X
Business

SaaS in 2007: reaching out to offline users

Earlier this year, I got flamed for daring to suggest that universal connectivity was just a pipedream. But now some of the poster children of Web 2.0 are bringing out offline clients. Expect many more in 2007.
Written by Phil Wainewright, Contributor

Just a few months ago, I got flamed by several usually friendly bloggers for pointing out Why Office 2.0 will never go wholly online. Timed a few days before the Office 2.0 conference, my posting was lambasted as "wrong and naive" by Stefan Topfler, while Dennis Howlett said I was both "ignoring reality" and "deny[ing] history"!

Yet by the time the conference had wrapped up, quite a few attendees seemed to have come around to embracing some kind of hybrid model for on-demand applications. Now, as we approach the end of the year, two of the best-known vendors associated with the Office 2.0 movement have announced new versions of their applications that support disconnected working.

Socialtext's unplugged icon
Thinkfree, one of the longest-established Web-based productivity suite vendors, yesterday announced that users will be able to buy a paid version of its application that allows them to work on documents while offline. Earlier this week, enterprise wiki vendor Socialtext launched Socialtext Unplugged for "the occasionally connected user," as CEO Ross Mayfield put it (thanks to Socialtext for the unplugged icon pictured).

Seems these two vendors have been listening to customers and have acknowledged their customers' need to work on documents even while disconnected from the Web — exactly the need I had previously highlighted.

My prediction for 2007 is that we'll see more and more on-demand vendors coming out with clients that support offline working. In fact, I think we're going to see an important debate developing in the SaaS industry about the best way to deliver functionality to users. Most vendors still aim to do it in the browser, but I think there are limitations to technologies like Ajax, especially when compared to the 'smart client' capabilities of emerging platforms like Microsoft Vista and Adobe Apollo. Those limitations have already led RightNow Technologies, for example, to introduce a smart client to improve the end user experience.

I'm currently writing about some of these issues in a series of articles commissioned and published by on-demand CRM vendor entellium, which recently introduced its own smart client platform (entellium is a client and I also hold a small equity position. See disclosure page). The latest article has proven to be very timely in view of the news from Socialtext and Thinkfree, and in particular takes issue with those who argue that broadband connectivity will soon be as ubiquitous as electricity:

"... the more mobile our computing becomes, the less comparable it is to the dependable utility power supply that comes out of the wall socket. Frequent interruptions on wireless networks are the norm, not the exception. They're cellular precisely because mobile devices disconnect and reconnect all the time, constantly switching to stronger signals as weaker connections fail. If this is the environment the devices are used in, then it's far from logical to design the software that runs on them as if connectivity will be a given."

The key point at issue here is that intermittent disconnection is an inherent fact of life in a network environment, and that a truly robust on-demand application design will accommodate disconnection in a transparent and non-disruptive manner. Anyone who designs their clients to be continuously dependent on their connection to the center is actually perpetuating a hangover from the days of monolithic, centralized computing. Clients aren't second-class citizens that are only worthy of consideration when they're visible over a live network connection. In a truly net-native architecture, clients are part of the network too. Even when they're offline.

This is the first of three articles in which I'll be making predictions for SaaS in 2007. What are your predictions? Same or different? Share your views in Talkback.  

Editorial standards