Re: "The beginning of the end of the desktop"

11 Jan 2005 - 1:38am
11 years ago
4 replies
1812 reads

On Jan 10, 2005, at 4:22 AM, David Heller wrote:

> But I have no backup, so I won't even go there. I would just spout more
> truisms. HTML, Flash, Java, C++ ... To be honest, in the end I just
> don't
> care. What I really care about is the interaction model that is at my
> disposal. In the past in this discussion, I have thrown out a behavior
> list
> and Ziya has said it can all be done in HTML today, but to be honest I
> haven't see any example of it.
> Since this list is about "interaction design" and not a technological
> debate, maybe we should assume that technology is what we need it to
> be. So
> I ask the question ... What do we need? Not what should we build on?

On Mon, 10 Jan 2005, Andrei wrote:
>Technology is very much at the heart of what we do IMHO. Maybe I'm out
>of touch with the list, but if designer's aren't focusing on technology
>just as much as they focus on user needs and business requirement, then
>I fear those designers will fall behind the curve over the next five to
>ten years. Of course I could be wrong, I don't claim to be right. But I
>would take a wager with anyone who wanted to on the lack of importance
>the browser development platform will play out five years from now.

My 2cs; as interaction designers, we need to design best within the
technology choice thats been made. But I also agree with Andrei, that we
need to also look for and suggest that best technology solution because more
often then not Tech Architects would try and work the technology within the
contraints of their knowledge space. To break that, it helps once in a while
to push them to the extreme to make the right tech choice. Ofcourse, as
interaction designers, we might not know how to code but we need to fully
understand what all it can do.

On the techonologies that we have been discussing, here's what is my view
(though this is purely in the apps dev space and might not be true at all
for products):
- most of the predominantly used technology, be it HTML, asp, .net, jsps,
java, etc. are in a sense open source - as in a techie can take over the
environment, decompile jars to figure out code. This I think contributes a
lot to any techonology becoming popular, because the source code is also
- in Flash for instance, one cannot do anything with the swf but just try
and recreate the entire code from scratch. I see customers and even the
development team pushing us back on this hugely on larger maintenance type
projects where for a change, one has to figure out where the corresponding
fla is and who has it. Basically huge Config Management issues

So apart from having to download an engine or a player to view an app, lack
of access to the source code is a big hinderance for a technology to become

Ramesh Krishnan


11 Jan 2005 - 11:18am
Juhan Sonin

Our design and engineering team is struggling with this same issue of
browser-based apps versus fat clients. Two words hold our wrists from
the browser flush lever: Deployment [the massive problem of every user
downloading/updating their local app] and Trust [trusting downloaded
content/app]. Flash is an option... and Java WebStart. WebStart
addresses the deployment issue: when the user starts the app, it checks
to see if there is a newer version on the deployment site. It also
addresses the Trust issue since the deployment package can be signed.
Finally, WebStart can control the installation of an appropriate JVM so
one doesn't have the worry about incompatible JVM version like one has
with a browser-based Java applet.

We're just in the preliminary stages of investigation... Delivering
remarkable, zippy, and ultra-cool apps in a browser is damn hard (we're
all pushing browser-apps to their limits). Java WebStart or some sort
of trusted, easy to update fat app strategy is needed.


Do you Yahoo!?
The all-new My Yahoo! - What will yours do?

11 Jan 2005 - 5:50pm
Dave Malouf


I have a few problems w/ Java WebStart from my perspective:
1. If I am not in control of the deployment environment, there may be blocks
for installing new signed or unsigned components in the target system.

2. Java is definitely flexible, but at the cost of a HUGE foot print if the
user doesn't already have SWING installed.

I have seen interesting deployments of WebStart functionality. I think that
WebStart is the technology that eTrade is using w/ this example:

The deployment issue as you stated is definitely a critical one to solve,
and it will be hard to solve the most complex case of it:
Uncontrolled/standardizable users (You cannot dictate the environment; user
Highly secure administration settings (the environment is out of the control
of the user themselves; corporate policies)
Low-bandwidth, or distanced bandwidth
(not enough of a market share in what you do to make a hard decision on
My requirement requires file system level integration with the client.

I haven't found a solution that deals with these 4 issues in aggregate.

-- dave

14 Jan 2005 - 1:10am


> - in Flash for instance, one cannot do anything with the swf but just try
> and recreate the entire code from scratch.

There are decompilers (of various capabilities) for SWF; for example:





If you want your code protected, keep it server side.

Nullius in Verba

14 Jan 2005 - 1:11am

David Heller:

> My requirement requires file system level integration with the client.

One of the most attractive aspects of web apps is the fact that they do NOT
provide file-level access to the OS. Just imagine IE as an open gateway to
your files!

Picking up from a few days ago...

To see how popular desktops (vs. web apps) will become in the foreseeable
future, imagine, over the course of a year, a moderate online user
installing perhaps a couple of hundred desktop apps (potentially anytime he
visits a dynamically-driven site). I just don't see people doing that.

I may want to buy a $29 item for my PC from a company somewhere in the lost
state of California. That may take a total of half dozen clicks. I'm not
going to install a separate desktop to do that. Not gonna happen. Now
multiply that situation by 100 or 1,000: we're not looking at a huge market
for desktop apps, certainly to be dwarfed by web apps.

I see, however, a much smaller market for well-designed desktop apps that
can aggregate a smooth user experience for a large number of sites -- in the
sense that iTMS, eBay or Amazon can dominate their respective markets for a
large swath of the commercial activity they cover. And, of course, for
non-trivial content creation apps.

Nullius in Verba

Syndicate content Get the feed