[SystemSafety] Development vs maintenance

[Derek M Jones]

Paul,

>> General software has a short lifetime and it is not cost effective
>> to invest too much up front:
>> https://shape-of-code.coding-guidelines.com/2017/04/20/average-maintenancedevelopment-cost-ratio-is-less-than-one/ 
> 
> Sorry, but this is patently not true. A lot of software has a 
> surprisingly LONG lifetime. I'm aware of VAX-based systems and IBM 
> mainframe systems still in critical use after several decades.

Yes, a lot of software does have a long lifetime.
But, there is an awful lot more that has a very short
lifetime (at least the according to the one dataset I am aware
of that measures such things).

The long time time software gets noticed.
Short lifetime software is not around long enough to get much notice.

> It's not clear to me how anyone could distinguish, precisely, between 
> development and maintenance in any case - if you have a definition I'd 
> be interested to read it.

Yes, this causes problems for data analyzing.

There are people who apportion costs to various work categories.
I am not aware of any data that that includes both development costs and
carefully apportions subsequent costs.

-- 
Derek M. Jones           Software analysis
tel: +44 (0)1252 520667  blog:shape-of-code.coding-guidelines.com