Driving desktop forward

Being a desktop distro in Linux world is hard: servers and cloud drive not only the development, but the possibility of development (aka money) too. Enterprise desktop customers do exist, but their goals may be completely opposite to regular users. Same time the demand from regular desktop users is higher then ever and competitors are stronger then ever. And by “regular” I mean non IT specialists, non professionals, or non Unix hackers. Their primary products gone free of charge (macOS) or going to be here (win10). So, being free to download isn’t an advantage anymore. Being open source still is, but it alone doesn’t make a whole product.

There should be a path to drive desktop forward while being helpful for regular users. This implies solving real issues like:

  • H/W accelerated video and good battery life
  • Working Chrome/Chromium snap (refresh-awareness, xdg-open, passwords, etc)
  • HiDPI support (without re-configuring GDM, grub, console fonts, etc)
  • Nice store app (maybe a web app from snapcratf.com/store for unified codebase)
  • Better H/W support, faster Nvidia driver updates, standalone GUI driver app, fwupd UI
  • Easier installation with encryption by default (and one password instead of two?)
  • Recovery partition
  • Many more

There so many of those that it’s a hard task for a relatively small team (even to find those issues). But there’re clues in what some other small distros do.

Let’s start with allowing any person to help and influence the movement: open a Patreon account only for Desktop team. Let people to contribute $1/2/5/10 a month to the product they love to make it better. Small (1, 2) amounts are important, but even so with thousands of contributions it will make possible to add a couple more of desktop-focused developers on board. And interact with patrons.

You’re a company and probably thinking about doing it without 3rd party and save on fees, but it is not so simple too. Patreon (or similar platform) will give additional media coverage and discoverability (same as Snap store gives to apps) while letting people to see contributions in one place and single payment.

“Desktop Team Updates” is a good candidate for Patreon updates. There could be other useful perks too. Making monthly poll from patrons is one of those. Not only feedback will be better, but there will be more engagement. Doing yearly poll like it’s now is good, but not good enough (especially with half of questions about servers/vms).

Using more established channels like GitHub/Twitter/Discord/YouTube/Reddit and using them more will be appreciated too. Linux power users feel comfortable in IRC, but that’s not how it works now. You can joke about “Maybe we should post TikTok videos too?” but teenage users will not laugh :slight_smile:

Maybe even mind blowingly buy a couple of macs and surfaces with everyone one the team required to spend a month a year on another platform to look out of the box.

Anyway, the post is quite long already and there many ways/ideas to make desktop experience great. It only needs a little push.

3 Likes
  • H/W accelerated video and good battery life
  • HiDPI support (without re-configuring GDM, grub, console fonts, etc)
  • Better H/W support, faster Nvidia driver updates, standalone GUI driver app, fwupd UI
  • Easier installation with encryption by default (and one password instead of two?)
  • Recovery partition

PopOS is doing it the right way. We should just copy and paste

Nice store app (maybe a web app from snapcratf.com/store for unified codebase)

We really need a better store:

  • Searches in gnome-software give often bad results
  • Snaps are still installing after the progression bar reaches 100%
  • Apps are not well presented
  • Overall dated UI: it feels old and crusty even if it’s actually pretty new
  • No “click and forget” support for flatpaks

I don’t think a web based app is what most people want.
I think we should learn from what Elementary is doing and maybe adopt their software center… wait a minute, PopOS did that as well!
…guys I just realised I should change distro :slight_smile:

2 Likes

I love what Elementary team is doing. Other distros have great ideas and strengths too, just like PopOS. The reason why I’ve created this post here is that I think because of Ubuntu’s wider adoption it can create a wider momentum within community. And Debian is too upstream :slight_smile:

While I appreciate the conversation, can I just warn against turning this into a giant spiralling wishlist of topics. It will go on forever, and no resolution or action will come from it. If you have specific actionable things, that’s great, keep them in separate topics. Creating a great list of things will often result in no forward progress, which I imagine is not what you want to achieve. Some of these are literally just bugs that need fixing. It’s all very well saying “just fix these things”, but there’s human beings needed to achieve these.

