, but also, don't ignore the empirical evidence if it shows your users are having trouble reading your site. If your time or budget doesn't allow for a usability test, then point to other sites you have done that use a similar size font and were also tested. Also, look for popular, well-respected Web sites in the same industry that cater to the same audience. Perhaps they use a similar size and/or color font. This of course goes against my statement that design must be treated as a whole, and these arguments are not concrete, but maybe can provide some supporting evidence that your font choices are OK. Your client wants reassurance that they are not making a mistake.

Too often I have seen the comments of a few "squeaky wheels" push a client into decision that has a much larger, negative impact. Remember, you need to help educate your client about design and usability, just like someone most likely helped to educate you.

[1] Let Users Control Font Size, Jakob Nielsen, http://www.useit.com/alertbox/20020819.html, posted August 19, 2002, viewed February 1, 2006
Does font size and color affect usability?

Sun Java Studio Creator 2.0 released

January 31, 2006

The latest release of Creator, which is Sun's answer to rapid application development for the Web, has just been released. If you don't know, Creator's primary focus is creating apps using JSF—Java Server Faces[1]. Creator contains a neat data-binding model that allows you to bind JSF components to just about any data source. Based on NetBeans, Creator contains many easy to use and quality team tools, inspectors, and editors. You can find from information about the release here: (Web site)

I tried out the first release (I believe it was a pre-release) and wasn't enticed enough to devote time into figuring out how it worked. Actually, I tried to import some existing JSP projects, which it turns out wasn't the best approach. Firstly, the designer wasn't too kind to my CSS and secondly even though JSF is based on JSP, it's a completely different way of programming an app. I was thoroughly lost.

This time around, I started with a blank project, added an existing data source, and within a couple minutes I had created a Web page that contained an table bound to a table in the data source. All of this without writing one line of code! Of course, the page didn't respond to any user events or do anything particularly interesting, but it was much quicker than if I had set up a JSP page that did the same thing.

My goal is to spend some time this coming weekend exploring the tool in more depth and writing about my experience. I have not used JSF at all, and I think my background with JSP and Microsoft Visual Basic should provide an interesting reference point. For all it's short comings, Visual Basic was a fantastic tool for creating GUIs and I am anxious to see if Creator can provide that same GUI building experience with the power of Java and J2EE under the hood. I am going to try to stumble through without trying any of the tutorials or without reading the documentation.

I should mention that The Java Posse (Web site) is a fantastic podcast for anyone interested in Java. They have some great interviews, lots of good news, tons of practical experience, and they really get you psyched about using Java.

And check back next week if you are interested in learning about my experience!

[1]=Information about Java Server Faces Technology at Sun.com

Sun Java Studio Creator 2.0 released