On 18/03/2022 14.01, sebb wrote:
On Fri, 18 Mar 2022 at 12:58, Rich Bowen <[email protected]> wrote:
. I find it
very discouraging to both devs and users if we write "don't use this
software".
+1
Point to a list of open/known issues for sure. But putting "don't use this"
in the readme has long term effects on user trust. So many times over the
years I have heard people cite long-fixed issues as reasons for avoiding a
particular package because it was made part of primary messaging.
This kind of blanket "don't use this" messaging is weird and damaging.
The point is that all software has limitations, and we should be open
about the known issues.
Users can then decide whether or not the limitations affect them.
It's not whether there are limitations or not, it's a matter of wording.
What was written was a very wide opinionated statement that this
software doesn't work well and should not be used in production, which
is patently false and VERY discouraging to read.
If there are *specific* cases where archiving doesn't work (I don't know
of any specifics here), put that in a document, maybe make a
LIMITATIONS.md file and link to that. But putting in the main readme
that people shouldn't use our software is just rude.
We strive for close-to-perfection, sure, but it's not a dichotomy where
software is either good or bad and nowhere in between, and we shouldn't
act as if that is the case.
