About HFI   Certification   Tools   Services   Training   Free Resources   Media Room  
               
 Site MapUser Experience for a Better World   
Human Factors International Home
Free Resources
Bookmark and Share

Ask Eric: Questions & Answers

Each month Dr. Eric Schaffer answers selected questions on usable interface design. Recent Questions
Archived questions and answers about ...

Software Presentation and Visual Design Challenges

December 9, 2002 – submitted by Brett King of Chicago, IL

Question: I have a situation where after a user enters an ID into a form field and the field is blurred, a server hit is required as a validation to the freshly entered ID. We call this a "Trigger Field." The validation is necessary because other actions on the page depend upon this ID being valid. The problem is that blurring of the ID field can be caused by the clicking of buttons that seemingly are ignored when the server is accessed for the ID validation. So if the ID is invalid, then an error message is displayed and the user is forced to fix the ID in the trigger field (no problem here). But when the ID is valid, the user must again click the button for the intended action.

Eric's response: I am very uncomfortable with the idea of blurring or graying out fields. It means that the user can see it, but can't change it. But then what is the point? If it is display only, show it as display data (not in a text box). It is better to have the field shown and then provide an error message (or warning) if the user changes it. This is far better then leaving the user staring in frustration at a field that cannot be accessed.

In the case you describe I would consider putting the ID into a separate part of the taskflow. Maybe on the preceding screen. You are doing an I/O anyway, and that would avoid the problem entirely.

Top

November 26, 2002 – submitted by Erin Bradner of San Rafael, CA

Question: I'm having a usability problem interpreting Bob Bailey's November 2002 article on line length. All of the references in the article refer to line length in inches, but line length will change depending on the monitor size and resolution. I am very interested in these findings, how do I interpret them? What is the screen size and resolution. ALSO: a line of 12 point text in MS word IS NOT THE SAME length as a line of 12 point text in the same font in Photoshop although they print the same. If you discuss the font size, please tell me the presentation media/application you're using to measure.

Eric's response: Good point. I think it is better to use "characters per line" in gauging width of text. This solves the problems you have identified. For optimal reading speed use 100 character wide columns. For optimal user preference use 55 character wide columns.

Top

November 19, 2002 – submitted by John K. of Massachusetts

Question: Hi Eric. I am trying to document a legacy system (banking software) that has an inconsistent use of icons, labeled buttons and picture command buttons. Unfortunately, the trend for new development is using graphics for command buttons like a pencil point for "add" and a pencil eraser for "delete".

What is the best approach and why? I need some facts to back up my request for imposing standards on a very large system. Ultimately, I would like to recommend hirng a UI Designer who can assist with this early on, but that might be a tough sell, too.

Thanks.

Eric's response: Hi John. Using command buttons with descriptive terms works well. Adding icons to a command button will tend to improve search time by about 10-300ms. This is not a lot of time and can be easily offset by increased download times.

Using icons alone creates terrible comprehension problems. Tool-Tips can help. But the user ends up searching through icons waiting for tool tips. A sad situation. The only real advantage is that the icon alone takes less space.

In summary there are only two reasons for using icons. To save space for expert users (who will learn the meaning of the icons). And using labeled icons (added to command buttons) to support the theme or metaphor of the site.

Top

September 25, 2002 – submitted by Brian Melancon of Houston, TX

Question: What is the Web standard for table header and data alignment when the data types are mixed (text, numbers, date/time, etc.)?

Eric's response: With tabular data, left align the data, with the exception of numeric fields which are being compared (e.g., dollar amounts). These numbers should be right aligned. The headers should follow the alignment of the data in their column. There is one odd situation where the field label is long and the data is short. This can look odd, and might make one want to center the column.

Top

September 3, 2002 – submitted by John Pavlin of Toronto, ON

Question: In your experience, what are your thoughts on where to place a mandatory disclaimer page on an interactive pension modeling Web site being created for one of our clients.

Basically, all the employees will be allowed to project their pensions using their own data from a database. Because the calculations are estimates only, the user will need to understand and agree to a disclaimer to be able to play with their results. We are trying to figure out if we should do it in one of the following three areas:
1) Just after login page but before Home page.
2) After Home page but before seeing their Personal Data
3) After their Personal Data page but before seeing Assumptions page.

