Proposal: sunset the backports pockets

[Erich Eickmeyer]

Hi Robie,

On Monday, July 19, 2021 5:05:49 AM PDT Robie Basak wrote:
> Dear Ubuntu Developers,
> 
> As far as I am aware, the Ubuntu Backports Team has been inactive for
> years now, and backports requests and uploads just languish in
> Launchpad. Thomas Ward last proposed an effort to revitalise it over two
> years ago, but with no response. I believe he's no longer available to
> contribute now.
> 
> My concern is that the backports documentation and process still exist,
> so users and contributors are being misled into thinking that something
> will happen, when it won't.
> 
> I would be very happy for the backports process to continue, but I think
> it's time to accept that it isn't happening, call a spade a spade, and
> shut the process down and document this properly to stop misleading
> users about it.
> 
> For example, I just handled LP: #1902198 having received an out-of-band
> ping, and looking at https://bugs.launchpad.net/focal-backports there
> are multiple recent requests that we know are not going to be answered
> from a backports pocket perspective.
> 
> Any objections? If you do object, please provide an alternative proposal
> that will mean that users stop getting misled.
> 
> Robie

I'm throwing my Flavor Lead hat on for this one.

I hate this idea.

The Backports have historically been a very difficult process, so difficult that 
at least two flavors (Ubuntu Studio and Kubuntu) have created their own 
backport process and a PPA for those who wish to opt-in. In my opinion, this 
should never have been necessary and those flavors should have had a better way 
to backport. These things don't happen because people are unable to follow the 
process, but because the process is a bottleneck.

Bottlenecks and restrictions are great way to make ideas die because those 
that would use it get fed-up with the extremely high bar that must be met in 
order to do anything. 

Additionally, sunsetting things like this are sure ways to permanently kill 
them. I look at Edubuntu as a prime example: it was sunset, and those that 
have wished to revive it have been denied that simply because "it died before, 
what's to prevent that from happening again?". Somebody made an "Ubuntu 
Education Remix" because they felt they could never revive Edubuntu because 
they'd be denied that ability.

With that, I'd rather see the Backports reformed, not sunset. As with any 
process, if it's not working then perhaps the problem is in the process not 
the idea. Throwing away a good idea because it's not working now is just 
giving up, aka quitting. I was raised never to give up. I didn't give up on 
Ubuntu Studio, and now it's thriving. I believe giving up on Backports would 
be a mistake.

Now, with my OEM (Kubuntu Focus, aka my day job) hat on:

I still hate this idea.

At Kubuntu Focus, we rely on the LTS because those in the scientific and 
artificial intelligence communities rely on packages in external repositories 
that are only made for the LTS. It is astoundingly frustrating when we get 
support tickets asking why XYZ isn't in Ubuntu when a newer version of the 
software that includes a feature they need isn't available to them without 
going to an unsupported repository. That's bad customer service and makes us 
(Kubuntu Focus AND Ubuntu) look bad, because now we're scrambling to provide a 
software solution for a customer rather than just having it for them.

Now, with my Community Council hat on:

How much of the community was aware there even is a backports process? I'd say 
very, very few. To me, that means somebody (I don't know who) dropped the ball 
in the community management aspect of this. As somebody who has a degree in 
this very thing, I find that unacceptable on many levels.

All of that to come to this:

Do not do this. That's basically "thowing the baby out with the bathwater." 
Instead, reform the process. Make the bar a little lower. Instead of 
bottlenecking the process, maybe allow people to actually backport. Not with 
direct repository access, but by lowering the bar in the process. The backport 
process, as it stands, it very, very restrictive. In some ways, it's more 
restrictive than an SRU, which seems backwards to me.


TL;DR: Don't sunset the backports, but reform the process.
-- 
Erich Eickmeyer
Project Leader - Ubuntu Studio
Member - Ubuntu Community Council
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.ubuntu.com/archives/ubuntu-devel/attachments/20210719/6c197a24/attachment.sig>