Another angle on best of breed in the web office game

After reading Richard MacManus' post about the best of breed Web Office products, I started thinking about my experiences with some of the applications he listed. He pointed to a few I was either completely unfamiliar with or had only taken a brief look at. After doing some catching up, it became pretty clear to me that most of the tools listed are in a pretty early stage of development Upgrades are not delivered in massive chunks. Incremental updates are the order of the day. and only a few had achieved a level of stability and usefulness sufficient for me to consider them ready for serious consideration.
The fact of the matter is that we are still very early in the game and it will still be a long while before a credible alternative to the PC-based Office suite is ready to challenge the way we have been working for the past couple of decades. Note that I said alternative and not replacement. In my opinion, that's really the key to what will define the web office.
Simply migrating Microsoft Office (or any other PC-based suite) to the web makes little sense.  I think that's one of the reasons why straight web-based knock offs of the core functionality in Word, Excel, or PowerPoint are doomed to fail. Simply replacing what I already have isn't likely to get me excited enough to switch. Give me the ability to do something I can't do (easily or at all) and I'll pay attention. For example, a web-based Outlook clone is problematic. Microsoft Exchange Server can provide Outlook Web Access (OWA) which does a fine job of provisioning Outlook functionality and look and feel in the browser. Its biggest drawback is licensing cost - Exchange ain't cheap. Microsoft Office Outlook Live, which I wrote about a couple of days ago, provides a less expensive option but still doesn't break any new ground.
Enter Zimbra - one of Richard's selections and one of the few that I think has a real chance at pushing the envelope and providing a credible alternative to Exchange. Zimbra is more than an AJAX-enhanced version of OWA. It adds some very clever smarts to the game by incorporating some nicely conceived mashups of your data and web services and raises the bar by providing an open and well-documented API that will allow developers to add even more interactive value. By talking the open source route, Zimbra all but eliminates the cost of entry although the company happily reports a positive early acceptance of the fee-based supported version as well.
Similarly, the applications offered by 37Signals redefine what we have traditionally done with Office apps rather than simply trying to relocate them to the browser. Collaboration, interaction, and a variety of notification options typify the company's offerings along with very affordable pricing. I've used Basecamp in a couple of group projects I've been part of and it does everything it promises to quite nicely. The secret, I think, is that Basecamp doesn't try to do too much and retains an elegant simplicity and brief learning curve. This is becoming the hallmark of the company's offerings.
And then there's Gmail. I agree with Richard that Gmail is one of the most useful of the Web office alternatives currently available. While still a beta, Gmail has become pretty indispensable to may (myself included). That quality reveals one of the inherent challenges facing the Web office - availability. On a couple of occasions, I've had serious problems accessing my Gmail account. Twice it was due to being locked out of my account due to experiments with Greasemonkey scripts that did something Google did not like. To the company's credit, an e-mail sent from anther account explaining what caused the unintentional violation of Gmail's terms of service resulted in a prompt restoration of access. But while I was locked out, I was dead in the water - unable to access e-mail I was expecting and reference messages I needed. I've been able to address this by using Gmail's POP3 access to maintain an offline archival copy which I update on a regular basis in a local Outlook .PST file. Google Desktop's indexing of Gmail also provides a way to keep a local cached copy of Gmail.
The availability issue needs to be addressed by any vendor whose model is based on keeping my stuff on their server. If I can't get at it when I need it you've lost me. Of course not everyone can afford the massive infrastructure Google has in place to provide uptime. And even at that, there are still any of occasions when Gmail complains the "system is busy" and suggests that I "try again in a few seconds". Annoying but not yet a deal breaker. Provisioning servers and bandwidth for a hew hundred or a few thousand users is pretty easy and inexpensive. Develop a killer web application and things get a lot trickier. Just as Bloglines, Technorati, Six Apart, and other services that have run into well-documented scalability and upgrade issues.
That's another thing that Google and 37Signals get right. Upgrades are not delivered in massive chunks. Incremental updates are the order of the day, with a new feature or two popping up from time to time and often on a rolling basis rather than all at once. Learning how one or two new features integrate with what I've already mastered is a lot easier to deal with than a 40-point bullet list of new things I'll forget about before I've had an opportunity to try them out. And when I'm not one of the first to get the new feature, there's a community of users who I know will be blogging and talking about it - a more interactive version of the ReadMe.txt file.
I think there's a lot of room to carve out success. Zoho is doing some very interesting things although much of it is still roughly formed. Writely has found some interesting ways to extend the word processing model for web office collaboration. Dan Bricklin's WikiCalc looks like a great approach. There are many (too many?) players in the calendar space trying to find the right mix between AJAX elegance and meaningful extension to what we already have available. I've been trying out AirSet for a while and think they're headed in the right direction for their target niches. But the fear of a Google Calendar looms over this space and until we see what comes out of the Googleplex, I'd say all bets are off on a clear leader emerging here. In fact, more than a few observers have opined, quite correctly I think, that many of the current calendar offerings have a decided "built to flip" flavor.
The latest example of a web office alternative done right is Google Page Creator, released today. It's not a Front Page "killer". As an aside, I'm growing increasingly weary regarding the notion that something has to be killed for something else to succeed at this early stage of the game. No... what Google got right was ease of setup and use It took me les than two minutes to make a sample page. I can see my mom and dad using this with no trouble (which translates to no tech support calls to me - always a good thing). Yes it's basic. But it works well. And only someone who has had their head buried in the sand would believe that what we got to play with today is anything other than a first volley. I fully expect to see features rolled out to Page Creator in the same way they have been for Gmail. If I were a betting man, my guess would be the addition of tables and simple style sheets are already in the pipeline.

Update: Michael O'Connor Clarke may have discovered one of the directions Google intends to take with Page Creator. He managed to repeatedly invoke a dialog box warning about the implications of another person's edits to the page. Sounds like a collaboration feature is in the plumbing.