Monday, February 04, 2013

People are the problem can we stop pretending its technology

A friend of mine the other day said an amazing thing
I like coding in C++
I mean, seriously?  The land of friends, of people writing C code and debugging nightmares, had things got that much better, I mean I know there are some good threading libraries now but seriously, C++ is nice?
All of the idiots code in Java, they don't know C++
And there we have the point.  Its not about what technology is best its about the people using them, I'll guarantee that if the idiots were in C++ he'd be having more problems but because they are scared of it he can get more done in C++ safely as for them its terra-incognita.  This for me is why debates around SOAP v REST are pointless and make me quite angry.  People pontificate on 'REST scales better' or something else that doesn't matter 99.99% of the time (as in yes it might, but if something else scales acceptably then its not an issue), its like the 'Assembler is more efficient' bullshit that those of us who dared to code in C will remember.

The worst piece about the technology marketing community, by which I mean analysts and vendors, is the ability to hype something that doesn't matter because its a new technology.  It isn't that this technology has to make things better, hell it can actually make things worse, but all it needs is to have some technical reason why its better than something else.  'Its faster' in a place where that isn't important, 'Its quicker to develop your first solution' but a bitch to maintain. We've heard them all down the years.

So as part of my desire to see Thinking is Dead proven wrong I'd like to start a simple campaign.  Everything an analyst, vendor, consultant or developer tells you that something is 'better' ask the following three simple questions

  1. How does it reduce the support costs
  2. How does it reduce the salary levels of my developers
  3. How does it have a measurable impact on its own to our top or bottom line
This last point is critical.  I've seen some crackers down the year around integration technologies in particular 'We used technology X and shipped $1bn in products, therefore X delivered $1bn in revenue' no it didn't, the only question is if it cost less to develop and support technology X, the best that a technology in integration can hope for is a cost reduction in integration TCO, it will never on its own deliver the value because the value is about the information or transaction it delivers.  If it does that more cheaply then its a cost saving, but its never a revenue generator.

There are places where technology can have a top-line impact but those are very minimal (Predictive Analytics and HPC are about the only two I can name) everywhere else its an enabler for people to deliver value.  So the goal of technology is to make the people work better, the people work more efficiently.  Having a technology that is 5% better than another technology at technology stuff but 10% worse from a people perspective is like comparing getting the horse with being driven to near extinction for native americans.  Sure its a benefit, but it really doesn't outweigh the costs.

No comments: