2013-03-15

Considering Windows 8

Since Windows 8 came out I've been using the release version on a brand new Acer laptop. The laptop itself is a reasonable one, more than up to the job of day-to-day use, with a touch-screen specifically intended to make for easier use of Windows 8. Beyond that, nothing about the machine is Windows 8 specific in any way, but I can also say that nothing about it is too anemic to run the O/S well. In short, I would have no complaints about recommending a similar or identical configuration to anyone using the new O/S. And rather than list those specs here, I'll qualify that what I'm really saying is that the following commentary isn't affected by any hardware-specific quirks that prevent the O/S from doing what it intents or desires.

Would I recommend Windows 8?

In short, no, even though it does have some useful O/S-level extensions that are improvements. The file copy visualization alone is handy, as are some of the fine tweaks to the file explorer and other applets that come as a standard part of MS O/S setups. But taken together there is not one compelling reason to recommend Windows 8, and there is one overriding reason not to recommend the O/S as an upgrade. (Having said that, it comes on new machines, and frankly there isn't much reason to avoid it either in those situations.)

The overriding reason to avoid Windows 8 is its "new" menu system (formerly called Metro). It is simply not a valid UI for a desktop machine, regardless of whether you have a touch screen available. (That is why I noted its presence above; it really isn't a compelling extra.) Besides it looking like a unicorn has vomited on the monitor (the colour schemes are awful), it is chunky and inefficient on a desktop. Ultimately, it looks like a candy-box cover, I suppose, and maybe it does work on tablets, but it simply isn't effective in a workplace machine environment. Is it a great production drain? No. But does it add anything at all? No. And yet it is the main reason to avoid the O/S for two reasons that are fundamentally unavoidable conclusions if you use the O/S for any period of time to do actual work.

The first reason is that as UI conventions go, on a laptop (or desktop) monitor the chunky and flat nature of the interface means a significant amount of scrolling. Yes, you can trim the iconography, choose small instead of large panels, and so on...but ultimately a menu is used to launch tools, and the absence of multilevel grouping is problematic. In the old Start Menu (or Orb, or whatever the damn thing was called) you could create sub-folders to hold ancillary programs, so that the main menu exposed your primary applications and you could drill down for the few occasions you needed a meatier interface. In this new model you have one left-to-right grouping pattern, or otherwise have to use the search UI to find what you want, or have to put up with an inordinate slew of visually boring iconography. Try installing MS Office 2012 on a Windows 8 box and you instantly see the problem, where every installed tool suddenly has the same level-precedence as the primary applications. And, really, the Microsoft Clip Organizer is not sensibly left on the main menu beside Word. It just doesn't make sense. But the menu system is flat, so there is no drill-down -- you either have it visible or hidden. And while for the expert user hidden is an obvious choice, inexpert users are not going to find that feature, and if they do they are not going to remember to search to find their applications once they are accidentally demoted from the main scrolling menu system. It's simply a bad design taken to ridiculous prominence, and renders a working machine clumsy. Yes, it probably does work with a web-only style box, but frankly forcing the desktop user to suffer it is an unnecessary step.

Probably the larger reason for this being somewhat a show-stopper though is about extended user experience. On a machine intended to work, the Windows Store is a poor experience in terms of depth of real utility. Sure, there are games and the like, but actual utility applications (things you do real, lasting work with) are few, and those that do exist almost always end up with the same launch experience: click the icon, and be punted to the old desktop minus the "start" button element of the UI. And that begs a question that displays the degree of almost arrogant foolishness of the UI designers who let this happen -- why give prominence to a UI convention like that new menu when the end result of clicking most icons in it is to pitch you to a familiar working desktop? It makes no sense whatsoever to layer on a shell that is functionally impaired to that degree. (The same could have been said for all experiences inside Windows, before Windows 95, but the difference is this clumsiness is not a paradigm-shift requirement.) So many better methods to engage the flat-model UI could have been chosen, while recognizing the transitional necessity of the desktop. For example, in the simplest model, reduce the start button to a flat panel button with the windows logo, which when clicked expanded the menu system. Keep the bar, folder and menu model in the left as a side-bar style, and expand the right pane region to show the flat iconography. Based the decision on where to display the icons entirely upon some new code detection scheme. That would have allowed the desktop world to coexist, reduced the transitional stuttering, and maintained the preeminence of the new menu model by giving it a visual showcase without interfering with functionality.

The problem with Windows 8 is small and large because of a fundamentally mistaken assumption about the need for change, and the nature of change. Yes, a new UI convention for menu management might be wise (ribbons are actually not a great trial once you get used to them, for example), but there is no need to break with the past when the transition is going to take time. People would accept that old programs remain in their menu model, and new ones absorb a new model. But by forcing away the menu model (Classic Shell can return it), MS created an experience where users doing real work on a machine will become aware that their required applications do not play well in that new menu; they are constantly shunted to a visually crippled desktop, and my guess is they will quickly gravitate away from new-shell friendly applications when they begin to find this experience jarring. That could have been avoided with almost no effort, but instead the path of "we know best" was chosen, only to rapidly prove otherwise when every useful application punts the user to the old desktop UI.

Would I recommend Windows 8?

No, because in itself it gives very little of immediate value to a user who is on Windows 7, or even Windows Vista. Upgrades are unnecessary.

But nor would I say flee from it, because if your new machine comes with it, with a few clicks (see Classic Shell) you can recover your old interface, and never actually have to see the "new" look again. That isn't much of a recommendation of a UI or Windows 8, of course, and that is a sad fact indeed.