Fog Creek Software
Discussion Board

Where is Workstation Development Going?

This is in line with the post today re: where the industry is going.

If you believe the hype, PCs and other devices are destined to become nothing more than display devices, and all workstation level development is absolutely irrelevant because all worthwhile programming logic will be concentrated on servers.

IE, the workstation is nothing, the internet and the web are everything, case closed. It's the kind of reasong that seems to indicate as a corollary that any server developer or admin is a big "stud muffin", inherently deserving of immense salaries and raises, and the developer or support person who concentrates on the  workstation level is inherently inferior, lower grade, and less "worthy". (y'know, that network guy swagger.)

The marketing & positioning problem with this contention (IMO) is that:

Users often want complete control, and some users are always willing to sacrifice some administrative time and spend some money to ensure that they always control their own data.
In line with 'control', many users want more speed, more control, and more security than they can find with distributed or networked applications or 'services'.

Some users just don't want the possibility of anyone else sniffing their data.

Some users are luddites about the internet and will not willingly use anything tied to a wire or a WAN.
No matter how many people claim that networks are everything: pretty GUIs *still* sell applications, and there are still many things that can be done at the workstation level that increase its value to users. For whatever reason, the GUI seems to get *NO* or little attention by product planners, recruiters, or managers these days, even though a bad GUI usually indicates a poorly thought out internal design and can often break a product in the market.

But I'm open on this. I'm trying to determine whether to dig in and embrace .Net, being a hardcore C++ and Delphi workstation person...

Take it away....

Wednesday, January 29, 2003

I have been a "server side" developer for 5 years (before that I did "workstation based 3-D simulations") and I think you are setting up a straw man. The general trend of server side developers being "stud muffins" and commanding "immense salaries" has already died. 

As a server-side engineer, you can maybe make a lot of money if you take on a lot of high-stress responsibility: like, being the DBA AND server-side app developer; in this position a company is paying you to guard over their data, and figure out ways to get at it. However, just being joe flunky .NET or J2EE server-side application programmer will be (and probably alread is) just as bad as being an old workstation developer.

That said, it is always a good idea to learn new stuff. Oddly, I'm getting into client side development, mainly out of sheer boredom with server side development.  I think the real way to become a "stud muffin" and command an "immense salary" is to build your own thing, whether that is server-side or client side, and then sell the shit out of it. 

Wednesday, January 29, 2003

Well, I am old enough to remember when using a computer meant having a dumb terminal on your desk, which was wired to a big box in an air conditioned room tended to by guys in white lab coats.  When the beast went down, everybody sat and did nothing until the beast was restored to life.

Then the PC concept hit - basically everyone had their own "intelligent" terminal or workstation, and they were no longer tied to the beast in the air conditioned cage.  Users rejoiced, as there was no longer a single point of computing failure that would idle an entire department.

Now, the PCs are of sufficient complexity that distributed application deployment and maintenance, along with PC OS maintenance, are getting people to re-think the whole situation.  We may in fact be getting back to having a "dumb" graphics terminal (now called a "PC") on everyone's desk, with apps running over a wire connected to (once again) a beast in an air conditioned cage.  The only difference is now the "mainframe" is called a "network server".  We are slowly heading back toward a system with a single point of failure, only now the users will sit and play Solitare on their PC until the server comes back up.

You see where this is headed.  The only constant is change, and most things in life are cyclical.  Bank on it.

Mitch & Murray (from downtown)
Wednesday, January 29, 2003

Short term, who knows?  Long term, though, capitalism drives entities away from what they want and toward what they need.  People need:

- Fast, responsive, productive, consistent UI (for data input/manipulation).  i.e. not "HTML everywhere"

- The ability for their machine to be trashed by a hardware failure, virus, user indecretion, etc. and not lose time, data, or settings.

- Data and application perminence, redundancy, security, and safety.  Will my data and application _always_ be available, from any machine, regardless of who goes out of business?

- Reasonable platform consistency throughout organization.  Cough, Windows.

- To have small IT departments

In other words, people need thick clients without dealing with installation, configuration, registry corruption, varied security models, RAID drives, backups, etc.

All of this is possible with .NET WinForms deployment.  Unfortunately, right now, it not possible to universally deploy as it is with HTML.

This is what people need.  Hopefully, they'll WANT it soon...

Bill Carlson
Wednesday, January 29, 2003

"IE, the workstation is nothing, the internet and the web are everything, case closed"

Heh, Heh - Curmudgeon, funny little double entendre there in your sentence. Do you see it?

Wednesday, January 29, 2003

I'm with Bill. You want the advantage of both the PC power and the maintenance free webwindow. .NET smart clients, disconnected datasets, auto-update etc. are examples of  steps in this direction, there are many more. We are only starting on this path, but the destination feels good.

This means the workstation/network/server bounderies will fade and you will just be on "the system".

Just me (Sir to you)
Friday, January 31, 2003

I think that applications in the future will fall somewhere in the middle between client-side and server-side.  Certain applications will be server-centered, while some won't.

This is happening today:  Everyone runs an e-mail application, but still perform many tasks locally.

Brent P. Newhall
Monday, February 3, 2003

*  Recent Topics

*  Fog Creek Home