Dear Editor:
Reading Pratik Patel's article "XML: Extensible Markup Language" (July 1998), I couldn't help but repeat to myself: "Finally." Finally, the world may put the emphasis on the content rather that the formatting.

The problems Patel mentioned with XML support in the field are very real and very prohibitive to the implementation of an XML solution, but as always, there is a way.

The hard part, as I see it, is getting the owners of the content to go to the expense of converting their documents to XML (creating an adequate DTD as well as the conversion itself). Who would want to incur this cost while cutting off 80 percent of their consumers because of inconsistent XML support at the browser level? One way is to have the benefit of XML and still cater to all browsers. You could create an add-on to your Web server that would, based on a "rules" file for your XML document type, convert the XML to HTML on the fly. For browsers that don't have XML support, send straight HTML; for search engines and browsers with compatible XML support, send native XML. No one loses, and information is taken to a new level. Finally.

I'd also like to add that this is the first time I've seen SGML properly explained in print. Congratulations.

James Sauve
Ottawa, Ontario

 

Dear Editor:
I just now finished reading Pratik Patel's article "Middleware and JDBC" (May 1998). Finally, somebody explained in human language what application servers actually do.

Today, when we are fulfilled by "buzzword-compatible" terminology of long and boring brochures and product information sheets that actually say nothing, it was a real pleasure to hear a human voice instead of a lot of "white noise!"

Dmeetry Raizman
President, CreDO Software Technologies Ltd.
Tel-Aviv, Israel

 

Dear Editor:
I just read the letter to the editor from Louis Langlois ("It's important to integrate mining and statistics," July 1998), and while I applaud him for his innovation in finding a solution for a particular situation, I would like to express my opinions concerning Y2K.

It has been my experience that many people do not take the Y2K problem seriously. This, in and of itself, contributes to the problem. A commitment must be made at every level of an enterprise to inventory, renovate, and test all items that will be affected. It is a lot of tedious work, the kind that programmers do not like to do.

The job should be done right, and code and files should be changed. Anything else is asking for problems during later maintenance. Some valid windowing techniques being used require extensive documentation when maintenance is required in, say, 2005.

Testing is the largest part of Y2K-compliance work in time as well as money. Environments must be set up to handle future date testing, and while this is costly even in the client/server environment, it can be overwhelming in the mainframe environment.

Like everyone else, I find humor in much of the problem that faces us but also think we must keep in mind the magnitude of the tasks ahead.

Mike Smith
Dallas, TX