[ Home | Contents | Search | Post | Reply | Next | Previous | Up ]
Remote Name: 220.127.116.11
That assumes that 'Linux' is the kernel alone. But that is not what I was referring to. Consider all the components that go to make up the average distribution - GNU tools, rpm, Samba, Apache, KDE, GNOME, the GIMP, various shells, mtools, Netscape, apt, etc etc. Many of these may well have been developed using methods/methodologies, but many are created by lone individuals, or groups exchanging only source code. Moreover, there is no single entity responsible for ensuring that all these things work together, but they co-exist quite happily in enterprise-level servers the world over. A complex system without an overarching methodological approach, no?
Re Life Cycle I'll give you that! That was bogus.
Re Management Depends if you think 'Project Leader' qualifies as management. It does in the senses you describe, but this is management after the fact, not management as control, coercion and monitoring. A subtle distinction, but not irrelevant. The point is, even though it might be useful to have a final point of arbitration, shepherding is not necessary (admittedly, this is not directly related to the discussion of methods).
Re Is Open Source a radical new concept etc Mu. The positions put forward are not necessarily antithetical. It is a semantic argument. Define 'radical new concept' then tell me what any conclusion would be worth in real terms (not a lot, I think).
Re Have any studies been done? I would guess no, or else we would have heard about them! As it stands, we have only the examples of individual project failures to adduce evidence from. But to do this without taking into account the abundance of contextual factors (not least of which is the criteria for ultimately judging 'success') results in sophistical arguments - 'Project X sucked. Project X used SSADM. Therefore, SSADM sucks'. FALSE.
How the hell you WOULD conduct a valid study is a tough problem, but we need to start somewhere.. Any ideas??