All Posts in Policies and Procedures

|In: Industry Commentary|Last Updated:

Microsoft’s posturing over AOL’s “obligation” to open up its proprietary instant messaging protocol was pretty funny; Micron Electronics’ plan to buy Micron Internet Services from their mutual parent Micron Technology is just plain bizarre. But Oracle’s Larry Ellison, topping the rest of the industry, wins in the definitions category by describing his company’s planned $150 (plus monitor) diskless Linux desktop system as a “network computer”.

This is a system that has 64MB of RAM and runs a Unix variant and applications that are installed locally from a CD-ROM. That makes it a network computer? So, I guess, is the iMac. It’s OK. Ellison coined the term “network computer.” He can define it as he pleases.

Speaking of newly coined buzzwords, my recent columns on fat network architectures generated a lot of e-mail and discussion in InfoWorld.com’s forums.

Many respondents disagreed with my proposition that “thin client” has lost all meaning. Several explained what the term “really” means. Regrettably, no two proposed the same definition …

Other exchanges centered on my conclusion that the one thing thin client architectures all share is their need for a fatter network (hence the name). Either missing my inclusion of servers as part of the network or disagreeing with it, they pointed to products such as Citrix and Virtual Network Computing (VNC) that make efficient use of bandwidth. Since fat network is my term, though, I get to define it, and servers are in.

VNC’s advocates in particular were emphatic about the wonders of their thin-client solution, and challenged me to “prove” my assertion that thinner clients require fatter networks. (Answer: See my definition, above.) If the number of capitalized sentences and exclamation points in its proponents’ postings is a gauge, VNC is worth exploring. The number of capitalized sentences and exclamation points also demonstrates the need for a more businesslike approach to product advocacy if the open source movement is to succeed.

In the end, when the flames had died and the smoke finally cleared, only a few points seemed certain regarding thin-client/fat-network computing:

1. Misuse of the term “thin client” has rendered it worthless. It now means “non-Windows desktops.”

2. The only benefits common to all fat network computing solutions are that they’re easier to deploy, stabilize and administer (please, not “administrate”) than applications installed on Windows desktops. Note to all who wrote: They aren’t more stable, only easier to make stable.

3. Browser-based computing benefits end-users – it allows IS to deploy applications that otherwise would be impractical to create at all.

4. The lonely defenders of HTML-only interfaces acknowledged their limitations, instead asserting that end-users need nothing more. Anyone who has filled out an on-screen purchase order knows better, though – HTML lacks basic facilities, like scrolling regions. To make browser-based applications competitive with modern GUIs you need combinations of JavaScript, Java applets and servlets, ASP, Perl … Techniques for maintaining applications built on all this stuff don’t exist.

5. NCs let you build rich user interfaces at the cost of locking down the desktop – sometimes appropriate, sometimes not – but at the cost of incompatible file formats for office applications. Lotus’ latest version of eSuite may change that, though, opening the door for a mixed PC/NC architecture.

Finally, at least within my unscientific sample, advocacy of thin clients and disdain for end-users correlate strongly. In the “real world,” I’m told, the only software end-users install is games and screen savers; IS knows what end-users need to succeed better than the end-users themselves; and what matters is what’s good for the company, not what helps individual employees do their jobs. Then in the next sentence or posting, I hear that IS lacks the resources to take care of every need in the company.

Put it all together and it comes down to the same, tired formula: We won’t do it for you and we won’t let you do it for yourself because we can’t trust you with the tools.

In IS, I guess, we need to update the old proverb thusly: “Give a man a fish and he eats for a day. Teach him to fish and he’ll wreck everything.”

Comments

|In: Organizational Effectiveness|Last Updated:

The Gartner Group calls it the “Zero Latency Enterprise.” Regis McKenna’s new book is titled Real Time.

McDonald’s asks “Want fries with that?”

Although McDonald’s doesn’t dominate the fast food business the way Microsoft dominates the desktop, it’s been stunningly and consistently successful over the years, even though its food is about as good as Microsoft’s software (but don’t push the metaphor, please – I’ve never found a bug in a Big Mac).

One reason for McDonald’s success is that a long time ago it realized that customers don’t care that much about flavor. Speed matters a lot – McDonald’s customers want their food in real time, with zero latency between ordering and eating. They want a clean restaurant to eat in. They don’t want to spend a lot. They want to keep the kids happy.

And oh, by the way … actually eating the food shouldn’t be unpleasant.

I have a lot of respect for Regis McKenna, and have even gained one or two useful insights from the Gartner Group. You can learn a lot from McDonald’s, though, which figured it out long before RM or GG, and you can learn it for a lot less money. Not only that, you can get a Happy Meal in the bargain. For example:

  • Training: McDonald’s doesn’t leave things to chance – Hamburger U may sound funny, but it’s one of the reasons eating at McDonald’s doesn’t lead to unpleasant surprises. Is the training you give your leaders as effective as Hamburger U? If not, why not? The products you deliver are a lot more complicated than a Breakfast Burrito.
  • Leverage: People get all gooey about how the plains Indians used the bison so efficiently, but McDonald’s puts them to shame. When McDonald’s buys a cow, the entire cow – skin, meat, bones and hooves – gets put to productive use. In IS, leverage comes from reuse. Do you create libraries of reusable subroutines or objects? Do you require developers to learn what’s in it?
  • Procedures: There are lots of ways an employee could assemble a Quarter Pounder, but there’s only one way a McDonald’s employee does assemble a Quarter Pounder. When it comes to your core processes – tasks employees do over and over again – do your employees do things by the numbers? Or do they waste time trying to figure out the solutions to problems that have long-since been solved?
  • Willingness to customize: My youngest, Erin, asks for a fish sandwich with “… only the bun, the fish, and the sauce, please. Nothing else.” And that’s what she gets. We spend a lot of time in IS worrying about enterprise-scale problems, but the single biggest difficulty we have is helping small constituencies with specialized needs. Do you know how to help a small group solve a small problem, or do you tell them, “That violates our standards,” or “We won’t solve it for you and we won’t let you buy the tools you need to solve it yourselves, either”?
  • Hiring old people: Think there’s a labor shortage in IS? Compared to the food industry, you have it easy. McDonald’s hires old people. It hires young people. It hires recent immigrants with poor language skills. It even hires the cognitively challenged. It meshes its staffing needs with their scheduling limitations, it helps every new employee find a way to succeed, it pays enough to attract them all, and it makes each McDonald’s a decent place to work, too. Meanwhile, lots of IS shops practice age discrimination, screen out anyone without a computer science degree, and never consider hiring high school students to work the help desk part-time, even though they know PCs better than most “IS professionals” and would kill for the job, besides.
  • Killing bad ideas: The Arch Deluxe didn’t sell, and McDonald’s unceremoniously took it off the market. What really bad idea is still floating around your organization because nobody is willing to put a bullet into it?

And then there’s the issue that started this column, speed. Anything that might slow down a McDonald’s employee in delivering an Egg McMuffin to a customer isn’t there in the first place.

Comments