Web 2.0: SaaS Limited by Lack of Offline access

AJAX, Flash, Laszlo and FLEX2 are browser-upgrade technologies that have seen huge uptake over the last few years, with AJAX grabbing most of the attention.  All of these technologies have moved the web browser several notches closer towards the rich and responsive interfaces that are possible with thick clients, but for many applications, thick client applications still hold a significant edge. 

A major usability roadblock for onDemand applications is that web applications are only functional when you’re online.   For example, most business people would find it unacceptable if basic applications for creating documents and spreadsheets became unavailable as soon as they walked into an airplane.

With the rise in popularity of SaaS, solutions are starting to show up that allow some amount of offline access to applications.   One path is to hack/extend browsers, while another route is to create a new breed of online-aware applications.

A simple browser solution is to create single-page AJAX/DHTML applications that can be saved to the filesystem, like TiddlyWiki.  Zimbra went to the other extreme by moving the application server and application logic to the ‘client’ machine.

The Open Source Dojo Offline Toolkit is another project trying to extend browser capabilities offline.  The project is led by some developers from SitePen.  Their goal is to create a small cross-platform generic browser plugin that enables access to web applications.  It allows users to work with data even when no network connection is available.  Their approach uses a small locally running web proxy that caches files needed for later access. 

Dojo Offline Toolkit uses the Open-Source C-based web proxy called Polipo which is small at 150K and runs on Windows, Mac and Linux.  Another key component is something called Dojo Storage that run on IE, Mozilla, Firefox and Safari for caching megabytes of data within the browser process.

Microsoft has offered HTML Applications (HTA) that allow web applications to be distributed for use offline, but the applications are not cross-browser capable.

Firefox 3 will have a feature that will add local datastores that can be accessed by applications when run offline.  But the capability is limited to Firefox browsers.  The Opera browser is doing something similar by allowing clients to store data in Opera widgets.  To standardize these efforts, the Web Hypertext Application Technology Working Group (WHATWG) is developing a specification for handling persistent storage and offline web applications.

Non-browser-based network-aware applications are also evolving.  One example is Adobe’s Apollo toolset that lets developers create applications that perform like desktop application but also are able to connect to and interact with web-based data sources.

Another development tools called Slingshot lets you take a Ruby-on-Rails application and have it run offline.

With so many announced projects to attack the problem of offline access by web applications, expect this capability to become more common and something which will be expected from SaaS applications.