On Thu, Dec 10, 2020 at 05:08:36PM +1300, Joel Wirāmu Pauling wrote: > Streams is currently being used by several very large orgs, and it makes > sense to make it really a RHEL-next+ project, which is effectively what it I wish you didn't speak in absolutes, because I find it, in view of my current personal experience, frankly offensive. To be honest, I'm really glad it works for you. But if you try to tell others that it's all a great idea, I'll ask you not to patronize those who are having different problems that yours. The following narration can be considered a work of fiction, based on a true story. Somewhere Critical For Society, during a global pandemic, we're halfway through a Centos7 to Centos8 migration that's been going on for months and will go on for months more. This is part of a specific long term strategy that uses release jumps as opportunities to review and address obsolescence of a technically, socially, and organizationally complex production ecosystem. Halfway through that careful migration plan, IBM switches the annouced and established 10 years of support for Centos8 to "until next year". Suddenly we realise that our careful migration will likely last longer than the system we're migrating to. We however have the option of retargeting everything on the fly to a hypotetical distribution that doesn't exist yet. In our case, containers are not an option, but a nightmare scenario. We are in fact extremely careful NOT to use containers. If containers ever start to be used within the peculiarities of that environment, then 20 years from now we'll have accumulated countless unique black boxes each with their little custom container setup. At some point one of them will suddenly stop working because of whatever unexepected time_t overflow or something else that was managed by a libc transition 10 years before that however didn't happen in that container, and suddenly there's going to be 20 years of technological debt to catch up. Catch up will have to happen literally immediately, because the level of a river just raised above a critical threshold and the component that failed was part of a production chain that computes which of the towns downstream need evacuating. You and countless others will have wonderful ideas of how containers could be used to make even such a setup so much better. I would consider voicing those ideas, without having made a serious effort to understand the peculiarities of the system that is supposed to adopt them, an act of extremely bad taste. Enrico -- GPG key: 4096R/634F4BD1E7AD5568 2009-05-08 Enrico Zini <enrico@enricozini.org>
Attachment:
signature.asc
Description: PGP signature