I was glad to see last week that Sun president Jonathan Schwartz agrees with my analysis that the desktop PC is history, making Microsoft's dominance of the platform irrelevant. But then of course he went and spoilt it by assuming that everything gets sucked up into the cloud onto Sun servers. At least people are starting to realize something I've been saying for the past five years: clients are in the cloud too. We're talking about distributed computing here, after all. That means computing will take place wherever it makes best sense for it to happen, and some of the time that will be on the client device too — as reinforced by developments like Ajax.
But while Ajax is good for many things, all these wannabe Web 2.0 killer app vendors who are coming out with rich-client versions of Office applications really are missing the point. I suppose if the cloud is everywhere, then it's inevitable there's going to be a lot of foggy thinking about, but let me see if I can just cut through the mist here.
For a start, people are being misled by Microsoft's Office brand name, which is a complete misnomer. OK, you use the applications at work, in the office. But it would probably have been more accurate (if less catchy) to call the suite Microsoft Cubicle. There's a reason applications like Word, PowerPoint and Excel all used to be grouped under the catch-all of 'personal productivity applications'-- they're designed to increase the productivity of individual people working in isolation.
What the Halfbrain team did was way ahead of its time (the spreadsheet was even multiuser). It got killed off by over-greedy venture capital investors who didn't see the dot-com crash coming. As I described in a series of postings to my Loosely Coupled blog in July of last year, the company got bought by Alphablox, which in turn was acquired last year by IBM, while several of its developers went on to found Oddpost, the Gmail precursor that Yahoo! bought. In a revealing follow-up posting, IBMer Koranteng Ofosu-Amaah described his work in porting the Halfbrain/Alphablox technology to the Mozilla (now Firefox) browser platform to provide a set of "Simple Browser Productivity Components" in IBM's WebSphere Portal.
How come such far-sighted technology ended up somewhere in WebSphere as a set of cute browser utilities? That's the best outcome any of these 'Office killer' applications can hope for.
If I want to sit down at a computer and compose a letter or report, analyze data in a spreadsheet, or create a presentation, then I really don't want the Web to intrude (in fact, come to think of it, what I'm doing now is precisely an example of that. I wouldn't dream of writing this article online in my browser: I use Notepad to craft what I'm going to say, and then paste it into the blogging application when I'm ready to do my final edit — including looking up the hyperlinks — prior to posting the article). When it's just me as an individual using the computer as an aid to my personal creativity, I want the computer to act as a seamless extension of my mind. I don't want to have to pause my stream of consciousness while some router somewhere falls over midway through delivering a screen refresh or some dumb suggestion to 'improve' my grammar.
The singular exception is Outlook, which Microsoft grafted onto Office when groupware was all the rage in the mid-1990s, and we users have been paying the penalty ever since. Outlook is a great illustration of why personal productivity applications don't mix with collaborative applications. Word is complete overkill for composing emails — just as much as it is for blog postings. Microsoft has made a pig's ear of the implementation too, something I am reminded of every time I am forced to use Word to edit Outlook emails and am confronted with the idiocy that there's no way of enforcing consistency in menu layouts between the two instances of Word. But don't get me started on that now ...
Where the Web comes into its own is in collaborative applications, such as jointly authoring a report or an article or designing a presentation in co-operation with a virtual team of domain experts. Before we had the Internet, of course, people used to congregate in offices for the precise purpose of performing this kind of collaboration. That's why I say Microsoft's suite would be more accurately named Cubicle. It was originally developed with no collaborative capabilities whatsoever, and whatever capabilities have subsequently been grafted on are pretty lamentable on the whole, up to and including SharePoint. The product thus bears little or no relation to the true concept of offices as people experience and use them in the real world.
That's why I say that anyone who emulates Microsoft Office in an attempt to build the killer collaborative application suite of Web 2.0 is on a hiding to nothing. Workers who do a lot of work in isolation will most likely continue to use Windows and Office. The rest of us will use a completely new generation of applications that automate collaboration and integration rather than isolated individual endeavors. Whatever Microsoft has historically found success with on the cubicle-bound desktop is irrelevant to what is going to succeed in the collaborative, virtual workspaces of the Web 2.0 era.