Public infrastructure may be designed to last many decades, but at least around here (Seattle) it only lasts a few years before being 'upgraded' for whatever reason, usually due to growth. On a meaningless and cynical tangent, I think the reason is to keep the construction firms employed.
The same seems to be true of much of our software. The core concepts are solid; but the growth rate of datasets and improvements in technology warrant constant replacement of old infrastructure with new. Whether the improvement is for efficiency or fun, it happens.
Reading further down the article (find the bulleted list), this individual's thought process is quite frightening. The argument seems to be for the removal of 'societal infrastructure software' from the free market, existing instead in a subsidized walled garden of chosen businesses. Municipalities would then have to rely on the products of such organizations.
To me this looks like an argument for the establishment of Looters. I'll grant the author good intentions, but not my agreement. Software, even for infrastructure, needs to continue being produced in a market setting where fast-paced innovation is embraced and praised, rather than feared.
First reaction: I agree. The systems we're building today should be able to support us well into the future.
The problem is that technology has advanced so quickly that what was adequate a few years ago (up to 20, even), is not going to work. We couldn't really predict that we would have taken to this whole technology thing so openly.
...after all, "640K ought to be enough for anybody."
The same seems to be true of much of our software. The core concepts are solid; but the growth rate of datasets and improvements in technology warrant constant replacement of old infrastructure with new. Whether the improvement is for efficiency or fun, it happens.
Reading further down the article (find the bulleted list), this individual's thought process is quite frightening. The argument seems to be for the removal of 'societal infrastructure software' from the free market, existing instead in a subsidized walled garden of chosen businesses. Municipalities would then have to rely on the products of such organizations.
To me this looks like an argument for the establishment of Looters. I'll grant the author good intentions, but not my agreement. Software, even for infrastructure, needs to continue being produced in a market setting where fast-paced innovation is embraced and praised, rather than feared.