To: John Wright who wrote (9194 ) 12/16/1998 2:59:00 PM From: joe Read Replies (1) | Respond to of 19080
"I'm currently accessing all my financial apps(located on a central application server), through my Netscape browser. I do not have any of the actual applications loaded on my desktop/laptop. This "Internet" computing approach would seem to offer a much cheaper and more practical approach to enterprise computing than having a thousand copies of bloatware loaded on a thousand machines. How do you see this approach in relation to WIndows and MSFT's stranglehold on the desktop?" It's a good idea, and it depends on what environment it's used. The extreme case is a Walmart situation, where users don't have to know much about the computer. The idea is to reduce TCO (total cost of ownership), and this is done by lowering the maintenance of software. My understanding is that maintenance is a bigger expense than hardware. I can believe it. If everytime a software program gets a patch or an upgrade, and it needs to be done on 5,000 machines, that could save a lot of laborious system maintenance. This solution though is not a function of hardware. It's matter of setting up software control solutions so that make maintenance is easier. Anybody can write programs to do this, including of course MSFT. This leaves out the hardware as the problem and with computers coming down in price so drastically, it doesn't make a lot of sense to build special NC's. Now,if your financial apps are "bloatware", then they use up a lot of your laptop/desktop resources (whether it's efficient code or not). So just imagine a server having many, many,of these resource intensive apps on a server. You'll need a whole lot of them. But, let's say you go with a central applications server...the bigger decision you're making is a very sophisticated system designed with central processing (that's back to the mainframe era in a way). I'm not an expert on these matters, but I just don't see any quick fix for any solution. Big financial programs (such as Oracle) always want to be tweaked and I would think somebody could argue that the system is more flexible if each person has its own high level processing requirements. (If it's simple financial stuff, maybe it's not a big deal). One thing that I think is wrong is to "box in" a very sophisticated/complex database design. IMO, also, these things will evolve into being much simpler entitities, but then again, there will be "new apps" that stretch the computer resources at a new level. There will always be a certain part of the corporate world that will pay the big $$ for the big hardware/software (that nobody else can afford) because it's important enough for them. IMO, this is where Oracle database applications are at right now. Oracle people are always comparing present day MSFT SQL to present day Oracle database. Of course there's no comparison in performance. But MSFT is looking down the road 2 or 3 years from now. Oracle, IMO, has to figure out a way to keep ahead of the curve if they don't want to make their products less expensive in a reasonable way. The JAVA solution, IMO, is flimsy at the moment, and I'm looking for signs of progress. Hope this isn't to confusing, but it's a very complex issue with lots and lots of strategies and opinions. joe