If you're into Web 2.0, AJAX, or even just web development, you must read the rather paranoid "Danger 2.0" article in the October 26, 2006 edition of Network Computing (Vol. 17, No. 22).
I'll warrant many of the authors concerns are likely valid, but at the very least his presentation is overly intense.
One section mentions how the payloads in XML and JSON transfers increase traffic to and from the server because of syntactical requirements. Sending the entire page is better?? Granted he's coming from the SOAP world...but if anything has intense syntactical requirements it's SOAP.
True, there's no standard markup in AJAX, but that means companies can chose their own XML, JSON, or just plain XHTML for their transactions.
In the end there are security and load concerns with AJAX, but everyone switching to SOAP, Active X (herk), or back to fat clients is certainly not the answer.
Just my opinionated $0.02.
Well first of all you're sadly misinformed as to the gender of the author of this article.
Second, the article doesn't mention anything about returning to fat clients. It presents both the good and the bad about AJAX-based technologies and is a good reminder that you need to keep in mind the security and performance considerations before deploying such applications.
As far as no standards being a good thing goes, obviously you've never worked inside the data center and had to deal with integration in a proprietary world. The fact is that standards enable SOA and if the enterprise is going to adopt AJAX and Web 2.0 technologies then it is imperative there be standards to support that adoption.
Why do you think IBM and the Open AJAX Alliance exist? It's primary goal is to standardize the tag libraries for AJAX in order to ease transition from one toolkit to another.
Standards are not evil things, they are a necessity to engender the widespread adoption of a technology.
Posted by: Anonymous at November 27, 2006 09:34 PMDear Anonymous,
Thanks for posting your thoughts on my comments on the NWC article. I apologize for getting the author's gender wrong in my post. I'll be more careful about that in the future.
I whole heartedly agree about the need for analyzing the security and infrastructure needs of an AJAX app, just as one would with any other application. Until reading that article, though, I'd never thought of AJAX as being any heavier on the server than old-school web-based apps were. Certainly it's heavier on the client, but that's to be expected (and certainly planned for).
My post didn't convey my own thoughts on standards very well. This sentence, I think, was the misleading one: "True, there's no standard markup in AJAX, but that means companies can chose their own XML, JSON, or just plain XHTML for their transactions." The thought behind that sentence was not one of relief at their being no standards associated with AJAX, but the freedom for companies to select (or create) standards related to their industry or needs--I'm thinking the payloads between the servers and clients here, not the interface.
Companies certainly need to codify their methodology on how their AJAX platform will work within their company. Interface specs should be standardized as well, but I feel that's already begun to happen with the several AJAX libraries that are available. Granted, since we're in early-adoption mode, these libraries (and their proposed methodologies) get confusing and picking a winner is a tough choice. However, I'm not sure that a standards body is the best vehicle for picking the winner.
In the end, I would like to see standards bodies like OpenAJAX, the W3C, etc. continue to create more building blocks for the web (standardizing XUL comes to mind as a good idea). Then let the developers choose what they'd like to use. Giving developers and platform builders marks to hit as well as accountability is the right way to go, but even that effort should be kept from overrunning its own intentions by bitting off more than is chewable.
Just my very meager $0.02.
Posted by: Benjamin at December 1, 2006 04:15 PM