On 7/8/14, 9:57 AM, Dave Salt wrote: > I hear you, and I've heard that same argument many times. But it > makes absolutely no sense to me. If we take the example of a > company that spends 10 million dollars a year to employ 100 > mainframe developers, and that company licenses a tool that > improves productivity by 10%, that's a REAL saving of 1 million > dollars a year no matter how you look at it.
Dave, while I might agree with you in the abstract, I think reality isn't as cooperative as you suggest. The cost of licensing that productivity improvement is a hard, easily-quantifiable expense that must be paid with REAL money. On the other hand, the supposed productivity increase might be hard to measure, quantify and confirm. And the payback might also be in "soft" money ("improved staff morale", for example).
Using your example of a 10% improvement in productivity and assuming an 8-hour workday, your hypothetical tool will allow my current 8 hours of work to be performed in 7 hours and 12 minutes. What do I do with those extra 48 minutes? It's getting towards the end of the day, I might not want to get started on a new task that might take several hours to complete. Maybe I take a bit longer for lunch, add a few minutes to my morning and afternoon coffee breaks. Spend a few more minutes each day checking FB (or IBM-Main). Those extra 48 minutes might improve my morale, but am I getting any more work done? (I've been in meetings where someone tried to justify investing in something based on saving 5 minutes per day per programmer over several dozen programmers. Come on, am I REALLY going to get that much more work done in 5 minutes!)
Allow me an analogy: pretend I run a hospital where we frequently perform a type of surgery that takes eight hours. A supplier offers me some new technology that cuts two hours off that surgery. Do I go for the improvement? Maybe not. I only have one surgeon qualified for the procedure. With the new technology the doc could finish her first surgery of the day in just six hours. But start a second surgery? I don't think so. Two hours into the second surgery her shift ends; she can't just leave the patient in the OR and pick up the next day!
But if the new technology can cut the procedure time in half, heck yes I'll go for it! Now I can get two procedures per day instead of just one. My experience has been that software development tends to have "step functions" somewhat like that surgical analogy. I find that some tasks, particularly debugging, can take a fairly significant amount of time and don't easily lend themselves to "stop-start" execution. Once I get started with a challenging bug hunt I want to finish, because interrupting the process and trying to pick it up later adds overhead. (I might have to spend some time after the interruption reviewing where I left of.) If that new debugger allowed me to finish a daylong bug hunt with 48 minutes to spare, I probably don't want to get started on a new hunt. Maybe I'll take care of some administrative matters, maybe I'll just wander over to a colleague and talk about the Germans beating Brazil!
---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN