Other Open-Source / Linux Letdowns For 2018 From File Creation Time To Flatpaks

[rukur]

And docker fits in where ?

[cybertraveler]

[Creak]

Well, Spotify is on Flathub: https://flathub.org/apps/details/com.spotify.Client

But indeed, it is maintained by the community, while the snap package seems to be the officially supported way.

[Vistaus]

You forget MIR vs Wayland & Upstart vs Systemd. They love throwing money at toying with their own version of key technologies that they will be using alone for a while. Eventually, after having wasted everybody's time, they will turn to the superior solution that everybody else is already using.

I agree, although to be fair: Upstart development started in 2006 while systemd development started in 2010, so there's no way they could've used systemd from 2006 on as it didn't exist yet.

[tildearrow]

Can you guys drop the Mir vs. Wayland thing? Mir has been Wayland for long so you are now saying "Wayland vs. Wayland".

[jo-erlend]

It's better to implement a custom solution only for the least possible number of top layers... And, reuse common layers below, and work together to make them more robust...

For Unity, it's better to just make plugins for gnome and extend layers below if some customization point is missing, or improve performance of layers below... Saves time, makes everybody happy.

Also, for delivering applications, do the same with Flatpak... Just touch top layers, and extend common base to make everybody happy.

Isn't this a very strange argument? Unity is based on the long-existing Compiz, whereas Gnome decided to create a brand new compositor. Wouldn't it be Gnome that should've just built their desktop on existing technologies, as you say?

For Ubuntu, it was easy; we needed a stable, mature and modern desktop by April 2014. Look at the current status of Gnome Shell and tell me Gnome Shell was the best bet in 2010 in order to achieve that goal.

These days, however, close to a decade later, Gnome is about to become a good foundation for a modern desktop. But sometimes, waiting a decade and hoping that someone else will fix the problem, is not the best solution. Sometimes you just need to make the stuff you require.

[computerquip]

I fee like some of this is a bit unfair...
Upstart was made before systemd. It *was* a good idea at the time and it was well implemented. The problem is they didn't market it to any other distributions. When systemd gained adoption with similar or an arguably better feature-set, the point of Upstart began to fade.

Launchpad is still running and isn't that bad of an idea. It just had really bad implementation. I'd say it's less featured than Gitlab to an extent but outside of that, it's solid. There's a simplicity to it I think is lost now in most interfaces today. I wouldn't consider Github an alternative being closed off and all.

Unity is another where it would be solid but had bad implementation. It was made before Gnome and just so happened to look similarly. It was buggy, was less featured, and had less support than Gnome. Most people I know hated it as a result and wanted Gnome instead which was similar but at the least a much better implementation. In addition, there was no marketing to other distributions so most people switching to Ubuntu wanted something more familiar.

As for Bazaar, a lot of people don't know this but it's older than both Git and Mercurial. All three of the tools were released around the March/April timeline of 2005. I can't say as to why Bazaar is so lackluster (as I haven't used it) and why it has had virtually no success in adoption. There is a fork called Breezy which proves to be backwards compatible and avoids the Canonical CLA. Apparently, Bazaar is also considered a GNU tool which I just learned today. It's hard to say what went totally wrong here to the point most developers haven't even heard of it (though that said, I've met developers who haven't heard of Git so....)

As for Mir and Snapcraft, that was pure NIH. Both of those actually set us back due fragmentation in commercial backing. These really hurt and they deserve all the shit they get for it. It's neat that Mir is now a Wayland implementation but it's not so neat it took that long and is probably a solid cause for the lack of Wayland adoption in the early days. Thanks Canonical.

[Weasel]

Not sure if there's much of a point to be going on about a lack of Adobe, AutoDesk and Microsoft software considering that's always been a really irritating "reverse chicken-or-the-egg"-type* problem. In all honesty, as defeatist as it may sound, more effort should probably be put on getting these pieces' Windows versions to run really well under Linux than trying to just get native ports.

*To get these pieces of software requires a lot of users, but to get enough users requires these pieces of software.

It doesn't need users. It needs a sane development environment and userland.

[timrichardson]

I recently used a snap to install nextcloud. It was an incredibly good experience, not just the installation, but the ability to easily customise it (one line and then it is running in https with letsencrypt certificates automatically installed). If this is the future of server installations in Ubuntu, I'm all for it.

[cybertraveler]

I fee like some of this is a bit unfair...
Upstart was made before systemd. It *was* a good idea at the time and it was well implemented. The problem is they didn't market it to any other distributions. When systemd gained adoption with similar or an arguably better feature-set, the point of Upstart began to fade.

Launchpad is still running and isn't that bad of an idea. It just had really bad implementation. I'd say it's less featured than Gitlab to an extent but outside of that, it's solid. There's a simplicity to it I think is lost now in most interfaces today. I wouldn't consider Github an alternative being closed off and all.

Unity is another where it would be solid but had bad implementation. It was made before Gnome and just so happened to look similarly. It was buggy, was less featured, and had less support than Gnome. Most people I know hated it as a result and wanted Gnome instead which was similar but at the least a much better implementation. In addition, there was no marketing to other distributions so most people switching to Ubuntu wanted something more familiar.

As for Bazaar, a lot of people don't know this but it's older than both Git and Mercurial. All three of the tools were released around the March/April timeline of 2005. I can't say as to why Bazaar is so lackluster (as I haven't used it) and why it has had virtually no success in adoption. There is a fork called Breezy which proves to be backwards compatible and avoids the Canonical CLA. Apparently, Bazaar is also considered a GNU tool which I just learned today. It's hard to say what went totally wrong here to the point most developers haven't even heard of it (though that said, I've met developers who haven't heard of Git so....)

As for Mir and Snapcraft, that was pure NIH. Both of those actually set us back due fragmentation in commercial backing. These really hurt and they deserve all the shit they get for it. It's neat that Mir is now a Wayland implementation but it's not so neat it took that long and is probably a solid cause for the lack of Wayland adoption in the early days. Thanks Canonical.

I read your post. I'm not sure where you pointed out the meme was unfair. My take-away from the image is that Canonical have had a long series of market failures with many of their in-house products.

I know launchpad is still going, but it hardly has any traction relative to Git. It does however seem to serve Ubuntu & other Canonical products well.