>Ed,
>
>What's your advise on this issue? How would you handle it? Beside moving the to SQL Server or some backend database server.

I've gotten the vast majority of my clients to let me reengineer their apps to provide an n-tier architecture when they've been ready for a major change; the ones who haven't largely are very happy with their existing FPDOS solutions running in conjunction with Y2KFOX. Almost inevitably they end up needing some Win32-based functionality, and I tell them reengineer now, so that if they need to move to the Web, or they need stronger security, or they need backend services, it'll be easy to add to the reengineered application. I've got one of these going on now - a freight charter firm that does aircraft charter quoting worldwide; we've an extensive database of airports, facilities, charter firms, aircraft capabilities - the original data design was sound, so the logical data design and base functionality remains stable, but they wanted UI enahncements and better integration with O2K/Word2K and Excel2K, and the ability to deploy in a thin-client environment from a client's office, fax/email and do some scheduling of ground personnel. The business logic behind the routing remains the same - you compute great circle routes, aircraft still have fueld consumption, range, weight and cube limitations, and the basic information about aircraft, airports, aircraft vendors remained the same - we rearchitected to an object model, giving a more flexible UI for routing, and built a multi-route comparator based on being able to add pageframes that house a container object built around a session object and UI behaviors. I was able to give them a more flexible query engine, direct reporting to fax/email/Word, and profit analysis built on Excel spreadsheet templates that had been developed to work with an exported text file, now populated by automation. They ended up spending more money than they'd hoped to spend initially, but get a far superior product, and should the database get to the size that we need more capability than the native file system offers, it's all managed by the Session objects; more likely, they'll want improved security for their data accessed from the road by their sales and ground staff, and if they want to Webify it, I can convert the mid-tier layers to have a DHTML front-end - there's no UI in the business layer except for some maintenance screens that will have to run server-side IAC for backing up/purging/archiving data. Everyone ends up happy - they get an app that does what they need now, with a stronger UI; I have no worries about their outgrowing any given component, and the business rules already existed largely in the FPDOS app; I moved to an OO model, but the math and business goals remain the same, so the work was very straightforward.

I've been out of the sell them the hardware/setup the server/tweak the workstation role for this client for several years; I believe very strongly that my job is to empower them to handle the mundane maintenance tasks, whether bringing in outside service agencies, having them do routine maintenance, setup field offices, take their laptops out on the road and plug them into docking stations when the get back, administer their own NT Server environment with assistance from me or their service organization. It's gratifying to see them operating competently and confidently, taking the precautions of backup, power protection, AV and the like, and if someone tries to sell them on something that smells fishy, I get a call, and can help them with the buzzwords, pricing, interviewing people to provide them the routine service they need - they don't generally need someone at my level, who charges what i charge for an hourly rate, and when they do, they're aware of what they're getting. These people are my friends at this point, as well as clients, and the more they can do competently themselves, the longer I'll have an ongoing relationship with them. And the more they'll recommend my services where appropriate. And any leads that they generate are already aware of what my rates are like, and what types of services I can offer. It's nice being able to walk into a new client, make hardware recommendations, and not have them worry about my getting kickbacks or commissions, or trying to pry a few extra bucks out of their pockets. And they're comfortable with the knowledge that I'll produce a deliverable for analysis and design that they can put out to bid with other developers as well as me; I'm often not the low bid, and even where I don't get the development contract, I'm often asked to consult on the overall architecture and testing of the final product. it's a comfortable, professional relationship, that both sides are comfortable with their ability to walk away from the project without worrying about how the other guy is going to 'get back at you.'

I lose out for many projects; there's a lot of Oracle and Oracle Business Objects done in my area, and I'm not even close to the best person for that work; I know who my peers are and who the competition is - it's tough to be in the VFP market with Jim Booth 20 minutes away! There are serious Oracle and SQL Server houses like MetaGroup and Creeth, Richman and Associates who're better at what they do well than I am, by miles and miles, and there are things I'm better suited to because of domain expertise in logistics, publishing, shipping and similar areas, and some fairly intensive OS and hardware background that isn't commonly encountered in the database developer market, especially at the small to mid-size business operation level. I make a living in spite of several years of health issues, pursue a graduate degree on a part time basis, and still play on UT...

It could be worse.