WebMonkey.com has an excerpt from Jakob Nielsen and Hoa Loranger's new book Prioritizing Web Usability, highlighting eight problems that still exist in web applications today: including:

• Links that don't change color when visited
• Breaking the back button
• Opening new browser windows
• Pop-up windows
• Design elements that look like advertisements
• Violating Web-wide conventions
• Vaporous content and empty hype
• Dense content and unscannable text

Oddly enough, Webmonkey itself violates some of these (and your Humble Editor makes no comment about where TSS violates them as well.) The authors make no mention of Ajax being a problem (unless perhaps implied in "Violating Web-wide conventions," although the section explaining this problem makes no specific mention of Ajax.) What do you think of the list? Seth Godin, in his "How to get traffic for your blog" and "Maybe you don't want traffic so badly?" entries makes the point that following the rules can yield boring sites, and users tend to prefer sites that aren't boring; where's the balance?

Seems that he got out of an oxygen chamber that had been sealed since 1997. Or maybe he just wanted to release a new book with recycled content. After all, he would have felt badly for throwing away all the precious data gathered during usability tests almost a decade ago. Anyway, I do not agree with some of the points that he makes. Looks like he treats web as hyperlinked collection of pages circa 1994, not as collection of pages *and* applications or even as all-uniting medium that will ultimately congregate web and desktop. For example, marking a link as "visited" does not make sense in a system that generates dynamic up-to-date information every time a user clicks a link. Same with the back button: if transaction has been committed, there is no point in rolling back the user interface since it would not correspond to server state. About new windows and pop-ups, I hate them as much as he does, but I understand that before Ajax gained acceptance this was the way to retain current page content while displaying some additional info like series of pictures. "Play Nice" matra is as meaningless and general as it can get. There are other signs of age, like he does not explain the benefits of semantic web development and proper use of CSS (yes, this affects usability); does not mention bad accessibility and less than stellar degradation of Javascript-based websites and how many of these sites do not even bother to check whether Javascript is enabled; does not explain liquid vs fixed design issues; does not critisize unscalable fonts, etc. Ajax is not mentioned as well. Maybe post-Navigator4 issues are disscused in the rest of the book, but the exerpt Webmonkey chose to publish makes one think that the whole book is just a facelifted reprint.

Same with the back button: if transaction has been committed, there is no point in rolling back the user interface since it would not correspond to server state.

You understand (or postulate) this. What Nielsen says is that it is a problem that you assume that your users share your feelings on this subject. I don't know if he has made any new surveys to validate his assumption. His statements match my experiences, however.

Same with the back button: if transaction has been committed, there is no point in rolling back the user interface since it would not correspond to server state.

A good explanation of why "Breaking the back button" is actually a requirement for Enterprise Web Applications. Valery&Galina's Blog

I too think hard-and-fast rules are usually counterproductive. Rather they should be looked at as trade-offs. For example, Zimbra breaks the back button, but the other functionality it gains is well worth it. And I find non-ad/non-survey pop-up windows to be just fine (so long as they are not a whole new browser window). I think it is interesting that some of the most cluttered sites on the internet (MSN, Yahoo, etc.) are also some of the most popular. I absolutely hate them myself (TSS is about as cluttered as I can stand), but it doesn't seem to affect their use too much.

I too think hard-and-fast rules are usually counterproductive.

Rather they should be looked at as trade-offs.

+1 big time

I too think hard-and-fast rules are usually counterproductive.

Rather they should be looked at as trade-offs.

+1 big time

This puts AJAX in context. A well-designed AJAX page can provide a very rich, interactive experience (making the big assuming that it is designed for usability) WHILE at the same time breaking several of the 'rules'. The guideline should be "you need to have a very good reason to break the usability rules". PJ Murray, CodeFutures Software Data Access Objects and Service Data Objects

I too think hard-and-fast rules are usually counterproductive.

Yes, that's why they're called guidelines, not rules. The understanding is that a guideline may apply in many, but not all, cases.

I actually agree with Neilson in so far as most users (i.e. people who do not read this forum and do not really understand what the implications of the back button are other than it makes them go backwards and forwards) really are simpletons and web sites/web applications must work as the user expects them to work from them to be a rich experience to which they will come back over and over again. For example, if a user clicks on the back button, they EXPECT to be able to get to the screen they had before. If the application becomes unpredictable due to this happening, users are going to get frustrated/leave. For me its not about the technology, more how the user expectation. Yes Ajax applications are cool, and yes we are more than a collection of hyperlinked pages. However, the end users have not really changed even though they have become more used to technology. (for example, how many people can program there video recorders even though the technology is 20 years old!!!)