The return of rich clients and the twilight of “HTML applications”?

Tomáš Kouba’s The Decline of HTML Applications?
got me looking into what a rich client is and thinking about it a little. But first things first. His post includes
a table with the percentage share of different clients for distributed applications in 2002 and projections
for 2005. I copied the table from his post.

Client type — 2002 / 2005. 1st Generation Rich Client — 24% / 6%. Stand-Alone Rich Client — 7% / 30%. Browser Rich Client — 5% / 18%. HTML — 64% / 46%.

The post and the table use the term rich client. A rich
client offers the convenience of conventional desktop applications using widely adopted
web and internet standards. Convenience, or usability as I’d rather call it, is the key.
Users want an interface and capabilities like the ones
in their operating system and its applications. That includes a standard graphical interface,
drag-and-drop, good printing, consistent state when moving between pages, and
personalization. Why use web applications that output HTML and, in most cases, depend on the browser?
That dependency matters if they’re expected to meet even some of those requirements.

Web applications as we know them are reaching and often exceeding the limits of HTTP, HTML, DHTML and the other
capabilities browsers offer. That’s where rich clients come in. They’re platform-independent
programs that offer more functionality and convenience than today’s browsers, or
practically any browser using commonly available technology.

Stand-alone rich clients are expected to account for 30 percent
of all client types in 2005. Java and .NET are very likely to be the main
platforms. John Yu points to some evidence for
that in Rich clients emerge as alternatives for Web applications.
Browser rich clients will gain share too. They’ll likely be built
on widespread technologies that are or can become part of a browser or be integrated into it.
The best chances belong to technologies already widely used, such as Flash, XUL (Mozilla), ActiveX and Java applets…

If we’ve got this far, let’s ask what kind of web applications this will affect. Will it be
ordinary B2C (business to customer) or complex B2B (business to business)? It will start, and already has started,
with the more complex applications, then move slowly toward widely used ones
such as online banking. The trend will probably bypass the web interfaces used
by free services such as email.

I’d conclude that pure HTML applications were here before, are here now and will still be here in the future. What will change
is that they’ll become more specialized for particular kinds of work.