I had a colleague swing by my office and ask to settle an argument for him.  They were building a new site and was going back-and-forth with his boss trying to convince his boss that they should open all links in new windows.  His reasoning was that opening links in new windows would help users remain engaged in their page.

"So, if they don't like the link they just opened, they could just close the window and then, bam, they're back on our page," he said.

My first inclination was to say, "Uh, they can use the back button, right?"

Lachlan Hunt, a "an idealistic, young web developer" writes

Basically, the author of a website should not have any control over the users application, in any way whatsoever. It has been known for long time that opening new windows confuses users. Not only that, it affects users in the following ways:
-- A user has to close the window, or at least switch windows, to get back to where they were ...
-- The back button does not work ...
-- It clutters the desktop and taskbar.

Now, I'm not saying that Lachlan is the de facto *expert* on the issue, but I tend to agree.  Generally, my thinking is this: users can generally expect when they click a link the new page will open in the same window.  Having any other interaction than that should be very carefully considered.  If a user *wants* to have the link open in a new window or tab, it's very, very easy for them to initiate that by simply pressing Shift or CTRL.  By keeping the interaction simple, the user is empowered, and that's a good thing.

I think, however, there are instances where we actually have to consider the action that is appropriate, and I, personally, think that there might be a time where opening a new window is the right way to go for the project (gasp!)  For instance, after talking a bit further with my colleague asking about the new-window issue, I found that one of the reasons he wanted to open a new window was to save the state the previous page was in.  Specifically, they had a collapsible panel that utilized jQuery to expand and contract on user-action and acted as navigation.  In this instance, clicking a link and then using the back button to navigate back would clear the state the panels were in and the user would have to "find their place" again.  Yes, of course there are technological ways to save states, but this project had neither the time or resources to implement that.  They considered what they had pretty advanced, and had to work with it.

As well, one concern with this particular project was that some of the links pointed to very small bits of content that helped define terms, provided contextual information, or linked to a larger image.  It was their thought that they could control the size of these windows to fit the content.  They didn't want to a blank page with a photo in the left-hand corner or a little bit of text at the top of an otherwise white canvas.  I tended to agree.

My solution?  Use a little jQuery to create rollover states for the items with simple contextual information and "lightbox" the images.  By using these interaction-models, they users stay on the page they were on, therefore retaining the state of the jQuery panels, and the problem of the blank canvas is solved.

Sometimes it's hard to convince other stakeholders of your reasoning, and even more difficult at times to convince others to spend budget and time to implement the right solution.  There are times you have to compromise.  However, by thinking through the problem and trying to determine what the user expects, what they're trying to do, and they many tools we have at our disposal, it's our job as designers and developers to be advocates of good user-experience and educate those we work with when building sites on the reasoning and process behind the design and development decisions we're making.

And on other side of the coin, we as designers and developers have to listen to our clients.  By listening to my colleague and understanding the full scope of this small problem he was having, I was able to provide him with informed counsel.  I could explain my reasoning and make him feel confident when presenting his thoughts on the issue to his boss.  In the end they used a hybrid approach, spawning a new window for a very certain set of links, and using Glimmer to produce rollover tooltips for some of the content they were planning on spawning a resized window.

I guess the long and short of it can be summed up in a few bullets:

• Generally, don't spawn new windows off your links, but don't be dogmatic.
• Listen to your "client," whether external or colleague
• Present them with your reasoned thinking and process for arriving at the advice your giving them.  Educate them on why.
• There is always more than one way to solve a problem
• Be a nice person.

So, what do you think?  Is there a "right" solution for particular scenarios?  How do you handle convincing your clients of doing things the "right" way?  Let us know, and be sure to follow us @mixonline on Twitter.