Site MapUser Experience for a Better World Site MapUser Experience for a Better World 
| Each month Dr. Eric Schaffer answers selected questions on usable interface design. | Recent Questions |
| Archived questions and answers about ... |
|---|
|
December 4, 2001 – submitted by Abdo Moh of Malaysia |
|
Question: How are the issues of long-term memory, short-term memory, and user expertise relevant for the understanding and use of the URL of a Web site? |
Eric's response: The design of a URL is a significant usability engineering task. It is mostly about optimizing the ability of users to work with that code. In many cases the user must remember the URL for a long period. For example the URL may be presented in a radio advertisement. In this case we need to make the URL easy to place in long-term memory. That means associating the URL with a well known event or object. It means keeping the information content low (www.qt349mw.com is hard). It means avoiding unexpected items (e.g., www.humanfactors.biz is hard because users expect ".com"). There can be mnemonic tricks (just press the bottom three keys on your keyboard - www.zxc.com). In other cases you are concerned about the ability to transpose the URL from a direct source. This might be entering it from a paper bill. In this case short-term memory is a major concern. If it does not fit in short-term memory, then the user must look back and forth to enter the URL (www.UTCS.com/Tirade/Giveup/Tank). Errors become likely. There are many other issues in selecting a URL besides memory. The likelihood of touch typing errors is a concern. Any time you have a user press the same key twice in a row it will increase the likelihood of typing errors. There are issues of visual description. Showing "L" and "1," or "O" and "0" in a URL will increase confusion errors. There are issues of brand perception and emotional response. URL design can be a significant source of human engineering challenge. |
| Top | |
October 31, 2001 – submitted by Satya Vallepalli of Harrison, NJ |
|
Question: I am student working in the field of HCI and usability. I am currently doing a project where we are developing a user friendly Webboard (forum) to our university. There are a lot of Webboards in various designs and structures on the web. Some of them use no frames and this is how it works. After entering the id and password we are taken to a page with all the main groups from where you can go inside a particular group. There you have the original postings or topics. Once we click on the topic all the replies to them are shown in the next page one after the other. So there is no hierarchical structure (in the sense). Some are in the fashion of using the hierarchy structure (+ and - signs) using frames (for example the way newsgroups are shown in Microsoft Outlook). I know using frames is very tricky and it is not recommended in your earlier reply to a question on frames. Is using frames justified for this kind of application or is the no-frame solution the best way of doing it. |
Eric's response: First of all frames can be a GOOD solution to very specific types of situations. These are basically situations where there is a set of material (for example a table of contents) that needs to be held constant where another body of material (for example a display of text) is scrolled. This is a classic good use of frames. Regarding your Webboard design it is hard to comment without pictures. You must consider the limitation of scrolling and the number of postings you expect to get. Providing a menu of topics and then a single page with all postings on the topic is appropriate only when there are a limited number of postings (say usually 3-5 scrolls worth). If there are more postings you may need another level of hierarchy; where the topic leads to a selection of threads. Then provide a list of all postings on the thread. The problem with using frames with these situations is that once you have a left Web frame wide enough to show the topics, it may take up so much space that the user must now scroll the right frame too much. These challenges define the chess game of good UI design. |
| Top | |
September 17, 2001 – submitted by Ann-Marie Lind of Lexington, MA |
|
Question:In developing a decision-aiding system, I know that it is important to provide information to the user on the confidence of the automation/algorithms in the information provided. For example, in a system that predicts earthquakes, microbursts, etc., the user needs to know the confidence level of the prediction so that he/she can decide whether to take action or decide that it is probably a false alarm. So my question is not regarding whether or not to provide confidence level, but appropriate ways of providing confidence level given certain applications. Have there been some studies on various types of presentation of confidence level and their effectiveness? Can you point me toward some literature on the subject? Thank you. |
Eric's response: The best display depends on the user and situation. In general, there is a substantial value to providing an indicator that is a direct reflection of the actual measurements. For example, the Stormscope weather avoidance systems tells pilots each time it detects a lightning strike. If there are only a couple of strikes shown, there is not much to worry about. If there are more than three, there is a good chance of a thunderstorm in that area. Moving the display to a greater abstraction leaves the user with less information to make a decision and a simplified task. An example is the use of "idiot lights" on cars. The user is no longer told the actual oil temperature. Instead a light comes on when temperature (or some other measure) exceeds a threshold. The task is simple... get to a service station. But there is no chance to take more sophisticated action yourself (we used to turn the car's heat on to cool the engine a bit). Once you decide the specific content you want to share with the user there are lots of methods of display. The human factors literature is full of criteria for using analog displays vs digital; bar charts vs. line charts, etc. |
| Top | |
August 2, 2001 – submitted by Vana Trudeau of United States |
|
Question: I'd like to know your opinion/thoughts on using frames on a B2B site. |
Eric's response: There is very little difference on the issue of frames between B2C, B2B, and Intranet. There are a set of problems with frames. They can disrupt search engine categorization. They break up the mental model of 'web pages. In older browsers they can be very awkward to manage (functions like BACK and PRINT do not accommodate them well). But, as is true for most interface design facilities, there is a good role for frames. Frames are worthwhile when there is a need to break the scrolling connection between elements on the page. The most classical case is a long scrolling Table of Contents (TOC) or Tree View on the right, combined with scrolling material in the body of the page. If the user is reading pages from the TOC sequentially, think of what happens without frames. The user pick the first item. The material displays and the user reads it scrolling down. But by scrolling down the user has moved the TOC too. The desired item has now scrolled off the page and the user must scroll back to select the next item. That type of interaction is very awkward. So for that type of page, given the improvement in frame management by current browsers, you will generally use frames. |
| Top | |
May 9, 2001 – submitted by Henry Eakland of United States |
|
Question: Currently our pages are designed for an 800x600 page resolution, and the display at this size is just fine. But when the user prints this page out from MS Explorer (especially), the right edge of the text gets cut off. What are some of the best practices for dealing with this printing problem? Do we have to provide a "Print This Page" option for all of our product pages, for example? |
Eric's response: To make printing easy, use a "Print This Page" link. More details |
| Top | |
March 28, 2001 – submitted by Earl Hurd of Springfield, NJ |
|
Question: To what degree do clients/users comprehend written materials differently than on-screen materials? |
Eric's response: In the early days of CRTs the quality of the display was so poor that users generally read about 25% slower online. That is no longer true. With improvements in format, resolution, and anti aliasing, the reading speed is quite close. However, users do seem to change mode in online environments. The Web appears to be the technological equivalent of attention-deficit disorder. So expect users to scan rather then read. If your design has lots of paragraphs of text, that is usually a mistake. Use less. Use bullet lists. Use decision charts. Use graphics. Legibility has become less of an issue, then, but activity style is still a major factor. |
| Top | |
March 12, 2001 – submitted by Erik Racho of United States |
|
Question: What is the preferred Font Type/Size for viewing programming code? |
Eric's response: A simple way to calculate the ideal character size is:
For example, if you're sitting 24 inches from the screen, the character height is about an eighth of an inch high (1/200ths of 24 inches). That corresponds to about a 10-point type size in most fonts. (There are 72 points per inch.) Here's the scientific background for the above:
|
| Top | |
| © 1996-2012 Human Factors International, Inc. All rights reserved | Privacy Policy | | 
biber hapıbiber hapı
|
