If I ever. . .
If I ever start a poultry-based shop, Flock IT already has my business.
An imploding software industry could be just what customers need. You're going to end up with cheap, specialized software and all kinds of wonderful services to support the code. We're pretty sure that's what executives, researchers and developers told us yesterday, during a Carnegie Mellon West sponsored event at Microsoft's …
Besides the suspicion this is another one of Larry Elison's attempts to convince us to go back to mainframes and dumb terminals (after all as a famous IBM spokesman once said in the 80s - no one wants their own computer).. no the thing that concerns me is how speedily this software will perform.
Will it download a portion to your desktop and run locally, or will it run on their servers? Will it have to compete with net radio and all that other bandwidth clogging crapware that invades corporate networks, or will your internet connection have no bearing on performance?
See to me there was a reason we went away from centralised computing, and it had nothing to do with being convenient for corporate software houses.
It was that chestnut about competing for CPU time on mainframes, it was about having more control over your software, it was about flexibility to run things as you wanted them run.
A lot of that has been taken away already, but the idea that we all fall into neat little boxes has me a touch dubious.
Obviously I can see the benefit to companies like IBM, Oracle and even Microsoft. What I don't see is the benefit to me.
Office 2000 (in fact Office 95 really) works today pretty much the same way as any version except Office 2007 (which is pants - what sadistic muppet "designed" that interface? The exact same program as all the others, except you think it's different because you can't figure out where anything is..).
But the thing is once I've bought it, I get to keep it - I don't really have to continue to pay money to use it. So on-demand services may appear financially beneficial in the short term, but the reality is they suck big time in the long term - unless you really do need your menus slightly re-arranged every 3 months.
If I'm wrong I'll be more than happy to sign up - if it costs 2p a month to run Word or Oracle online, great. But somehow I doubt it.
This post has been deleted by its author
I think no serious business will ever likely go for on-demand services, even public facing web pages are placed on in-house servers. Maybe public data could go outside to lower TCO, but even if I use a couple free internet services - like mailing and webpage - I don't use as permanent storage, because keeping data as close to the place of application as possible is the best way to reduce lags and means tighter security.
I read the recent stories about webmail user who lost years of mailing due to account deletion, and my first question was: Why anyone - with a decent private home PC - keeps so many info on the web? If you travel a lot, or have no home access then it is understandable, but I doubt the person in the news was one of them.
Hardware on the server front and the cost involved with it is enormous, but in-house data management is still better in my eye for security reasons.
"See to me there was a reason we went away from centralised computing, and it had nothing to do with being convenient for corporate software houses."
"It was that chestnut about competing for CPU time on mainframes, it was about having more control over your software, it was about flexibility to run things as you wanted them run."
You've got a very good point about large businesses, they may well want their own computing power to do stuff without having to mix with the hoi polloi.
With small businesses and consumers though, they never ever had access to mainframes. Most pre-internet homes and small offices never had cables to central computing services, so they never had a chance to see its benefits. For them, the only kind of computing they've ever known has been offline, PC-based. Judging by the success of webmail, there's clearly a consumer market for online software, although it's still unclear just how big a market.
With the cost of client computers being so low it doesn't make sense to me to have all the power and functionality at the other end of the network bottleneck.
I believe a client – server model where the applications are downloaded to the client is best. The next time you start the application the checksum is verified against the copy on the server and if there is an update it gets downloaded otherwise you continue to use the version on your computer. With a big enough hard disk the client could store different client profiles and so hot desking would possible.
Also, all date is stored on the server so it is accessible from anywhere and is backed up by the enterprise solution rather than expecting everybody to remember to back up their files.
All the control of server based computer but with the power and functionality of client computing.
On the argument about a bomb hitting the data centre. What about if a bomb hits your office, a single copy of your data gone. Even less extreme, a hard disk failure. In a data centre the drives will be in a RAID configuration, so no issue, but in a desktop there will be no backup. Also, any reliable service provider will have off site backups and also replicated servers that can go live in an instance.