stormbind,
I agree that there may be marketing or simple human issues that are unforseen that could make this a less viable option. And definitely the winterm model is not for every user's desktop. But remember--our rollout was for 300 machines that do 1 thing and 1 thing only--Order Entry. I'm not advocating that this is a viable option for the accounting department or sales or maybe even most users. But in our organization, of roughly 1000 pc's, I'd guess that over half are candidates for winterm/citrix. The majority of the remaining PC's are here in the main office, so the prime advantage of winterms--remote support--is somewhat moot for nearly half of our pc's, so they'll likely not be converted.
IBM terminals lost favour because they weren't scalable: They lacked sufficient processing power to meet next-generation software expectations.
I'm not sure I understand this point--the terminals, for all intents and purposes had no processing power and weren't meant to. The apps themselves were typically on mainframes, mini's, or AS/400's, which were very scalable.
Yes, the terminal front-end client apps (the familiar green- or amber-screen apps) lost to both the flash and actual increased functionality of a gui. Neither of these is really an issue with a winterm/Citrix setup, at least not that I can see. Again--this setup is essentially identical to a PC, but instead of having your keyboard, mouse, and monitor cord go to the box under your desk, they go thousands of miles away. When a new OS comes out--yes Citrix will have to deal with whatever idiosyncracies may come with that, but the fundamental layout stays the same. It's a pc in every way, shape, and form, with every bit of functionality--with the exception being limited control/access of attached devices--printers, scanners, any usb device, etc.
Soon, Citrix will have to cope with next-generation software expectations which may be more about presentation than functionality.
IMHO, the risk is that intergrated Citrix client/server solutions inherently have more potential bottlenecks. If the 1) server, 2) client, or 3) physical connections fall short of requirements they are going to be offering something that presents itself as being long in the tooth.
I think the issue here is bandwidth if I'm understanding your premise. If the terminals need a higher refresh rate to get a more immediate feel, this is already configurable, and my opinion is that in the future bandwidth will be more plentiful.
IMHO, other risks are: will upgrading the server software force you to upgrade all of the clients (hardware)? Does it remove the possibility of following a slower upgrade path? What are the effects on staff moral if they don't get the latest GUI?
This is a possible item for concern, but I can't see where the issue would be. The winterms are dumb as mud--in a good way--that is, they make a physical connection to send and recieve data. We could walk next door and upgrade one of the citrix boxes from win2k to XP and as soon as it boots up those 20 users now have XP and all the new flash that comes with it. They want the latest upgrade to MapPoint? We throw in the CD here and 2 minutes later they've all got it.
The winterms see essentially 3 kinds of data--key & mouse events, and screenshots. There is no OS-specific logic, nothing but raw data. I'm not sure on the lowest levels the server-side specifics about the format of this data, but unless the entire TCPIP spec changes (and these machines support the upcoming tcpip version) I can't see what the upgrade issues would be, at least for our situation.
I guess for me the bottom line is that feel of 'ownership' Winterms are like renting. It's not really 'your' PC, and that is definitely an issue that the users may have phsycologically, but most of our order-entry people truly don't like or care about pc's, they just want something that works day in and day out.
--Jim