One unforeseen benefit of the rise of netbooks is the rethinking of desktop interfaces. Compared to workstations, netbooks have smaller screens and less memory, and developers generally assume that users do less demanding tasks on them. As a result, developers are starting to redesign interfaces to function better within these constraints.
These assumptions are questionable -- the hardware distinctions of netbooks are blurring at the high end, and many users, especially travelers, are using netbooks for more than email and web browsing. However, by rethinking interfaces based on these assumptions, developers are highlighting the question of what users actually need. Considering the tendency of all applications to bloat with each release, this is a welcome re-examination.
Few netbook interfaces illustrate this re-examination better than Anjal. Less than a year old, Anjal is a light interface for Evolution, GNOME's default mail reader, and one of free software's main answers to Microsoft Outlook.
Evolution has been part of GNOME for a decade. However, the last major changes came with the 2.0 release in 2004. Since then, Evolution has become more stable, but has changed so little functionally that users could easily conclude that it is a low priority for GNOME.
Evolution, GNOME's default email reader
According to the project's home page, "Anjal" is Tamil for "mail." Srinivasa Ragavan, Anjal's developer and Evolution's maintainer, as well as a member of GNOME's board of directors, puts the new interface firmly in the emerging netbook tradition when he writes:
Anjal doesn't replace Evolution. Anjal is a lite UI built on top of Evolution and uses 95% of Evolution code for its experience. Anjal was built to suit the smaller screens and low memory devices. Particularly Netbooks and the mobile segments. It has a very interesting UI, that is TOUCHable and consumes less memory and deals efficiently on the real estate for the small screens.
Currently, Anjal is in rapid development. However, it available as source code and from Ubuntu's Launchpad, and installed by default in Fedora 12, regardless of the target computer -- an alacrity which indicates just how overdue a rethinking of Evolution has become.
The Great Simplification
The first time that you run Anjal, like Evolution, a wizard steps you through setting up an email account. When you are done, you are on a Settings tab, where the only option is to create another account -- a reasonable end point, considering how many people have multiple email addresses these days.
Anjal, a light interface for Evolution
Still, even at this early stage, you can hardly help noticing how fewer features Anjal has than Evolution. It is a mail client plus an apparently unimplemented address book, and nothing else. The calendar, tasks, and memos apps that are part of Evolution are simply not there.
Then, as you go back to the Settings tab, you realize that this is not all thats missing. Besides mail accounts, Evolution's configuration options include another six general categories: Autocompletion, Mail Preferences, Composer Preferences, Network Preferences, Calendar and Tasks, and Certificates. Most of those in turn have at least a dozen settings, and some have several tabs. I haven't bothered with an exact count, but sixty or seventy settings must have been jettisoned by Anjal as unessential.
The same is true elsewhere on the application. Not only are no menus to be had, but the toolbar is reduced to six buttons: Check Mail, New Mail, Search, Sort By, Setup, and Close. Reply, Reply to All, Forward, Print, Delete, Junk -- half the default buttons in Evolution are gone.
Some of the missing functions show up when you select the Reply To button. For instance, whether you select plain text or HTML email is a matter of a combo box, and the choice of address fields to display (To, CC, and BCC) has moved from the preferences to a couple of buttons on the taskbar. But, aside from the basic formatting options for HTML mail, the only other choices are to show the attachment bar, add an attachment, save a draft, and to send the email.
Another simplification in Anjal is in the names on the interface. Instead of Evolution's Preferences, Anjal displays "Settings." Similarly, "Check Mail" replaces "Send/Receive" with its more geekish forward slash. Instead of "Contacts," you have "People."
Without exception, where Evolution has a longer and more abstract description, Anjal has a shorter, more concrete one. That may seem like a small change, yet you would be surprised how Anjal's more direct language makes it easier to navigate, even if you are already familiar with Evolution.
The same is true of the interface. Except on the list of mail folders in the left hand pane, Anjal makes no use of right-click context menus. In much the same way, by using web browser-like tabs instead of Evolution's buttons for changing tasks, and separate windows for new emails and configuration settings, Anjal achieves a unity that Evolution cannot match. Although the main reason for these changes may be a netbook's smaller size and lack of resources, the user experience also benefits from Anjal's radical simplifications.
Re-evaluating Feature Creep
Anjal's simplifications can be exaggerated. In some cases, Anjal assumes a reasonable default, or moves a feature rather than eliminates it. Nor can you always be sure if a feature has been removed or rearranged, or simply has not been implemented yet, although the amount of code that Anjal shares with Evolution does suggest that the feature set is reasonably complete -- unless I missed something, Anjal does not include any unique features.
At any rate, Anjal provides a case study of how a standard application can benefit from a thorough housecleaning. Essentially, Evolution's interface represents the thoughts that GNOME had ten years ago about what users needed. Back then, GNOME's users were more likely to be stone geeks, while now the majority are unlikely to have seen, much less written a line of source code. Considering this shift in audience, it would be strange if the requirements for a mail client hadn't changed as well, with fewer people demanding a full range of options.
Also, while each new feature over a decade may seem like a good idea at the time, in retrospect, some might need to be re-examined. For example, do most users really need a separate basic search field and an Advanced Search? How many will choose which plug-ins to enable or disable unless they are having specific problems? And does anyone change the color of quotations from the default dark gray?
Once you start with these questions, you might also wonder what might be automated. Instead of specifying which address book should be used for autocompleting addresses, is there not enough power in a modern computer to search them all (assuming that the average reader has more than one?) If Expunge is a useful feature, couldn't it be done automatically -- especially since most users probably have no idea what it is? Does a mail client really need a separate function for replying to a list, or could that be taken care of behind the scenes?
Naturally, for each user, there will be a list of essential features. Experienced Evolution users will probably miss in Anjal the ability to set message filters, or to set a standard signature. But the point is, once you get over the surprise that some features are simply not present in Anjal, you may find that the list of essential features is smaller than you imagine.
Whether Anjal will succeed on its own merits is uncertain. Even if the limits imposed by netbooks' size and resources require a new interface, users might very well not want that interface to be had at the expense of features. Yet, even if that proves to be true, Anjal may still be useful if it leads to a streamlining of Evolution itself. If so, then users could easily feel that Anjal is an exercise that is long overdue.