Hi Aaron
These technologies seem to be getting lumped into the class of Rich User
Interface. Another interesting option is Curl (not to be confused with the
language curl or the PHP method). (I?d really love for IBM to purchase curl
and make it the replacement for 5250)
I agree with Nathan that they look like desktop technologies - and to me
that is because a rich interface needs a lot more code infrastructure to
support it than a browser does. The richest, most productive interface you
can find is the Windows environment - forgetting the support issues and
deployment and management complexity. The UI Widgets supplied by Windows are
incredibly productive and efficient. If only you could deliver this to the
desktop without all the accompanying problems of Windows itself.
Having said that, in my opinion the simple browser interface is a dead-end
technology for data entry and any meaningful, involved interaction with an
application, it just doesn?t work well enough unless it is loaded down with
a ton of Javascript. It may be that it can be made to work with some kind of
standard library such as Webkit or whatever, but that just proves to me that
it was the wrong answer in the first place - the functions that the
additional javascript provides should be an integral part of the client, not
tacked on. I just see all this javascript library stuff and additions as a
cobbled together answer that will eventually break - it's an incredibly
brittle environment.
Packaging and delivering Flex and Silverlight and Curl as browser plug-ins
just takes advantage of the fact that the browser is a delivery mechanism
par excellence. Users seem happy enough to deal with adding a plug in, when
otherwise they might wonder about downloading and installing an additional
application. It also has the advantage that the rich client operates from
within the browser making it "part" of what is an already familiar
environment to the end-user.
Any meaningful application also needs to consider the requirement for local
sandbox environments and connected/disconnected operation. It seems to me
that the clients designed from the get-go to do this can safely provide this
capability but the simple browser can't (safely) be made to do this. It will
just get more and more unwieldy. I guess it pays to keep in mind that the
browser was really designed to deliver and present content not work with it.
Popularity isn?t king - ubiquity is king. You can count on any user on the
planet to have a browser and therefore you can count on them having a
predictable interface that operates in a well understood way. The end user
doesn't need training and can be immediately productive.
Don?t discount Silverlight - the battle between Flex and Silverlight is
really about capturing developers not end users - and Microsoft has proved
on a regular basis it understands how to get developers to use its products.
Regards
Evan Harris
-----Original Message-----
From: web400-bounces@xxxxxxxxxxxx [mailto:web400-bounces@xxxxxxxxxxxx] On
Behalf Of Aaron Bartell
Sent: Saturday, 20 December 2008 11:52 a.m.
To: Web Enabling the AS400 / iSeries
Subject: Re: [WEB400] Business Developers was ->Re: IBM Gives RPG Devotees
Their Own Café
>You mentioned Silverlight and Flex. They may be advertised as Web
technologies, but they look more like traditional desktop technologies to
me.
That's an important observation. I think there are a couple criteria for
the way "the next big thing" will be successful:
1) Easy to deploy on existing infrastructure (both Internet and intranet)
2) Popularity (unfortunately)
Commenting on #1:
What makes things easy to deliver these days seems to be anything that
can
be sent from a server to the desktop via browser. Silverlight, Flex and
JavaFX all have the power to deliver through the browser both the initial
framework and subsequent business applications (compiled or what not -
i.e. swf files for Flex). But the uptake of any of those will be
dependent on popularity which takes us to point #2
Commenting on #2:
Popularity seems to be king these days regardless of whether a technology
deserves it. For example, just because Flash already exists on 90%+ of
all PC's out there it will be able to get a solid share of Flex apps to
be
successful (even though there may be an xyz technology out there that is
way better). Silverlight will only be as good as whether Microsoft
recognizes that the world is no longer made of Windows desktops and that
Mac is catching up (and Linux a little too).
I mention JavaFX whenever I bring up Silverlight and Flex but I have yet
to see a real good example. Seems like it is escaping me or something
whenever I try to research it. I get the feeling that it is going to be
slow and bloated (like the 3D desktop from Sun that I tried to run awhile
back that nearly brought my quadcore 4GB PC to it's mercy).
On my TODO list I have an intention to dig into the TN5250 and TN5250J
projects to see how they evaluate the 5250 stream and eventually render
it
to the client. In reality the 5250 data stream works for a majority of
my
needs, I just need better UI widgets so I can "sell" it to the people I
am
building web apps for. I just want to process "records" in and "records"
out, and wait for user interaction (i.e. events). I think this is what
the EGL team is attempting to accomplish, but the mess of open spec
technologies is introduced and the lack of OS400 integration is also a
downer (i.e. job control vs. everything running in a JVM).
Aaron Bartell
http://mowyourlawn.com
Nathan Andelin wrote:
midrange.com