Our techie would like to get rid of it as soon as possible while our communication expert wants to place the disclaimer after the Personal Data page.

Which of the three do you prefer? Any other thoughts? If you know of any Web resources in regards to this matter, that would be beneficial as well.

Eric's response: We know from thousands of usability sessions that people do not read these disclaimers. There is in fact nothing you can DO to MAKE them read the disclaimers. So they must be adequately accessible from a legal viewpoint, but out of the way as much as possible. If the user is only logging in to use this modeling system the best solution is to include the disclaimer on the login page. This is "free" space anyway. And they only see it once. If this is not feasible put it up front where they will only see it once and it will not recur. It is also common to have the disclaimer itself in a scrollable list box. In this way the user can get to it (if they want to, but they don't) yet it does not take much space on the page.

It may also make sense to try to have a SINGLE sentence that summarizes the gist of the disclaimer if it is not all meaningless legal stuff. Something like: "You are only guessing what the stock market will do. Therefore, your results are just a guess". Put that at the top of the disclaimer area.

If the user must be presented with the disclaimer before reaching the calculator, then make the controls they use to agree/disagree indicate the consequences. Perhaps links like:

I AGREE

I DO NOT AGREE(calculator will not be available)

Top

August 14, 2002 – submitted by Clare De Cleene of Washington, DC

Question: Help! We are currently redesigning our Intranet site and are grappling with the issue of when acronyms should be used and when not. Sometimes we need to be aware of new employees who may not understand an acronym. Sometimes we think a specialist audience will be using some pages and they work better with acronyms. Are there any guidelines we should follow? Is it acceptable to use an acronym and then have the complete term display in a mouseover?

Eric's response: If an acronym will be a WORD for 90% of your user population you MUST use it. For example we would not abbreviate International Business Machines as INTER. IBM is a word for most users. There are even acronyms that many know, but they may not even recognize the words. For example TCP/IP is often understood, but the words are not known.

However, in general avoid using acronyms. They are non-mnemonic and people misinterpret them. If you decide to use one because most users know it and use it as shorthand, then define it the first time you use it. Use the conventional method please, with parentheses.

VOPAR (Vendor Oriented Payment and Reporting)

Top

August 6, 2002 – submitted by Scott McDowell of USA

Question: I'm trying to solve a debate in our office in regards to the most usable form layout based on actual user studies. The three layouts are:
Layout 1:

Address 1-----City
Address 2-------State-----Zip

Layout 2:

Address1----Address 2
City---------State------Zip

Layout 3:

Address 1
Address 2
City
State
Zip

Eric's response: Address format #3 is easier as it segregates the labels and allows the user to scan down a left justified set of data. However I would not use "Address 1" and "Address 2". The word 'address' refers to the whole thing. Also it clutters the display unnecessarily. So use...

Name: _____________________
Street: _____________________
_____________________
City: __________________
State: __
Zip: _______

If you are short on space use...

Name: ______________________
Street: ______________________
______________________
City: _______________ State: __ Zip: _______

Top

August 5, 2002 – submitted by Meny Lev-Sagie of Jerusalem, Israel

Question: We are trying to improve the way a medical decision support system presents its output, while allowing the user to perform his other tasks of using the system to document his actions, etc. Instead of modal pop-up dialogs that we sometimes use, suppose an alternative specific area would be used to display a list of suggested items. What options do you see for us to use a small screen area to show a considerable list (up to 20-30 items)? One idea we came up with is a running text banner window – like news headlines. How effective could that be? Is it better to rather hide the suggestions behind an indicator button?

Eric's response: A titled area of the screen is OK for monitoring secondary information. It is used extensively in the financial arena (for brokers and market makers who watch the market move on a screen with many tiled displays). It allows the users to maintain situational awareness. However, it is NOT good for critical alarms. The data is likely not to be noticed fast enough. It is also not good for supplemental task information (like online glossary or help) as it permanently takes screen real estate. For critical and time sensitive alarms use modal popups with auditory alarm. For supplemental task information use popups, which will go away and not take screen space.

