Taking the Devil's dollar

Reviewing the comments on the section summaries for my putative Sun Ray book has been interesting.

Arnout Groen hit one of my concerns bang on the head in the first comment on the series:

It smells in here Murphy..

The issue is that I often write about Sun stuff, and here I'm asking a publisher to ask Sun for co-sponorship money on a book about a Sun product. It does smell, and I share Arnout's concern, but don't have a compelling answer other than the obvious one nobody's hiding anything.

Not very satisfying, but either way I have to give up something and I think most readers will understand the trade-offs involved - and lets remember that nobody has agreed to anything yet, so the book may never happen.

The most predictable reaction, of course, came from the people who see thin clients as synonymous with the loss of whatever personal empowerment they get from the PC and therefore fight tooth and nail to stop progress in this direction.

Frequent contributor TonyMcS put this very clearly:

This has been done before by various scamsters, thin clients are dead and gone, but they keep popping up under different names - they still reek however.

There's plenty of vendors that will offer the thin client solution for Windows - usually cheaper than you can build it with your smart displays.

But go ahead Murph, isolate and de-empower the users. Teach them that IT is all powerful and the only purpose of Business IT is for IT to keep all its ducks in a row. When the users point to all the apps and solutions they could be using, just tell them what you've decided they can use for the next 5 years.

You best idea is to try and approach the guy at the top as he/she will know little about IT and will let you wreak havoc on their business.

See also the very lucid restatement of this theme by James@... on August 2nd.

They're both right and wrong here. They're right that SOX and thin client control agendas are a very serious threat to user control of IT, but wrong when they think that the PC empowers users. Balmer to the contrary, today's corporate PC deployments are absurdly constrained through central management of what gets run, what gets loaded, who has access, and when things happen. In looking at this what I'm generally seeing is in fact the opposite of what these guys do: PCs being used far more to monitor and control users than to free them - and help desk people being ostracised from real IT decision making or out-sourced as too limited in scope and too influenced by users to be listened to.

Part of this may be an issue of scale, I tend to see PC deployments at the hundreds or thousands of users scale and not the mom and pop shop level where the PC may indeed still provide some feeling of empowerment relative to not having anything or using a 1970s style time share service -although I would argue the using Linux or the Mac would serve such customers far better.

They're also totally wrong on both costs and software. A well set up Sun Ray system has roughly the same capital cost as a Microsoft client-server system, but spreads those costs out over a much longer stable operating period. More to the point, that Sun Ray system can deliver Windows software services at lower costs than Microosoft's client-server systems can - and concurrently deliver the Unix (Solaris, Linux, BSD) software that Microsoft can't. So there's more software, and it's cheaper.

I thought Richard Flude's Third Way response to Roger Ramjet's guest blog nicely summarized why the solutions Roger and I see aren't that different: distributing the servers (something I generally recommend) eliminates that single point of failure Roger is concerned about while allowing client service via Sun Rays.

Jorwell's comment Why it won't work was off topic but interesting. Here's a key bit worthy of some future exploration:

On the whole I would say OO creates more work rather than reducing effort. If you use methods more firmly grounded in logic rather than wishful thinking then the need to write all those individual methods to maintain data items simply disappears.

A plurality of the other comments raised issues that I've thought about -stuff like change management in complex organizations or transnational data access - but the one big issue that caught me by surprise was the leadership one. I should have known, but didn't.

JPlatt39's comment title When I hear the term "Leadership" I reach for my gun... says it all. The problem here, I think, is the classic chicken and egg one that if you haven't seen leadership at work, you haven't seen it - and you certainly can't teach it to bored managers in "attendance required" seminars. This is both a key issue and a very difficult one - and therefore something I'll have to give a great deal of thought to if this project goes ahead.

But whether it does or not, I want to thank those who contributed thoughts to the discussion and invite further comment here - after all, there would be little point in going ahead without first considering other viewpoints and finding a fit with other people's ideas.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.