|
The only type I have experience with.
0 absolutely no support for older versions, use clickonce and force the update when the deployment is completed.
Even when I did get into commercial software a CD (yes it was that long ago) was sent out, the users update or they did not get any support.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Funny how many respond with over 7 years. It is my experience that if you keep an OS or application running to long without upgrading it "increases" the risk of instability. 
In the end you get things like "don't touch that critical software, because we're worried it will break otherwise."
I've worked with critical systems for a good few years, including upgrades.
* don't upgrade right away (AKA don't jump to the new gadgets the moment it comes out)
* don't wait too long with upgrading
* redundancy, redundancy, redundancy (this is the golden rule  )
* have a full functional and independent test system that reflects the real situation
|
|
|
|
|
We still have people running DOS program on 486. 
If you sold software as part of an instrument it must last at least as the hardware. People spending tens of thousands dollars for a system that does not evolve (for example laboratory measure systems, software for opening a bridge). It's hard to explain that they have to upgrade when legacy systems fits their needs.
Remember that Space Shuttle uses 386/16 when pentium where on all desktop..
bye
|
|
|
|
|
Davide Zaccanti wrote: If you sold software as part of an instrument it must last at least as the hardware.
Funny because that is not what Apple does  .
I somewhat agree, but systems that don't evolve probably won't need many upgrades and very little support, right?
|
|
|
|
|
Yes, some little differences from Apple policies...
and again, yes, little improvement on system when a new version is available, but some fix (most related to hardware changes, even on embedded systems) will be delivered.
Think on changes on serial port in last 10..15 years (yes ther are still PC that have on-board a couple of RS232).
When I test a new chair for dentist with imaging software in a real environment the primal told me that the prevoius one last for 25 years...
.. you are right, probably when Apple will build industrial equipment something may change.. you will have to replace a blast furnace if the USB cable fails.
|
|
|
|
|
Think it depends a lot on the type of system you're talking about. For example, if you're talking about a critical system that goes on a space system, the last thing you'd want is something that just came out.. you'd want something that has been proven to be long-term stable and well tested. A similar case would be military equipment that does a specific job. Sometimes those systems take so long to get from concept to being deployed and in use that you wouldn't want to be swapping parts just because a new OS came out. Those also need to be long-term stable and supportable, probably for at least ten years.
On the other hand, on your laptop at home, you probably want something that's more recent and can make the best use of the new hardware you have. You'd probably expect that to be updated more often.
|
|
|
|
|
On the contrary, code doesn't decay with age. If an application requires significant training, is critical to the business, and to make it interesting is subject to government regulations (casino industry is a good example) the fewer changes the better. Maintenance is limited to keeping up with changes in legal requirements, otherwise customers just want it to run as is forever.
I've done casino back office accounting systems that have been in place for over 15 years with virtually no changes. The PC hardware gets an upgrade, if necessary a few tweaks to run under Win7 instead of NT4, but otherwise it stays as is.
The requirements don't change, the software doesn't change, the procedures never change, so where would the "instability" come from? An obsolete OS? Switch to a virtual machine. About the only problem is the programmer retiring and no one to replace him.
|
|
|
|
|
I don't disagree -- far too many companies wait too long to upgrade. Been there and done that numerous times.
But let's look at it from their POV:
For a large (20,000-100,000 person) corporation let's use Windows as an example, going from WinXP to Win7. First there's going to be a wait-n-see period, often until the first service pack is out and vetted. The negative effects of being on the bleeding edge is magnified by the number of people affected. [Personally it can be fun, but professionally I never recommend it.]
As a bare minimum, dozens of applications must be tested to ensure they function in the new environment. The number of applications could be in the thousands, depending upon the nature of the business and the breadth of areas it encompasses. This could take several years before the confidence level in the transition is high enough that management signs off on the upgrade. This process consumes significant resources.
Then comes the actual upgrade, which may take 3 to 6 months (or longer). Plus training for the affected workers, PLUS heavily increased help desk calls until people settle into the new environment. Adding up the costs, which will include new software, personnel to handle the upgrade, lost productivity -- plus upgrades of hardware that worked ok with the old system but not with the new. Plus other software that does not work with the new system and must go through its own upgrade cycle.
The upgrade process easily takes 2-4 years from the time Win7 was released, and certainly cost millions.
For major software, 7 years of lifespan is a minimum. It's too expensive to upgrade more often than that, for something that is a TOOL. No one wants to immediately start again after completing the last upgrade.
So management puts off the next upgrade until they feel they have to. Like when the current OS is going out of support.
Windows is the extreme case, but it demonstrates the problems.
For the vast majority of the companies for whom software is *not* their business, this never ending update cycle is a serious drain on their core business. Looking at it from that POV, it's easy to see why most end users are resistant to change. This doesn't mean it's a good idea, just that it's understandable.
|
|
|
|
|
Quote: assurance your app will continue to run on the platforms / OS
I need to support over 7 years.
I don't think, will OS remains the same (at least) for 3 years
As OS change I need to extend my support too 
if OS is not constant for 3+ years ...Forget about my APP/Software
Find More .Net development tips at : .NET Tips
The only reason people get lost in thought is because it's unfamiliar territory.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
