Netcentric View:

By Bernard Cole
iApplianceWeb
(08/09/01, 010:10:52 AM GMT)

In the last two or three columns I've been dwelling on the subject of names for two reasons. The first is that I can't avoid the issue when I talk to many of you about your products and technology.

In the new rapidly changing connected computing environment virtually all of the terminology used to describe events and things have been bent and mangled to reflect particular viewpoints.

I find I have to interrupt the conversation and ask questions like, what do you mean by embedded? What do you mean by real time and determinism? What do you mean by wireless? Networking? Communications?

In the case of the word embedded, I find that some of you just use it as a convenient catchall description into which you put things that don't fit anywhere. And as such, embedded yesterday does not necessarily equate with embedded today. Sometimes you have a very specific meaning to the term, but it's not the same meaning that everyone else in the conversations is using. Often I find that the use of the term obscures the views that people have about the changing computing environment.

I am not saying we have to replace the word embedded to describe a market, but we need to modify it because the nature of embedded in one environment has little or nothing do with its nature in another.

The second reason I have got a thing about naming is partly philosophical and partly practical. What I think I bring to technical journalism is an ability, carefully nurtured over the years, to see things clearly, to see connections between things unfiltered by misconceptions and previous conceptions of things.

And what I find is that language often gets in the way if I don't use it properly because language is thought, and vice versa. Using old language to describe new phenomena will doom us to thinking about the new in old terms.

Aside from leaving engineers and developers in the dark about what to design and how to design it, using old terms such as embedded, desktop, or server and router imposes a linguistic and conceptual filter through which it is not possible to see clearly.

In the case of the new net-centric computing environment, when viewed without such filters it is clear that traditional, well-understood sales, marketing and OEM channels defined by such terms as embedded, personal computers, servers and even wireless phones, are disappearing. New web-based service channels that bear little resemblance to the old ones are replacing them.

The channels of common services have changed so that the diverse makers of PCs no longer share the same channel, nor do the makers of cell phones, servers, handheld internet or information appliances. Depending on how they are used, everyone in a Web services-based channel, whether they build the servers, handheld devices or wireless telephones, may have more in common than each individual community.

For example, consider the builder of an application server in an enterprise that has as its main function providing services to millions of net-centric embedded devices in refrigerators, ovens, washing machines and the various consumer electronics systems in our homes. The company has more in common with others in this channel of related web services than it does with the designers of, say, application servers for web-enabled cell phones and other iAppliances.

Think about it. In building application servers for embedded devices in the intelligent home, the concerns of the server builders and of the end-device developers are the same, and those concerns have little to do with human interface. The device builders are focused on how to collect information in the device on the system operation and format it for delivery to a server, the opposite of the traditional view of a client-server relationship.

The builder of the device hardware and software has got to be concerned with the control application and its real-time and deterministic operation as well as delivering information to the server and dealing with downloaded upgrades to the system. Similar deterministic issues are of concern to the application server developer, who has got to be concerned about how to respond to millions of refrigerators that are about to break down. The developer also has got to figure out how to respond in something approaching real time, not only with the device but with all of the other on-line business-to-business resources that will be needed to complete the repair transaction.

By comparison, the developer of an application server for a web-enabled cell phone or wireless internet- connected PDA has more in common with the developers of those devices than he does the server targeted at managing connected embedded devices.

Where the earlier environment was much more about delivering data from the device to the server, here it is now exactly the opposite. There will be some use of the wireless information appliance to send and receive emails. But the excitement in this community is on how to deliver compelling content from the server to the device.

Maybe you think I'm going off the deep end about this language and definition thing. It does have a serious purpose. It helps me, and you too, I hope, think in new ways about the new computing environment.

Bernard Cole is site leader and editor of iApplianceweb and an independent high technology editorial services consultant. He welcomes your feedback. Call him at 602-288-7257 or send an email to bccole@acm.org.