Friday, March 24, 2006
  Fat Users
Jaime Cardoso talks - once again - about "network computers".

First of all, I don't quite understand the concept behind the "network computer" designation. The last computer I had that wasn't on a network was back in the 1980's.

Besides, when I read "network computer" I tend to think of something that acts as a compute node for the Network, something more like a grid node than anything else. So, Marketing has it all upside-down. The oh-so-modern "network computer" is, in fact, an X terminal.

And those, my friends, I also had in the 1980's.

The current definition of "network computer", from an architect's point of view (which is from where I stand, want it or not), still doesn't differ much from the 1980's X-term. Central processing, central session management, remote display. You can log off from the machine in your office and continue your work, as you had left it, from any other terminal in the building, in the company, in the network.

Hmmm. I could do this in 1989. I still can do it today.
The benefits are obvious, and many people have spoken about them. My question, the real question, is Why didn't it catch on?

And the real reason is Fat Users.
Fat Users don't like thin clients.
Fat Users like to think they're special, that their machine is brighter-bigger-better than the guys next door's.
Fat Users like to think they have a distinctive activity that makes them need special tools, configured in a special way. Than makes them need extra tools, that aren't even company-standard.
Fat Users like to change the colours on their desktop, because it makes them more productive.
Fat Users like to have a special wallpaper on their desktop, because it makes them more motivated.
Fat Users need to sync their phones' agenda with Outlook (or something), because Fat Users have this oh-so-very-important life, and they can't miss a meeting.
Oh, and Fat Users need their own printer, that only they can print to. And it has to be connected to their machine - that's how they know only they can print to it.

Network computers aren't the answer - they aren't even a question.
Getting companies to standartize on a platform and on a configuration, getting users to stop worrying about detail and concentrating on being productive - those are the main issues today, as they were in the 1980's.

Fix that, and the door is open for the return of the X-term.
The strategy appears to be somewhat different. I totally agree with you - Network Computing and Thin Clients are not the same thing. Thin Clients are in fact (as you put it) glorified VT Terminals, with better graphics, sound, etc.
The idea was sound (but too "idealistic" for some markets) - take your budget and invest in servers and backup solutions for the servers, and proper storage for the applications, etc. Then have your users access those servers/apps with plug-in stateless appliances.

But this was in it´s genesis. Then the dumb-thin-clients have been constantly evolving to meet the "Fat Users" requirements, to the point of PDA Syncing being possible, USB ports for scanners, cameras, etc.

To put it bluntly, todays thin clients are a local harddisk away from being a PC... (Wait - i think some already do so that they can have applications like a citrix client...)

It still is a worthwhile idea - but for how long?
Call it what you'd like, the SunRay is a pretty good idea and, it has a great advantage over the old VTs.
It's not up to the Hardware provider to decide what I want to see in my desktop.
I also never saw a "company standard" regarding I.T. to be anywhere near something smart and, I have absolutely no problem with an user that changes it's Desktop wallpaper, changes it's screensaver or likes to open the spreadsheet on an icon or executing it with a shell. Also, to sync one's contacts between his email client (why not the server I ask) and a mobile is a safeguard that I think It should be encoraged, not restricted so, this devices HAVE TO support all those things and do it on an user level.
Lack of choice loses, dictatorships usualy end in blood, AT&T rotator phones were finished, green leters on a black screen are history and, having to have the active window moving to the front of the screen will be the next to go
I agree with both of you, that the idea is a good one, in many respects, but most implementations are flawed - for one of three reasons:

1. It's not attractive enough to draw people away from the "fat client where i can be a fat user" option

2. Bad architectural design - implementing "thin clients" with "fat software", e.g. using Windows (and derivatives) as a thin client

3. Cost + Installed Base - shifting paradigms implies - in most cases - changing the installed base. Not many companies are big enough to do it by steps and still be cost-effective on step 1, and not all companies can do it in steps at all - banks, for instance, tend to like the "big bang" approach, since they can't support two different branch architectures for 2 or 3 years.

Any of the ways, there is no clear and defined road ahead for something like the triumph of a SunRay architecture, in numbers, in the next couple of years. The same has happened with other thing in the past - COBOL comes to mind - and still we keep being bound by old technology and paradigms because... they just work, worse than the new things, but still just work...
Post a Comment

Links to this post:

Create a Link

<< Home
Juliao Duartenn's thoughts on people and technology

My Photo
Location: Portugal
March 2006 / April 2006 / May 2006 /







Powered by Blogger