It never ceases to amaze me that I will have a conversation about something esoteric with someone and it will appear the next day in someone else's blog who was totally unaware of the conversation.
We are currently planning an enterprise scale piece of software and have been going back and forth about what our UI strategy is going to be. There are many choices for us, as this applicaiton involves web services, geographic information systems, asset management, real time data visualization, hardware control interfaces, diagnostics and still more.
It would be nice to make this application run on a central server and provide a browser based UI or a set of browser based UIs. This means that nothing needs to be installed on local machines, that multiple users across an enterprise can have access, and that people could easily work remotely.
On the other hand, browsers place serious constraints on how people can interact with their data and with the UI itself. One of the main points I made yesterday was that browser UIs are lousy for giving us context sensitive right click menus. Then, I wake up this morning and Robert Anderson shoot off this post.
Now, in the interest of full disclosure, Robert is my cousin - so it's possible that there's some direct link between us on the astral plane. (Though this means I'd be even more closely related so some of our more ... interesting ... relatives.)
Given the higher evolution of the Mac in usability - I'm not sure I buy the single-button mouse concept, but implied in his criticism of Web 2.0 is one that I, myself, have weighed in on. Web 2.0's focus on "web" also focuses it on "browser" which is currently a highly limiting platform under which to create applications.
An internet-aware java desktop application can be much more powerful and lend itself more toward data integration, than are all these stand-alone web 2.0 apps we see today.
We keep trying to force the browser out of its limitations with AJAX, .NET, and so forth. But it's like adding levels to a house sinking in quicksand. We want it to be more than it ever was meant to be. Sort of like Windows and DOS.
Having said this, we still want to keep a lot of what browsers represent:
- A highly standardized, open platform for development
- A familiar look and feel
- A customizable platform for users
- An easily understandable environment for non-techs
- Freedom from installation contingencies and support for widely varying system configurations
What we want to get away from is:
- A system built thin for computers with 16 meg of RAM and running at 166 Mhz
- A system built for access to the internet through a 1200 baud modem
- A system with a highly limited user interaction paradigm
Ultimately, for me, web 2.0 is about creating portability in data, ownership of your own information, and freedom to associate. None of these elements are reliant on the browser. Perhaps it will rely on the astral plane.
Photo: Clara Natoli
Technorati Tags: web 2.0, astral plane browser, firefox, p2p
I just chanced upon this page thru some random technorati search and found it interesting.
"One of the main points I made yesterday was that browser UIs are lousy for giving us context sensitive right click menus."
How about creating custom a extension to handle the client-side UI functionality? (I know browser is a generic word, but i'm encouraged by your 'firefox' tag to make this suggestion :) ) Agreed, it requires some client-side installation of the extension, but I think it is quite straight-forward and intuitive, and could be as simple as clicking a link (of an XPI file, on your main page).
Posted by: Ravi | 15 February 2006 at 14:27
We've been running down that road as well. That may be how we end up.
I can't help but thinking we're still lipsticking the pig though. I wonder when you cross the line between extensibility and shoring up something that doesn't meet your needs.
If we create the extension, it works for us - but it's not standardized. That makes the solution somewhat stovepiped.
Posted by: Jim Benson | 15 February 2006 at 15:08
If you create the extension and then open-source it -- or contract out the development of the extension to someone under the conditions that they open-source it -- then you might just get beyond your "not standardized" problem in time. Ping me if you need a name.
Posted by: Edward | 15 February 2006 at 21:35