On 4/16/22 18:49, Johannes Zarl-Zierl wrote:
Am Sonntag, 17. April 2022, 00:01:53 CEST schrieb o lu:
Developers having information like this will eliminate the need for
conversations like this...
Not really. Usage data could play a role in informing part of the discussion
("which plugins are actively used?"), but won't change the big picture:
FWIW, I just jumped in part of the conversation because it was the first
message in my inbox...
But what I was trying to express was that if a significant number of
users are still using, let's say, a plugin, then you probably don't want
to remove that because it is (in my mind) a breaking change. If I
upgrade my app(lication), and all of a sudden something doesn't work,
then I'm pissed (very disappointed for non-English speakers).
1. There is a legacy technology (kipi) that used to be great but was in
decline for many years before it was abandoned by its authors.
Nobody stepped up to rescue the old technology.
You could vendor it and incrementally move it to the new technology.
2. There is a newer technology (purpose) that fits at least the part of the
use-case that is discussed here (export plugins).
3. The new technology has not yet(?) implemented part of the functionality of
the legacy technology [1]
[1] https://phabricator.kde.org/T10525#189748
4. Nobody stepped up to port/implement the missing functionality for the new
technology
Right, but is this a good enough reason to remove the old technology. I
don't know if it (kipi) still works, but if it does, you should leave it
until the functionality is equivalent (or at least competitive in most
cases -- most here is I guess subjective).
So, yes, telemetry data could help us in making an informed choice where to
put the effort. Still, somebody would need to actually do the work.
We do not disagree on that.
Telemetry data would not change the discussion on two of the following three
points:
a. Do applications need to support both technologies even if they are very
similar in scope?
I think it (almost) directly addresses this.
b. Is having two slightly different plugin systems an acceptable user
experience?
If it is transparent to the user, then ok.
c. Is dropping support for the legacy technology an acceptable user
experience?
My retort is, is having a lesser functionality an acceptable user
experience. I would say no for KDE. This would go against my expectations.
Cheers,
Johannes