4 Likes

…guys I just realised I should change distro

Just kidding of course! :slight_smile: but I seriously like what those guys are doing. I wish we could join forces

1 Like

The concept of micro-targeting donations already has at least one earlier thread already.

There are serious issues that would need to be hammered out if this is a serious idea.

Couple examples:

  • What changes to make to the existing Ubuntu donation page?
  • Who’s in charge of the donated funds? How often do they report to the community? Do they get paid for their work or reimbursed for their overhead?
  • What happens when a targeted bug is fixed but the donated funds were unused? Who keeps track and makes those decisions?
  • What happens to funds when a bug or feature is done by an upstream project instead of by Ubuntu?
  • How many years are funds kept before being returned due to inaction on the target? What happens to funds that cannot be returned?

As somebody who dabbles in business administration, this has potential to be a nightmare: Misunderstandings, hurt feeling, bad press, high overhead, etc…if done poorly.

Elementary and other projects use Patreon without financial reports. People not necessary want to see a detailed records on which buck fixed which bug. Platforms like OpenCollective provide simple reports on spent money and successfully fund well known projects like Webpack/Babel/etc, but only few look at these reports (I remember doing this couple of times). “It isn’t about money, but about sending a message” – more specifically, current donation page shows $15 from the beginning (which is too much) and it’s also one-way transfer without feedback. One-time donations are pointless too: Ubuntu is not a fixed time project, it’s a process potentially infinite in time. As someone from GNOME team said on GUADEC that “people started creating issues and posting comments after migrating to GitLab”. That was exactly what I was saying in the first post: less enterprise administration and mail lists, more “normal” communication tools, let people contribute on Patreon monthly, hire additional desktop-only developers (Desktop Amplification Team, DAT) on these money, let DAT communicate with patrons about what issues should get focus (RFC like process?). DAT may create a discord/spectrum, actively use twitter/reddit, make a couple of posts a month like “Desktop Team Updates” and Daniel (gnome-perf) or Didier (zfs) blog posts, upload a video like Alan does on YouTube (maybe stream), make it part of their job (people pay). This will be a start. Improvise, adapt, overcome. If it doesn’t go well at the end, just close the campaign and use remaining money for Ubuntu anyhow.

1 Like

I think in 14.04 the store had payments? I think I bought Numix theme and icons. I think even bastion was there.

It would be cool if snapstore supported something like itch.io payment model, specially for tiny indie games.

Other thing it would be cool if there was a proper way to pack music on snapstore.

1 Like

You’re right, the “14.04” Ubuntu Software Center did indeed have paid app support. It wasn’t a tremendous success for a number of reasons.

  • Minimum payments per application was too high (~4 GBP)
  • Developers had to build debs - which are hard, especially for people who are just trying to bring their application to Linux, and don’t want / need to learn all the intricacies of debian packaging
  • All package updates were gated by a (very small, and massively overloaded) team of human reviewers, which led to delays, and constant back-and-forth between review team and developers to refine the debian packaging
  • Software needed to be re-built for each release of Ubuntu. So if a developer depended on a library which changed name or version in a new release, they had to rev the packaging going through all the hoops again, above
  • There was no option for In App Purchases, Downloadable Content, Pay What You Want or crypto payment

Developers found it hard to publish software, they had to set prices people wouldn’t pay, and were then signed up to maintain multiple revisions of their application for different Ubuntu releases. For these, and a bunch of other reasons, the USC paid apps support was not a success.

Cut forward to Ubuntu Phone. We had a paid app store model there too. A small number of applications were published as “Click” packages (the spiritual forerunner of Snap packages) but had many of the same issues as the desktop store. In addition, the market was very small, given the thousands of devices sold, limited markets, and new development platform, it wasn’t a success.

