May 11, 2005
Proprietary AJAX?
When I first read about AJAX methodology, it sounded awfully familiar to me. Because I'm a Siebel developer, and Siebel has been using what they call a "highly interactive web client" or a "Smart Web Client" since 2001. DHTML. JavaScript, request and response. I'm not the first to say that this sounds like a fancy new name for things that have already exist.
However, I don't believe AJAX is all hype, as some do. We need better web clients. Especially in the enterprise application space, data driven applications that are essentially front ends for databases. Everybody wants to do this, and it's hard. A bloated behemoth like Siebel web client does the job (after a few years of tuning), and pretends its web based, and really isn't as good or fast for the user at all as the Siebel 6x fat client version, but it does have some architectural advantages.
So now you really have a highly interactive AJAX web browser with a Siebel 7.x client. And Siebel was doing it a long time ago, allowing any 22 year old Accenture consultant to configure and publish a highly interactive web client without even understanding a HTTP POST request. Which I can't prove to you because you have to be a mega monster corporation to actually use Siebel. And this is the most interesting point to me, which is the value of openness in technology development. No AJAX isn't new. The pro engineers who put it together for Siebel and the product development team who asked for it sure knew about it. But when your corporate strategy is based on secrecy and non openness, as any Enterprise Application software company is, you can't share what you've created with the world, and improve it. You can't set standards. It's not science. I would argue that the best of new (or old?) software development techniques centered on true open source and developers is a model much akin to science, and proprietary software companies are headed for some pain if they can't adjust to new trends related to open source.
Will the AJAX ideas come down heavier on the marketing side or real developers? Lets say it's a step ahead of Siebel, which is falling apart right now as a company, and a step behind Linux/Perl. A little hype is good for the system anyway.