What will software be like in the distant future? One of my main interests (and vocations) for the last 25 years has been artificial intelligence - both "real AI" (creating consciousness) and "practical AI" (expert systems, Baysian networks, neural networks, etc.).
Right now, I am not thinking of AI however, I am thinking of more mainstream software: operating systems, productivity tools, games and entertainment, and software for embedded devices.
The current offerings for commercial software are driven by market economics: companies must continue releasing new versions of software with new features - just fixing bugs is not in most company's economic interests.
I am imagining a distant future when everything "just works" and works together. Think about it in today's context: how many of us would be willing to use software that was feature frozen 5 years ago if it never failed and always performed exactly as expected. I would, for the most part.
I started thinking of this because recently I was tempted to update the Linux distribution on my server. Although I am running an old version of SuSE Linux, I do do regular security updates - but nothing else.
I decided to leave my server alone. This made me think about a future where software was very old and very reliable. Even though as a software developer this is a scaring thought, it makes sense in a lot of ways.
3:41:39 PM
|