What I am thinking about software development

[Joel A. Adams]

To some extent, the golden age of software development is over. Starting in the 1960's and going at least through 1999, most companies and corporations spent a large proportion of their IT budgets developing and maintaining their own software. This was a time when the number of software developers expanded from near zero to millions of people in the US and many millions more around the world.
That expansion is now over. I believe that the absolute number of software development positions may not grow and the percentage of the population engaged in software development will drop. Of course, the number of software users will continue to grow until it includes most of the population of the world.
But this does not mean that the opportunities for software developers is limited. The software development has shifted from being in every company in every industry to being located in companies that are in the software industry. Software is now more likely to be developed so that one standard version is sold to many customers.
With this maturing of the software industry, comes a need for better, more sophisticated software and software developers. The Agile trained software engineer is a better, more productive professional than his counterpart of twenty years ago. There will always (well, we better not look farther than twenty years, but...) be a need for the experienced, competent, sophisticated software engineer.

[Joel A. Adams]

That would be very cool. Banking software, I know, uses lots of built-in configuration to allow each bank to have a custom looking system. The person who figures out the right amount of easy configuration to give the users the options they need could be very successful. But that goes back to my point that the best software engineers will be needed by the software companies. And other companies will have less need for talented engineers.

[D. André Dhondt]

Interesting. If software is going to be sold more often as "one standard version is sold to many customers" then we have to reconcile that with the trend of ultra-customization: lean makes it affordable to support many, many products, and to churn out new products quite often. This has had ripple effects from clothing manufacture (Zara), retail outfits like Sears, computer manufacturers (Dell), microbreweries, car building. How will that affect software companies? Lots of out-of-the-box configuration parameters???

[Joel A. Adams]

Devon is a staffing company, and not a very big one now. And, yes, there is my sampling bias. ten years ago we were five times larger and did huge amounts of work for end using companies like pharmaceuticals and banking. Now our customers consist of sometimes smaller software companies and companies that rely heavily on on-line systems. The big pharma's still use consultants, but prefer less experienced, lower priced, programmers. Most of the custom big projects are outsourced to...software companies. A small software team is still an incredibly efficient way to develop software. Mom and Pop shops have always been good at their work. But usually fail in the sales and marketing side of things.

[D. André Dhondt]

Just curious--could this perspective be influenced by sampling bias? I know my response is--I've worked almost my entire career for in-house custom software projects, and my customers have felt their tailored software easily justified the higher Total Cost of Ownership. Still, if we were to talk about custom watches from only 50 years ago, we'd see lots of craftsmen earning a living amongst commodity watch manufacturers. Eventually this generation of craftsmen died out, and all that's left if the big industrial manufacturers. So maybe that's what will happen to us? If we're a 'mom-and-pop' shop, we're going to die out and the reasonably large consulting shops like TW, Valtech, Devon, as well as the COTS vendors, will go on?