Ubuntu 17.04: The bittersweet Linux release | Hacker News Reader

Ubuntu 17.04: The bittersweet Linux release

zdnet.com

53 pointsrbanffy59 comments

The article is just a reword of release notes[1]. Half of it is listing of included versions of software. It's not a review, authors didn't even tried to run it in VM. Don't bother reading this "shareware press".

[1] https://wiki.ubuntu.com/ZestyZapus/ReleaseNotes

From the article: I tested Ubuntu 17.04 in an Oracle VirtualBox virtual machine (VM) on my main Linux Mint system. I also installed it on my primary Ubuntu PC. They ran it in both a VM and on bare hardware.

then did no testing of the features they listed to see how well they worked or didn't work or did any real review of anything.

It's not a review. It's news and commentary. Your link doesn't have any of the commentary, which I found useful.

This Ubuntu version has also replaced Unix/Linux's ancient swap partition with a swap file. The net effect should be to make Ubuntu a bit faster in situations where the system is overburdened with applications and has to resort to using drive space in place of memory.

There are reasons a swap file might be preferable to a partition, or maybe to get rid of swap altogether, but it's not this.

actually, do you know why they changed this? I thought that a partition was always superior to a file (aside of the usability issues).

> I thought that a partition was always superior to a file (aside of the usability issues).

Pretty sure you've answered your own question there.

File-based swap will always be a bit slower because of the additional file system layer and possible fragmentation of the file.

On Linux, swap to a file bypasses the filesystem layer when writing and reading, so only fragmentation can be an issue.

And then, on SSDs it is a very minor issue

The release notes don't say why, but I'm going to go for flexibility. Swap to a file is slightly slower than swap to a partition, but if you're in a situation where you're swapping, then you've lost the performance battle already.

I would also guess flexibility. It's certainly a nice property to be able to resize the swap without having to move partitions around or deal with LVM.

I was just about to ask this. I've never heard that claim before.

Maybe so trim works for swap on SSDs?

"This means it supports the AMD Ryzen and Intel Kaby Lake processors. Microsoft, in stark comparison, won't fully support Windows 7 or Windows 8.1 on either high-end CPU"

This is a weird statement - comparing the very latest kernel release on Linux to older windows versions isn't particularly meaningful. If it was 3.2 or 3.4 then sure, but 4.10 is roughly 5 days old...

Linux long-term supported releases, including Ubuntu, have so-called Hardware Enablement, where support for newer hardware is backported to the LTS version during its lifetime.

Is Ubuntu deciding to backport Ryzen / Kaby Lake to the 2009 LTS release of 10.4 (Kernel Version 2.6.32.21)

Because that's how old Windows 7 is. If we want to compare apples to apples.

Windows 7 received patches during its lifetime as well, if you want to really get technical.

I think the implication is that Microsoft still hasn't released a reasonable successor to windows 7 yet (because of telemetry, etc). I'll buy that.

However, even if you grant me that, the statement still kind of falls apart, since you could argue that there is no ubuntu with feature parity with their last gnome based release (even the author of the article runs mint...)

You could allays switch the kernel and see how it goes, otherwise you're comparing oranges with bananas.

I thought that for x86, backwards compatibility was the name of the game. Why do these new processors require OS support?

On a purely technical level they don't per se, however platform features like power management, graphics, ISA extensions etc. needs OS support. The latter part is pretty much just used as leverage by Microsoft and Intel (which also stopped driver support IIRC) to force people into Windows 10.

Re the death of touch-everywhere:

" There was a fantastic universal sense that whatever we were doing was right, that we were winning. . . .

And that, I think, was the handle—that sense of inevitable victory over the forces of Old and Evil. Not in any mean or military sense; we didn’t need that. Our energy would simply prevail. There was no point in fighting—on our side or theirs. We had all the momentum; we were riding the crest of a high and beautiful wave. . . .

So now, less than five years later, you can go up on a steep hill in Las Vegas and look West, and with the right kind of eyes you can almost see the high-water mark—that place where the wave finally broke and rolled back. "

Alternatively, a blind delusion to differing design needs in differing contexts with differing audiences.

Design 101.

Some compromises were made, similar to how Windows sacrificed desktop experience with Metro. I was not a fan. On the other hand, a strong open community contender on mobile would be great.

Good work getting a HST quote to fit to Ubuntu. Never thought I'd live to see such a day

Hunter S. Thompson, Fear and Loathing in Las Vegas, for those who are ignorant as I.

Thanks for this. My first thought was Harry S. Truman.

With the privacy concerns about Windows 10 and Apple's diminishing focus on macOS this was the perfect time for Ubuntu with its desktop focus to thrive. The only missing link was a hardware focus. Every other Linux thread on HN confirms the underlying demand for a fuss free Linux hardware experience.

Inspite of the 'organized hatchet job' on Unity by folks never under any compulsion to use it to explain the 'hatred' it was the best placed to do this. It was by far the most polished and modern looking Linux desktop. It was Ubuntu's USP and its tragic they imploded at a time they could capitalize the most.

I fail to understand all these 'community' spokespeople who have suddenly cropped up on discussion boards who seek to act as 'gatekeepers'. This is not how open source works, people have always done their own thing. No one is in a position to decide what is NIH, fragmentation, duplication of effort or 'good' for the community. Individuals cannot speak for a community.

What is wrong with having Mir or having an alternative to Android on phones, all bankrolled by Shuttleworth? These would be welcome and how some have managed to convert these Ubuntu investments into a huge negative is simply inexplicable. No disinterested user or observer can be so bitter about other projects.

Ubuntu still needs to make money and serve some kind of audience. They have a philosophy and goals as an organization. Writing software that no one will ever use does not serve their organization. If a feature is just going to go to a black hole it is a waste of their time.

There is no real problem for the community, but there is a problem for Ubuntu. Also, I think you could argue that there are, effectively, gatekeepers for the most important projects (Linus/Linux, etc.).

Totally agree about the privacy concerns of Windows 10. I would even say with Sierra I have definite privacy concerns for macOS. The direction is more and more in the walled garden direction for macOS and I do not like it one bit. I ditched Apple/Win for a Linux system even though it meant several weekends of tweaking to get a Linux laptop correct.

An amusing anecdote regarding that, I think a lot of the time cost is just knowing which things work well on Linux. Things are not always "seamless" on Windows or macOS and once you know how to configure a Linux software stack into a productive desktop environment it doesn't cost you that much more than any other OS.

Thing is, most folks make an effort to migrate over and conflate two separate things of learning the hardware and the actually dysfunctional Linux desktop things with finding software that is similar to what they are using on macOS.

Sigh. Windows 8 was the opportunity. That massive incompetent stinkbomb, while simultaneously OSX was being ignored.

Google could have finished off Microsoft with a death knell investment of a relative pittance:

- get LibreOffice to 100% parity and compatibility with office

- pick and polish a desktop environment to a perfect OSX, Windows-7, and Ubuntu interface

- get CrossOffice/WINE to near-100% compatibility

- near-perfect android app use.

Instead... ChromeOS?

Google could have added a multibillion dollar industry for the investment of $100 mil.

Maybe they didn't want to be RedHat? As in, all those projects are open-source and if Google invested in them, they'd be investing in something they can't fully control. Google's goal is to make money, not fight for user freedom, after all.

> - get LibreOffice to 100% parity and compatibility with office

> - get CrossOffice/WINE to near-100% compatibility

> - near-perfect android app use.

> relative pittance:

> investment of $100 mil.

Wait, what?

It was by far the most polished and modern looking Linux desktop.

That is an opinion not shared by everyone. I found it to have questionable design direction. Unity was best on the platform where it was initially released, being screen-size constrained netbooks. It just stops making sense on the desktops everyone has. Hiding menu/scroll bars and the like...

No one is in a position to decide what is NIH, fragmentation, duplication of effort or 'good' for the community. Individuals cannot speak for a community.

It's called sharing one's opinion, and it happens everywhere. Mir had nothing to offer to the larger ecosystem, and neither did Unity. Uniqueness or being a "third option" isn't a virtue all on its own, and it looks like that lesson finally hit home.

True, it goes the other way too. Simply don't use what you don't like. I don't like KDE but many do. Nothing can justify launching a tirade based on personal preferences, so what explains the kind of negativity directed at Unity?

One is free to share one's individual opinion, but one can't speak for the 'community'. For instance just because 'you' personally see no value in Mir, Unity or Ubuntu phone doesn't make it true. Maybe others do. That's just your opinion. Don't pretend to speak for a community. As long as you are not compelled to use it you can choose something else.

Since there is no overarching Linux community no individual or project can declare 'NIH', 'fragmentation', 'duplication of effort' or the particularly measly-mouthed 'good' of the community? Which community? Debian, Redhat, Suse, Mint, Cinammon, Slackware, Gnome, KDE, Elementary, Budgie? Which community are they speaking for? These terms are being used to manipulate and push self serving agendas by bad actors.

>Nothing can justify launching a tirade based on personal preferences

I'm sorry, but why do "tirades" require "justification"? I can talk all day long about how Unity's UI is shitty. I have no obligation, legal, moral, or otherwise to phrase my annoyance in any specific way.

I hated it because it was incredibly user hostile, at least for users like me.

It broke the alt key. I need three things to work: emacs, web and shell. Two of those three need the alt key. Also, it was always flaky under virtualization (due to needing 3d acceleration).

Finally, the last time I tried to use it (16.04, I think), it was full of spyware by default. By the time I removed all the lenses, etc, apt was so confused that the machine couldn't boot. I have been running debian / ubuntu for nearly two decades, and this is the only time this has happened to me.

Fixing these problems would be trivial compared to the effort of implementing unity from scratch. I suspect the core problems with unity were symptoms of structural issues at canonical. Hopefully they'll be fixed now that Shuttleworth is back in charge.

I personally did not like Unity all that much, but I think this was not a wise decision by Ubuntu. Linux is all about the long game. And because of that, tablets and phones are here to stay. It does not make sense to abandon Unity it will be useful if not now then in the future. It's a question of time. Ubuntu should continue with Unity and maybe push it to a community model where it pays for a few developers to keep working on it and the rest of the work is continued by the community.

I want to eventually see a powerful Ubuntu tablet that I can use for my daily use. As it stands I don't trust Android or the manufacturers of my phones enough to actually put important data on there like banking info and passwords.

How is Linux "a long con"? Who's the conman and who's the conned?

It's not a con. I am hijacking the phrase. What I meant was that it's something that takes a long time to payoff.

so, 'long game'?

yea i fixed it.

Am I the only who likes Ubuntu's latest decisions ?

It seems like they are refocusing their efforts on their core, rather than being stretched.

I love their decision. The only problem is I don't really like GNOME. But I do like that Ubuntu is now targeting projects that are being developed by others and they are stopping their phone projects. Lately, I've had issues with Ubuntu in my VM not allowing me to use the tweak tool and some other odd issues I haven't seen.

Then install something else.

Like any of the 30+ window managers that exist.

The software center is two clicks away.

This ignores the fact that the default configuration is where most attention, and so bug fixing work is going to be.

That said, I'm more annoyed that Ubuntu users will now be exposed to GNOME's user-hostile design. I went looking for a good example or two, but some [1] kind soul did a lot of the compiling for me.

I still have no idea who G3's target user is. Worse, I don't think they do either.

[1]: https://wiki.samat.org/GNOMEIsAntiUser

The stuff on that list you linked to seems pretty minor...

(Disclosure: I'm running Ubuntu Gnome 17.04 on my Linux machine)

How many pretty minor things does it take before a pattern is established?

And of course, they're all minor - until they break your workflow, and the developers politely tell you where to insert your bug report when you rightly raise a complaint about it. I'd rather use a UI that has little-to-no danger of a feature I use daily being ripped out from under me because some engineer who doesn't know:

    * What user is using their product
    * What user they are designing their product for
    * How their product is used in real life

..decides to make unilateral decisions with no possibility for reconsideration.

My only problem is that I really wanted an ubuntu phone...

For the first time in years I’m considering Debian. Not because Ubuntu is bad in any way, there just seems to be less difference between it and upstream Debian these days. Many are praising this as a good thing and I don’t disagree. As an end user I’m just thinking “is there enough difference, why not run the original?”.

There really isn't that much.

I've been using Debian for a year and have to Google s lot of questions because I'm a beginner and learn a lot, hack a lot, fuck up s lot. I haven't found a solution that doesn't work on Debian. My searches are 'how to... On ubuntu', 'error this.. Ubuntu '.

From the article, emphasis mine:

> So, in sum while you may not feel compelled to switch to Ubuntu 17.04 on your PC or laptop, on the cloud you might well want to use it. Its improved network speeds look really, really good.

Does anybody use non-LTS Ubuntu releases for servers?

The only time I've come close is when Docker was first released and it required a newer kernel. And that was only for non-prod experimentation.

I ran 11.04 in production for Docker, ~100 servers. If you're in an environment where you know you'll be refreshing servers by the time the LTS drops it'll be fine. Just don't get stuck running on non-LTS when they end of life the release. I also had a ton of issues with degraded network performance while upgrading on a non-LTS release.

I use non-LTS for everything. I don't want once-every-two-years pain. I want to do a test upgrade on a dummy server with the same config (I clone the image in Digital Ocean), I ask questions from the maintainers, then I do the upgrade.

Sometimes I just start from a clean image of the new release if the project has a really solid testing suite. That way I'm not dragging along state.

I installed yesterday, on the surface there don't seem to be any changes. My dns stopped working every time after a minute though there seems to be some issue when dnsmasq is installed.

unity 8 session is a big change https://www.youtube.com/watch?v=-JxXMEFHz54

It doesn't matter since they're going to scrap it in 18.04 for Gnome.

Ok but apart from the wobbly transparent windows, does it ACTUALLY IMPROVE USABILITY? I don't see the point of adding in all the animations, transparencies, etc. if they aren't used to giving you the ability to get your work done with less keystrokes/mouse clicks and to show you more information or show you the info you want in a clearer manner. A good example of my gripe is the window switcher in Unity 7 - it shows window thumbnails (which is useless when you have a dozen black terminals, so you just get a dozen black indistinct rectangles) but doesn't show their names, which is really their only distinguishing feature. If it were built with less shiny graphics, literally as a list of text entries, it would be more useful!

I hit this all the time on ElementaryOS too. I wish there was a simple (under the hood), usable linux distro.

Why do desktops need to run their own DNS servers, anyway? What about all of the other random enabled by default services? Each one is another thing to break, and another attack vector.

Reply on news.ycombinator.com