Re: Code Reuse (was Re: The Continuous Amnesia Issue)

Lawrence D'Oliveiro <ldo@nz.invalid> writes:

Ben Collver wrote:

The discussion moved at a level as if nobody ever had said or written a
single word about reusability in the past 50 years.

>
That’s because most of that 50 years was spent talking about it, not
actually doing it.

The quoted blog is very vague about what the author thinks is being
ignored/forgotten.

He refers to a NATO conference, and the reports are online:
- http://homepages.cs.ncl.ac.uk/brian.randell/NATO/nato1968.PDF
- http://homepages.cs.ncl.ac.uk/brian.randell/NATO/nato1969.PDF

A cursory search finds only a couple of remarks about re-use.

- First an observation (nato1969.PDF p19) that re-using code in
  different environments would benefit from automation to adapt to new
  environments, which is very much not forgotten, I can think of
  multiple examples that fit: (i) compiler platform definitions, (ii)
  configure scripts (iii) hardware probing/enumeration by OS kernels.

- Second (p29) an observation that modules must be “isolated in an
  envelope” and encourage economical reuse of existing
  constructs. Again, examples are easy to see: (i) shared libraries (ii)
  class and modules in a wide a range of languages (iii) anything with a
  network-addressable API.

I think the problem is pretty much solved now. Open Source has become
the established way to develop most parts of the software stack
(except perhaps the most specialized bits at the top). And code reuse
follows very naturally from the ability to share, modify and
redistribute other people’s code.

Indeed. And we’re a long way into finding out the downsides too -
vulnerabilities arising from dependencies, supply chain attacks, etc.

--
https://www.greenend.org.uk/rjk/