You could put the 20-30 lines in micro-flyspec font and it will fit in a small tiled area. Unfortunately it will not be legible. You could scroll the tiled area. However at that point the data is not continuously displayed. Continuous display is probably the whole point of the exercise. Consider using a graphic presentation of the data. You might also consider textual aggregation of the information. So have 4-5 lines or numbers that indicate what is happening in the 20-30 lines.

Top

July 25, 2002 – submitted by Mark Bosco of Pittsburgh, PA

Question: We are working on redesigning our Address information section. Do you have any suggestions on how to handle zip codes that are used to populate city and state? Is it better to place the zip code before the city and state field? I really don't want to break our tab traversal rule and skip from the address 2 field over city and state to the zip code field.

Eric's response: You could try:

Name: _____________________
Street: _____________________
Zip Code: _______________  New York City, NY [wrong]

The "wrong" could be a button or link. It only shows up when you display the lookup city and state.

Top

June 14, 2002 – submitted by Emily Mueller of USA

Question: I am working with a team to design a customer-facing self-service Website, and a debate has come up regarding the terms "my" and "your". We will have several sections of the Website that are personalized, but we can't decide which pronoun to use (i.e. "My Account" or "Your Account"). Is there a standard that we should use or is it just a matter of personal preference? What is your preference?

Eric's response: In particular, the term 'My' has been quite overused in personalized Web sites. I have seen menus and headers that simply put 'My' at the head of each item. It is a silly waste of space. If you are going to have an optional personalized section you might use 'My' ONCE in the title of the section to indicate it is customized. I would use 'My' instead of 'Your' because 'My' is more widely used in that way and is also shorter.

I will also throw in a caution that few users are interested in filling out forms or doing other work to customize a site. It is better practice to design the site right the first time instead of hoping the users will do the design work for you. Personalization of this type is only likely to succeed in sites that are used EXTENSIVELY for a long period of time. Instead, use more subtle personalization that the user need not actively set. The best personalization is never noticed by the users.

Top

June 3, 2002 – submitted by Lynn Henry of Georgia, USA

Question: Is 800x600 is still the industry standard GUI for displays? Or has 1024x768 become the new config?

Eric's response: Unless you have a special user population I would indeed recommend developing shrink-wrap applications for 800x600, with elegant degradation to 640x480 and potentially good expansion to 1024x768.

Top

May 31, 2002 – submitted by Erwin Van Trier of Wylie, TX

Question: Is there a difference between login and sign in ? A lot of times you see links like "click here to sign in" and then the first data you need to provide is the login field.

I see login as an action to validate a userid and a password. A lot of times the term "login" is used for the userid or account.

I also would like to know what is the opposite of login. I found terms like "log off", "logout" and even "disconnect" (should be used in combination with connect).

Eric's response: There is a phenomenon in design I call 'Out Subtling Your User'. So many times designers come up with some feature that is very cool and subtle. The designer is excited about it. But the users never notice. There is no practical difference between Login and Sign In. If you make one up it will be your own private joke. The ONLY useful principle is to maintain consistency. So if you use "Sign In", then have the button say "Sign Out". If you say "Login", then have the button say "Logout".

Top

May 28, 2002 – submitted by Akanksha Malik of India

Question: Is it a good idea to have graphics on the left side of a page in a WBT and text on right even though you may want the user to read the text first and view the graphic later? The graphic could be interactive or might be representing an example explained in the text in a visual manner. The WBT is in English.

Eric's response: The idea that the images should be placed at the left was pushed by people looking at hemispheric dominance. The right brain is more attuned to processing images and due to the neural hookup of our eyes, the images to your left end up getting channeled to the right hemisphere. This has turned out not to matter.

Instead, lay out the page based on the FLOW of the users' task. People tend to start at the left and move right. If you want them to view or interact with the image first and then read the text for embellishment, put the image to the left. If you want the user to read the text first and then look at the image put the image at the right or below.

Top

May 27, 2002 – submitted by Hal Pawluk of United States

Question: Re: the 'Ask Eric' Question 'What is your opinion concerning "liquid" design? Any compelling reasons to adopt this to allow pages to scale to any resolution if you do not know your user's system?'

I like part of your answer to that: "Making the page resize to match the current screen size is a very good strategy."

