I work for a very large company and it would be impractical for the visual designers to code all their work. At some point it makes sense to hand the coding responsibility off. I do think their needs to be a close relationship between the designer and the person who is doing the coding. Also the designer should get code approval.

In the end I do believe you will make a better product if the designer codes his/her own work. I love coding in live templates where you can see different instances of the data you are laying out. Makes the end product much more refined.

Like every thing their is not one answer and 37 Signals tends to write articles from there narrow perspective. So take there opinions with a grain of salt.

What say you?

1. Eliminate Redundancy

In the old Windows vs. Mac OS debate, most of the time people talk about security, but my biggest pet peeve of the Windows platforms is its overwhelming myriad of ways to do the same thing. With all user interfaces there will be a learning curve, but don’t use that as an excuse to add duplicate functionality.

Be stingy as it relates to adding top-level navigational elements, and make sure adding them is the right choice. A higher-level grouping may be what is needed. I recently came across a book called The Paradox of Choice; it talks about our inability to deal with an abundance of choices. An excess of choice has a paralyzing effect, so keep the UI from overwhelming the users. Make sure everything that’s included is absolutely needed to accomplish the goals of the UI.

2. Nomenclature & Contextualization

A user should have a reasonable understanding of what they will find underneath top-level navigation. Name your navigation exactly what it is and try not to be cute when naming different areas. When developing navigation and content areas, group like items together. This will add cohesion to areas and straighten the users understanding of the content.

Contextualize navigation and certain content elements to limit what the user has to visually take in at any one time. Content having the appropriate context and related information can do a lot to add additional clicks to supporting content.

3. Consistency

Being consistent in placement of navigation is a must. Some of the worst UI’s are ones where navigation moves around the page. Iconography and color can do a lot to improve the UI. Stick with your design conventions and try not to have many variations.

Buttons and typography should have a consistent feel and any deviations should point out special situations to the user. If you have a case where there is a slight change from page to page, like the case of an error message on a form, make sure the user notices the change. This can be extremely frustrating to users who don’t notice the page refresh and the page isn’t noticeably changed.

4. Beware of the Visual Arms Race

When designing a site or application you will usually have input from a client or business interests. It’s important to translate their request as it relates to the UI. A lot of times there will be multiple interested parties with separate goals competing for UI importance. It’s vital to understand this; it happens so often I coined the phrase “visual arms race” to describe it.

A visual arms race is when multiple stake-holders compete for prominence in the UI. It tends to skew the UI in some funky directions with the danger of losing focus of the core goal. Remind them what the core goals of the UI are and make sure their suggestions are in line with them. I’ve had many UI’s fall apart because of competing interests fighting to be the biggest and boldest.

There will also be a tendency for people to provide solutions to UI problems — try to get them to articulate the problem first. Many times their suggestion will be to add something to the UI, but this may not always be the best solution. If you first understand the problem, you may be able to come up with a better solution.

5. Testing

Because we are so involved in developing the UI, it’s hard to look at it as if you were a new user. That’s where testing comes in. For most projects I don’t think you have to do extensive user testing, but for projects that have sophisticated UI’s, testing can eliminate a lot of customer service question and also save money.

For smaller projects, taking simple screen mockups and asking someone new to the UI what the primary and secondary goals of each screen is can give you a good idea if the UI is working. For larger, more complex projects, having a formal user test or focus group can also be helpful.

In Conclusion

None of these suggestions are ground breaking, but they are things that I keep in mind when developing a new user interface. You can’t go wrong with just trying to keep it as simple as possible.

As for a crisis in web standards, there is none. Yes, there hasn’t been a ton of movement on CSS 3 and HTML 5, but thats not a crisis. More and more sites every day use and follow web standards and it’s been an amazing shift in the last 4 to 5 years. Most of the talk of a crisis is related to having new tools for web designers. Because these are voluntary standards and are not owned by one single company that can force compliance, progress will be slow.

Lets not forget how bad it was back in 1999 with the browser wars wreaking havoc on the web design profession. The whole web standards movement is mostly responsible for the growth of Firefox and in helping Microsoft get off their asses and make IE 7. Lets not whine too much, because it could be a lot worse.