Definition of open source

[Rick Moen]

Quoting John Cowan (jcowan at reutershealth.com):

> Actually, djb's demise is not a problem for maintainability, since
> he doesn't object to the distribution of patches; his code is like
> QPL-licensed code in that respect.

I'm aware of the notion of maintenance via patches (having heard that
discussion a few times before):  I don't think that's a tenable
development model over the long term.

And, no, I don't think QPL is an apt comparison:  QPL permits
distribution of machine-executable forms of (any) derivative versions.
It also doesn't ban distributing the patches with the canonical source
and a patching script, whereas Dan's terms (typically) forbid any change
to the md5sum.

> The reason djb's stuff isn't Open Source is his refusal to put a license
> on it at all, claiming that copyright law already gives everyone all
> that they need.

The way I explain this is that the default licence inherent in copyright
law (absent a grant to the contrary) is proprietary; it doesn't convey
the right of redistribution or of creation and distribution of
derivative works.

(Quibble:  Explicit licence terms for some DJBware live on Web pages,
e.g. http://cr.yp.to/qmail/dist.html .)