However, you blow it in the implementation on your site and the "good example" you reference. In both cases, I have to scroll horizontally to see all the content because you don't use liquid pages.

You say: " We design for 800x600 resolution..." and while that may well be a good choice for your own software, it's not a good choice for a Web site.

I can see why you picked that number. Most monitors sold today are at least 800 x 600. Further, it you take a look at Web stats, you're likely to find that 95% of the visitors to any given site have monitors that are 800 px or wider. That's good data, but it's the wrong data for Web design.

The problem is that a large number of surfers do not maximize their browsers. Readability on the Web is worse than in print, so many surfers use a narrower window to get the line size down. The basic print design rule, by the way, is 1-1/2 to 2-1/2 English alphabets, or 39 to 65 characters per line.

You can see some sample browser and monitor stats here. This is a demo and they reset to zero at the end of the month, but currently they've got a fairly large sample size. Out of the measured 52,000+ surfers, over 90% have monitors 800 px or larger, as expected.

For Web design, though, they also provide stats on browser widths. In this case, 19% of the visitors have their browsers set to a width of 699 pixels or less. The measurements are for a gaming site, and gamers tend to maximize their browsers. On regular business sites, the number is usually quite a bit higher.

The bottom line is that 800 px is probably a good choice for software designs because you have a captive audience and can set the width to anything you want, but for the Web, you need to decide if you want to annoy and alienate 20% or more of your visitors.

Eric's response: Actually, the optimal reading speed is at 100 character length text lines, though people do like the 55 character wide text better.

It is absolutely good practice to design for 800x600 and then use liquid page design and other methods to avoid horizontal scrolling for viewers with lower resolution.

The HFI site is interesting. Note that our target users are people in the development community. They tend to use higher resolution displays. I think we may be being a bit conservative with our 800x600 resolution. The general point is that we need to examine assumptions like resolution for each user population we design for.

Top

May 17, 2002 – submitted by David Bryant of Lawrenceville, GA

Question: I am working on an implementation that will be converting data from an old mainframe application. Is there a standard for uppercase vs. lowercase for customer facing information? I know that the post office prefers uppercase but what about the other data that will be presented to customers in bills or on self-service Web apps.

Eric's response: In running text, upper case presentation slows the user's reading speed by 14-20%. In short segments of text, under 5 words, there seems to be little measurable effect. But in any case I would avoid all upper case text with the exception of codes (e.g., "AT23".)

Top

May 10, 2002 – submitted by Anonymous

Question: Is there a standard or best practice when it comes to formatting phone numbers for Web pages?

Eric's response: In the USA and Canada, if you are displaying phone numbers use:

Home: (212) 555-1212

If the user is entering a phone number in a form, use a single field entry. Do NOT break the number into three separate fields. Accept and remove blanks and other separators.

International phone numbers are a different issue. If it is critical, you need to know the specific country and then format for that country. Otherwise leave a large field for the user to enter what they want.

Top

April 2 , 2002 – submitted by anonymous

Question: What is your opinion concerning "liquid" design? Any compelling reasons to adopt this to allow pages to scale to any resolution if you do not know your user's system?

Eric's response: We design for 800x600 resoultion in most public applications right now. But certainly people view the pages in other resolutions. Making the page resize to match the current screen size is a very good strategy. However, it needs to resize within bounds. Making it VERY narow or VERY wide makes the page harder to use. Research shows optimal reading speed at about 100 characters per line, and user preference is 55 characters per line.

There are many coding tricks which make the dynamic resizing of your pages work better. Here is a a good example prepared by Jeff Lees, a project director at HFI.

Top

April 1 , 2002 – submitted by Len Conte of Northboro, MA

Question: I took your Web design course in 1997 and your UI design course years before ...I still refer to those notes :) I am interested in research and guidelines for selecting a good set of colors (6-10) from a "Websafe" palette for the display of multiple lines on line charts, multiple bars on bar charts, and stacked bars on bar charts. (We have the ability to code redundant cues as symbols on the line charts... we don't have a way to pattern the bar charts or stacked bar charts)... I'd also be interested in research and color recommendations for chart background colors and gridlines on these graphs.

