The salient difference between the “Genius Bar” and the proposed “Ninja” forum is that the “geniuses” are being paid. Almost two years ago, the WordPress hackers list discussed the issue of getting more code-savvy folks to help in the forums, and not much has changed (although there is an “advanced” section of the forums now). As some of the hackers pointed out, expert coders often find the “noise ratio” frustrating. But even assuming the signal-to-noise issue is eliminated, what will make the “ninja” forum encourage more expert help than does the current advanced forum? I’m looking forward to seeing what Mullenweg has in mind.

To summarize, with my comments in parentheses:
The first question asked about database abstraction in general and SQLite in particular. Mullenweg’s answer was that database abstraction in general was probably not appropriate for WordPress because it’s too slow, but SQLite might have some potential.

(Mullenweg seemed somewhat vague about how SQLite could or even whether it should be used with WordPress, but having followed past discussions among WordPress developers about porting WordPress to PostgreSQL, I think that the core developers are willing to abstract to the extent that it makes plugging in other databases possible, so long as it doesn’t cause performance to suffer.)

The second question was why the new WordPress dashboard, which debuted in version 2.5, split up so many of the menu items. Mullenweg mentioned two main reasons: 1) In an attempt to make navigation more intuitive for users, verbs appear on the left and nouns on the right. The idea is that the verb menu items are what most users want to do; the nouns (such as Settings and Plugins) are used less often and usually just by site administrators. 2) The dashboard was split up to allow plugins to put menu items in more places.

(Point 2 touches on a recent wp-hackers discussion, in which some argued that plugin options should always go under the “plugins” menu, and that this was the position of Mullenweg himself. The answer to this question seems to refute the latter notion. And the former notion is refuted implicitly for usability reasons, as some had already pointed out.)

The third question asked what WordPress was going to do to help users who aren’t technically savvy. Mullenweg had a three-pronged answer: 1) WordPress.com helps beginners get used to the WordPress interface, without actually having to set up the blog themselves. (Mullenweg made an interesting point that WP.com is geared toward users on both ends of the spectrum: beginners and those who need lots of resources, like CNN). 2) Mullenweg wants WordPress to be a “platform,” by which he seems to mean that like Firefox it should be extensible and able to upgrade itself. 3) Mullenweg said that he wants to create a WordPress ninja forum, in which WordPress experts can help each other out for free (I’m not sure how that differs from the WordPress.org support forum).

The fourth question was whether there were plans to include caching in core, to which Mullenweg replied with an emphatic “sì!” He mentioned that there are two related Google Summer of Code projects this summer and that he recommends using the WP-Super-Cache plugin.

The fifth question was whether WordPress would include a way to backup everything, to which he replied “no.” He said that it would probably not be possible for technical reasons and that it was really the job of the site’s host to provide such an option.

The final question had to do with whether WordPress would be incorporating SEO features, to which Mullenweg suggested that it was more profitable in the long run to focus on attracting human visitors rather than trying to rig Google pagerank.

Update May 12: Another brief video interview with Mullenweg.

More interviews.

More WordCamps are coming up soon, including one in Milan this weekend.

His comments came in response to criticism on Khaled Abou Alfa’s Broken Kode blog of the still inchoate WordPress admin theme redesign.

Apple’s “human interface guidelines” are highly regarded, so WordPress would certainly benefit from anything similar. Some of the material for the WordPress guide will probably come from the same research showcased in a talk about usability Happy Cog member Liz Danzico gave at WordCamp 2007. Unfortunately, I wasn’t able to attend WordCamp, but various blog notes and slides from her talk give me the gist, and the Happy Cog-inspired changes appearing in development WordPress provide further clues. Here are some points I expect to see:

• Make things easy to find. Apparently this means a navigational structure in which one uses verbs to find nouns. Here’s an example:
  
  
  The first image is from the new admin rework, still in progress. The second is the current design. Notice how in the first, newer, design, there is more of an emphasis on verbs in the top-level menu: “Design” has replaced “Presentation,” and though not visible in the screenshot, “Settings” has replaced “Options.” In the sub-menu, however, there is an emphasis on the objects being managed: instead of “Write Post” the new design just has “Post.”
• Be consistent. Users don’t like surprises, Danzico said. For WordPress, being consistent will probably mean similar ways of handling objects. For example, the lists of published posts, comments, and users should be arranged similarly, say by editing, viewing, and deleting them from similar-looking controls.
• Unconscious design. Good design doesn’t get in the way of users; it’s transparent in the sense that users don’t notice it, they just do what they need to do.

Those are mostly Danzico’s ideas, as much as I can glean second-hand. If someone were to ask my opinion on what WordPress should do to improve usability (and no one will), here’s what I would suggest:

• Dump the “Dashboard.” I’ve never met someone who actually paid attention to most of the stuff on the Dashboard except for the incoming links. It does appear from Danzico’s slides that there might be a Dashboard overhaul in the works, so perhaps it will soon become more useful.
• Improve attachment—i.e. image—uploading. Currently image uploading has a couple of serious drawbacks. First, there’s no control of size: you get either the thumbnail or original image. The problem is that many people just want to upload straight from their cameras, and the original image is too large practically-speaking for the web. Second, attachments must be made one at a time, which can be tedious if you have a number of photos. Unfortunately, there’s probably no way to get around the second problem without using a proprietary technology such as Flash or Java.
• Fix search. This isn’t an admin issue as much as an overall WordPress problem: you can rarely find what you’re looking for using the WordPress search box. According to usability expert Jakob Nielsen, bad site search is the number-one mistake in web design. Although that may be the most common mistake it’s not the most serious, in my opinion. I think most people get around it by just using Google to search a particular site. However, better search capability would give WordPress greater credibility as a CMS.

Mullenweg said we could expect that user interface guide in February, so I’m looking forward to seeing it, and I’ll be sure to discuss it here as soon as it’s available.