So now lets look at today. We have solved many of the issues above. Making a snap is easier than making debs, they are installable on multiple releases, and most of the time require no initial human review, and no review for each upload. These are compelling advantages over the old desktop store model, moving us towards being able to have paid apps in the Snap Store.

There are still some issues however. Fundamentally, the majority of software these days isn’t actually sold in one-off purchases. Users are typically more inclined to buy subscription services. Think of Spotify, Slack, Skype - some of the most popular applications in the store, none are gated by a one-time payment. The obvious exception to this is games, which often have a single payment, with optional DLC or IAP. Games have plenty of places to be seen though, with Steam, GoG, Itch, Epic and others.

While there is some popularity among indie developers for PWYW (Pay What You Want), it’s by no means universal, and I’ve seen no evidence it’s tremendously successful except for a few specific examples. There are indeed other distributions which have PWYW models for their developers, and perhaps due to the small audience, or lack of wide app availability, despite the press, they’re not super financially successful.

Crypto currencies are an option for uber nerds, but not for the average joe, typically.

There’s also the overhead of running the thing, dealing with refunds, complaints, failed purchases etc. No online store seems to do this particularly well, and we’d need to really do it in a way that’s profitable. There’s no point us having paid apps in the store, if only a tiny percentage of them actually result in completes sales, which generates revenue for the developer, and us.

There’s plenty of other issues along the way, making this hard. I certainly haven’t covered them all.

Maybe one day we will implement one or some ways to do it, I’d certainly like that to be the case, and do personally add it to the internal wishlist/backlog when I get a chance. Implementing one-time purchase, PWYW and crpto seems like a panacea, but it’s a lot of work, for potentially little profit. So we’ll have to wait until it’s determined to be worthwhile, I think.

5 Likes

just to give a ball-park number here IIRC the average review time was multiple months … (snaps are usually auto-reviewed in minutes (manual reviews when using a super-privileged interface take a little more time))

1 Like

PWYW may be as bad as it may be good. For example, looking at “Donate $10” message at website every time you’re downloading an ISO and have already donated is very annoying. Reminds me a South Park episode where Randy Marsh was persistently asked to donate every time he was shopping. That’s what my original post was about: doing something isn’t enough, you also need to use tools right. So, an integrated system into Snap Store, that remembers my donations and doesn’t show excess messages, gives me overview of my contributions, allow monthly payments, etc. – that’s a completely different thing.

1 Like

I very much like the idea. The points I raise are NOT to throw cold water upon it, but to help avoid some of the more obvious shoals. To help guide it toward feasibility.

Unfortunately what I suspect will show up in some of the more-sensationalist press is “Canonical Admits That It Tracks You!! We Knew It!” We all know that’s highly inaccurate and quite silly, but that’s what some loud voices will say anyway. And it will kick off a storm.

I remember the last time that kind of storm happened, years ago. AskUbuntu and UbuntuForums were deluged with haters and ranters and assorted crazies for a week, and they kept trickling in for months.

Upshot is that it might be worth looking at a third-party (new or existing) to handle the administrative overhead, adjudication, and the contributor profiles.

I just read what you said about turning this into a giant wishlist but I wanna mention a great way to achieve more community engagement which is to run a Matrix server instead of using IRC. It is literally becoming the next IRC and just yesterday Automattic invested a good amount in it. Has a bright future, very cheap to host, fully open source, and I really want to chat with the Desktop team using the Riot client :). Please consider it if you haven’t

Edit: @popey you actually packaged the Riot client lol who better to explain how great it is :wink:

PWYW doesn’t require you to pay anything. Logging in may be needed only in case if you want to contribute, additionally providing with saved payment info, subscription cancellation interface, favorite snaps, etc. Digital warriors may still anonymously use the donation page. But I doubt that it’s possible to make a convenient system targeting non-technical users without personalized experience. Especially when you’re contributing $1 monthly into 15 different apps.