Eric's response: Thanks Len. We love knowing that our courses make a difference. There is a rich literature on color coding. First it gets hard to have more then 6 color codes that people will recognize. Your suggestion of 10 is pushing the envelope.

When selecting colors you have two issues. First, the user must be able to discriminate between the colors. This is an issue when both colors appear together. But it far tougher if the colors are NOT show together. We are pretty good at comparing two colors shown together. We are much worse at independent identification. To make it easy to discriminate between colors be sure to pick ones that differ widely in hue (frequency), brightness (intensity), and saturation (how much of a pastel it is). You can also pick colors that are less likely to create trouble for color deficient users.

The second issue is the semantic content of the codes. Colors have associations. For example, we use green for 'Go' or 'OK'. We need to pick colors with the right associations. This can make them easier to interpret and remember.

Top

March 28, 2002 – submitted by Cynthia Spellman of USA

Question: I am interested in research concerning font and background colors on Web sites. I have several sources that say to pay attention to brightness and contrast, but I am specifically concerned with white lettering on a black background. Is this in general OK? Preferred? Good only for labels and large, sparse type but not so good for paragraphs of text? Can you please point me to any references?

Eric's response: The research is clear that fastest reading occurs with black text on a white background. Use this combination unless you have other objectives. In addition to being faster to read, it is more resistent to glare (the large white area overposes some glare).

Early research on early CRTs showed that dark background with light characters are less fatiguing. This does not appear to be true in the current technologies. Therefore, the only reason for white characters on a black background is stylistic issues. If it is important for aesthetics, to support a brand, or as a part of a theme... use it. It will create negligible performance effects.

Top

March 5, 2002 – submitted by Shane Jewitt of Rochester, NY

Question: Hi Eric, I attended one of your seminars last year "How to Design Effective Graphical User Interfaces" as part of a retraining phase for my job. My background is in graphic design, specifically printed collateral and now I'm dealing with screen graphics. My question for you deals with contrast deficiency. You recommend on your Web site to give text and its background a 90% contrast difference between foreground and background in order to read text clearly. How does one go about measuring this?

Eric's response: Hi Shane. You need a 90% difference between text and background in order to avoid degradation in legibility. This can be measured by a light meter in lumens. However, be aware that this rule holds for black and white only. Color is complicated by the differential sensitivity of the eye to colors. So red and blue look relatively dark. Yellow and green appear brighter. So pure blue text on a black background is bad because the blue will appear dark and you will have insufficient contrast. Measuring this gets complicated and interacts with user types (the lens of the eye yellows with age accentuating this effect). Bottom line, LOOK at the design you create and make sure you have plenty of difference between character and background brightness.

Top

February 25, 2002 – submitted by Christina Formentini of USA

Question: What screen resolution should I be designing for for Intranet applications? Recently I have heard many developers adopting 1024 x 768, however I can find no data to support this as a new standard (replacing 800 x 600).

Eric's response: In the public Web the numbers still support an 800 x 600 resolution. In an Intranet it will depend on your in-house configuration and equipment. It is a matter of looking at those numbers.

Top

January 16, 2002 – submitted by Subrata Basu of Toronto, ON

Question: Hi Eric – I attended one of your seminars last year and that really helped me to understand the importance of usability. Currently we are facing such an usability issue. We publish on our Intranet a lot of "Policies and Procedures" type of html documents for our organisation. These documents are read very often by our employees but they mostly look for changes made in the document. The software we use keeps track of the changes made in the document and marks these changes with an "Underline" and makes it "Green". With this marking it is easy to find out small changes made in a paragraph but when a couple of new pages are added in the document, it becomes difficult to read. Some of the text in our documents are already in "Red" so we cannot use that to mark these changes. Do you have any suggestion?

Eric's response: Try a vertical mark down the left margin of the page to indicate the area changed. This does not show the specific words changed, but it will eliminate the clutter problem. Also, the green unerline may look too much like a link. If you need to have the specific words indicated, consider a switch that allows the user to reveal the underlines to indicate the changes made.

Top

© 1996-2012 Human Factors International, Inc. All rights reserved  |  Privacy Policy  |   rss feed biber hapıbiber hapı