Re: Q: top three things you would like to change if that was easy?
On 2019/04/01 18:42, Lucas Nussbaum wrote:
> Which brings me to a question to the DPL candidates:
>
> We all know that discussions about changing practices or procedures
> inside Debian are long and difficult: there are many very opinionated
> DDs, it's difficult to measure consensus, etc.
>
> Let's imagine for a moment that you have three "full consensus +
> peoplepower" magic cards. Using them, you can propose something, and
> everyone will automatically agree (no need for a long mailing list
> discussion!). Furthermore, people will show up and volunteer to do the
> implementation work according to your plan (if there's a need for some
> implementation in your plan).
Oh man, I like the real-world scenario questions a lot more than the
deeply hypothetical. The problem with the "Genie gives you 3 wishes"
kind of questions is that instinctively, you're going to think of ways
to use those to get more wishes... or in this case, I'm wondering if
it's a better answer listing 3 things that could make consensus + people
power smoother going forward (kind of providing more of the same for the
future) or 3 projects that is sorely wanted/needed in debian now (I
think you intended the latter, so I'm going with that for now).
> So, what are the important things that you will fix or create within
> Debian with those three magic cards? Why? How?
Just off the top of my head, here are 3 things that I would use my
magical FC+PP Cards on:
1. Massively overhauled archive management
-------------------------------------------
There's a whole bunch of what we do now that's either old-fashioned,
weird, obscure or just plain unintuitive. Uploading, processing the NEW
queue, dealing with backports and updates all have great potential for
being more modern and efficient.
But we also need more features. The bikesheds address some of this, but
we have lots of 3rd party projects that want Debian PPAs the same way
that Ubuntu has them on Launchpad.net.
One of the biggest problems I face with my users on a day to day basis
is if they found some 3rd party packages that's perhaps either not in
Debian yet or the Debian version is older than what they require, but
the PPA that exists only exists for Ubuntu and then I either have to fix
their system that they broke by adding that, or find some other
alternative to get that software installed for them. Meanwhile, the
publisher of that software would happily maintain packages for Debian
releases in their PPA too if they had an easy and convenient way to do
this. Additionally, maintaining software in a PPA is a good gateway
towards eventually doing so in the main Debian archives.
I'm not going into too much further detail on this one, because I think
the average DD's imagination is easily provoked on this topic and we all
probably have a list of improvements we'd like to see for archive
management :)
2. Massively overhauled bug tracker system
-------------------------------------------
Our bug tracker is holding us back. Seasoned DDs can probably do
everything without having to put a second thought into it, but I still
have to reference the wiki pages every now and again just to be sure
that I'm sending the right commands to the right email address. I should
be able to spend my time on better things.
There's been a whole bunch of suggestions around the BTS lately, mostly
around building a straight-up new web interface for the BTS. I haven't
spoken against that publicly yet, but I actually think it's a bad idea.
IMHO, what we need is a strong API built on the BTS. From there, it
could be integrated nicely with existing sites such as Salsa, the DDPO
pages, you could make a nice phone APP and CLI utilities to interact
with the BTS without having to do lots of scraping, and ultimately, a
web savvy team could also build a very nice web front end for the BTS.
3. Make APT more competitive
-----------------------------
Snaps, Flatpacks, AppImages, containers, etc all have their place right
now, but they provide some functionality that I think APT + some Debian
tools could take care of even better.
For example, let's say I'm a user on a machine where I don't have root,
how cool would it be if I could, as a normal user type "apt install mc"
and my app gets installed under a namespace in my home directory, is
sandboxed and gets added to my path? Especially if I could have
different namespaces for different releases and have different versions
of packages installed from different sources. That would be *so* much
cooler than using the other type of app images and a lot more efficient.
Debdeltas would also be great in that regard, I know there's been some
work done on that over the years, but it's really quite horrible the
amount of the same data that we download over and over and over. And
they would be especially be nice for edge devices and remote locations
where internet isn't always cheap or highly available where an image is
currently a much better option.
4... Oh wait that's 3 done already... these wishes burn through so fast!
-Jonathan
--
⢀⣴⠾⠻⢶⣦⠀ Jonathan Carter (highvoltage) <jcc>
⣾⠁⢠⠒⠀⣿⡁ Debian Developer - https://wiki.debian.org/highvoltage
⢿⡄⠘⠷⠚⠋ https://debian.org | https://jonathancarter.org
⠈⠳⣄⠀⠀⠀⠀ Be Bold. Be brave. Debian has got your back.
Reply to: