not all change is progress
June 27, 2016
Direct download links:
MP3 &
Ogg
00:03:53 News
01:05:32 Devenir Gris
01:15:11 Feedback
01:32:14 ownCloud/Nextcloud
With the initial release of Nextcloud recently dropping, we took the opportunity to spin it up to see whether there were any major differences with the current ownCloud offering. Guess it depends if you count a colour branding change as major… And with one of your Luddites tossing aside such modern frippery, we also asked how much value colour actually adds to a productive computing experience?
00:03:53 NewsSony agrees to pay millions to gamers to settle PS3 Linux debacle
Nextcloud releases ownCloud fork ahead of schedule
Introducing the Nextcloud bug bounty program
PayPal screws over Seafile, backtracks
Samsung enters the cloud fray with Joyent purchase
Samsung considers using Tizen in all products
Acer introduces a Remix OS-powered laptop
StartCom
launches a new service – StartEncrypt
Let’s Encrypt passes the 5 million certificate mark
Defending Our Brand;
Comodo backtracks
Google makes it much easier to use 2FA on your account
“Gtk
4.0 is not Gtk 4”
“Gtk
5.0 is not Gtk 5”
Adios apt and yum? Ubuntu’s snap apps are coming to distros
everywhere
The
universal application distribution mechanism?
On Snappy and Flatpak: business as usual in the Canonical
propaganda department
Maintainers
Matter
Announcing
Flatpak – Next Generation Linux Applications
01:05:32 Devenir Gris
Do you worry about going grey as you get older? Paddy certainly doesn’t, and he’s dropped colour support from all of his regular computing devices — except, sadly, his main Linux desktop. So if anyone knows how to get a usable greyscale configuration working on his ThinkPad X200 with Xubuntu 14.04 and i915 drivers without resorting to hacks like running Compiz, he’ll be eternally grateful.
01:15:11 Feedback
Following Paddy’s gentle rant about the paucity of quality code out there in some projects, several listeners got in touch. Thanks to Campbell Barton, Will and Michael for your feedback.
And Will, along with Florian, clearly still hasn’t given up on Firefox. So thanks both for your thoughts there too.
01:32:14 ownCloud/Nextcloud
Having had our collective conscience pricked by listener Mike Tills, we thought we’d better have a look at the latest incarnation of ownCloud (and the initial release of Nextcloud) before slating this project any further.
Snap is a Snap. Nothing WRONG with Ubuntu’s release. Snap should not be compared to their tablet. Red Hat and Fedora are not making any technical complaints against Snaps. For ONCE I agree with Paddy’s take on this issue. YES Snap is solving a problem that NEEDS to filled! Great idea.
Esteban, are you SURE we talked about the Ubuntu
tablet as I don’t remember that – perhaps it was just
in passing, but I agree snaps and tablet ARE separate
so shouldn’t be merged. GREAT to hear that you share
agreement with paddy on THIS topic, hope you continue
to ENJOY the show.
Jesse
Grist for the mill:
While I’m writing about this before Linux Mint’s official
release announcement, it’s still disturbing that the
sigfile in the primary mirror does not check against the
official key ID for Mint ISOs, unlike the beta sigfile,
which does. Another security breach, poor release
management, or am I just jumping the gun?
Expected key ID: A25BAE09 / Actual key ID: 451BBBF2
This returns a “Can’t check signature: public key not
found” error.
Nice discussions on many topics in this episode.
I am looking forward to seeing how these universal packaging systems evolve. They have the potential to change the way software on Linux is typically distributed and installed and perhaps that could even feed back into how it is written and released. Perhaps it will become more common for projects to maintain several different release channels of varying stability and freshness once package versions are no longer tied to distro versions. I think there will still be a role for distros but that role will evolve.
The Uber/Airbnb analogy brings up a lot of interesting parallels. I would prefer to get my package updates buffered through package maintainers to provide a second set of eyes and make sure new updates are not broken, but I also prefer to stay in a hotel…maybe my feeling is just old fashioned. I respect all of the work that goes into maintaining a stable distro with tons of interdependent packages, but as impressive as that maintenance is if we can get all of the same packages with much less work and only need to use a bit more disk space to do it it is probably worth it.
I just want to point out that the packages in the official Arch repositories closely track official releases of upstream packages. Every time Jesse mentioned Arch, he also referred to beta software in the same paragraph as though Arch was beta software. You can get beta versions of many packages in the AUR.
You missed the definition on sprint there a bit. Sprints are part of the agile process and are basically just small chunks of time that you allot for tasks. It’s a way of scheduling and tracking progress. You made it sound like a weekend hackathon. I will say though because scrums are supposed to be short 1 to 4 weeks they can make it hard to structure long tasks. I think agile ite self does not lead to more bugs, but if you think it will magically give you continuous rolling releases it will, and i find the scrum mindset often opposes internal documentation to an extent that is not safe. Even if your product is constantly changing, having some even dated documentation can save your ass of you lose a team mate to another job and need to get up to speed fast
You asked why anyone would use the monochrome or other color features.
The hint is where it was found, developer options.
As a developer you might have perfect color vision but some of your users might not. This is a tool for the developer to see how their app appears to people who can’t see all colors.
Another reason you might use black and white is power savings. I imagine this isn’t a thing on all devices and turning black and white on through the developer options may not use the same technology but on the Galaxy S5 and above if you put your phone on ultra-power saving mode it makes everything black and white and does some magic to make the gpu work less. In addition to turning this on in ultra-power saving mode you can tweak your settings to have it activate this feature in normal power saving mode (which still lets you use all apps). The first time I did I discovered I liked black and white so much I left it in power saving mode for a week. Better battery life and nostalgic warm fuzzies!
I’m glad that you mentioned the Joyent acquisition by Samsung. Unfortunately the linked article admits what I consider to be the most significant part of the story: Samsung is investing in a cloud infrastructure not powered by Linux. Joyent’s Triton (formerly Smart Data Center) is based on illumos, which is a fork of Open Solaris. You should also know that Triton is open source and can be deployed as your own “private cloud” infrastructure (instead of Open Stack or similar).
When Oracle acquired Sun, they quietly ceased releases of Open Solaris and returned that product to being proprietary. Much of Sun’s operating system engineers quit around that time, and some started illumos as a continuation of the Open Solaris project, but outside of Oracle’s control. Bryan Cantrill gave a rather spirited talk on this called “Fork yeah!”: https://www.youtube.com/watch?v=-zRN7XLCRhc Cantrill and I believe some other Sun refugees work for Joyent.
I became familiar with illumos because an older, wiser fellow sysadmin at work swears by it. illumos includes Solaris goodies like ZFS and DTrace, and has a few notable distributions: omnios for servers, Open Indiana which has a desktop environment, and SmartOS for virtualization. SmartOS is a product of Joyent, and is the operating system which runs Triton and Manta (Joyent’s storage service).
So why would Samsung go to Triton, which is not based on Linux, for cloud infrastructure, when Linux rules the world? Triton has sever modes of virtualization, including running zones (containers pioneered under Solaris, similar to FreeBSD jails), Linux containers (with a docker-compatibly API), and KVM (for any general-purpose hardware virtualization). So you can deploy Linux containers and VMs on Triton.
This is probably the most significant advancement for a non-Windows, non-Linux data center technology in quite a while, and will be pretty interesting to keep an eye on now that they have a lot more resources behind it.
Spell-check takes a holiday! “several” modes…, and “docker-compatible” API.
Right, that was the first topic I wanted to comment on :)
Nathan pretty much everything already.
I may add folks like Bryan Cantrill are pretty deep
into Free software so (see “Fork yeah! (Boy, I love
this guy, he’s so much fun to listen to :D)) if
Samsung would go the Oracle way of “there’s no money
in open source, we’ll close this down” they would
leave.
Bryan Cantrill talked about the origin of Solaris Zones on BSDnow: They’ve seen jails, thought “that’s cool, we want that!” and took the concept quiet a bit further (SysV IPC namespaces, control over which ports of the zone are accessible from outside…).
Yeah, Paddy mixed up sprints and hackathons. And not all folks meeting for hackathons produce rubbish code, just look at OpenBSD. They have at least two of those a year but new code will get polished afterwards before being commited and what gets commited is stuff they’ve discussed face to face and agreed upon.
And things like technical dept is a thing the devops
space addresses. If the people deploying/running the
software have trouble with the way the database needs to
be configured they can talk to the devs and don’t get
dismissed right away. Why? Because smoother deployment
means easier to automate test environments ops can
provide for the devs so the devs can test more
efficient.
I discarded the whole devops thing as just on of those
hipster things for years. But the basic idea is to work
together and improve not only the product but also the
production and increase collaboration between teams to
work more efficient.
And Paddy: Go read “The Phoenix Project” by Gene Kim.
It’s a fun book with a lot of IT horror:
http://www.goodreads.com/book/show/17255186-the-phoenix-project
You’ll probably enjoy parts of it even if you don’t like
all the devopsy stuff in the 2nd half of the story ;)
Wow, not one of my finest works… shouldn’t comment from my phone, where I barely see a paragraph of what I’m writing – even if it’s all black and white ;)
Remembered the 1st thing to comment on: I use a Moto E
with the SDcard as adopted storage. The card certainly
isn’t slow (got me a new one with 30mb/s write speed,
tested on my laptop) but Android still warned me of
decreased performance. I guess it’s the phones
interface.
But no, I don’t see any impact on performance. I don’t
have to worry about internal storage for apps running out
because google decided to split another thing you need
for the playstore to work into two pieces as just as big
as the original 40mb thing anymore.
I’ve seen some hiccups in the beginning but I’m running
cyanogenmod nightlies and the feature was still pretty
new back then.
Great episode that took on a lot of important stuff. First though, mostly because both my wife and I have Moto E 2d Gen LTE devices, the default treatment of the external SD is problematic because the E comes with a maximum of 8G internal storage, of which over 1/2 is used by the O/S and Google apps. As Paddy indicated, you can get back some space by disabling the Google apps that aren’t needed and downgrading them to their smaller shipping versions. Unfortunately the latest from Motorola is that unlike other parts of the world, those of us in the US will NOT be getting the upgrade to Android 6, and so we can’t adopt the external SD as internal storage. Sadly there isn’t a CyanogenMod build out yet that is trouble-free enough to rely on (it still has issues with recognizing the SIM and with the GPS).
On Agile, sprints and other development paradigms there’s no question in my mind that at least half the people who talk about them don’t know what they’re saying, and at lease half of that group who actually try using them are predictably, failures. People in business often operate without any checks on their own personal reality distortion fields and consistently misuse terminology and concepts without any risk of being challenged by underlings who understandably are reluctant to make career limiting moves like contradicting the boss. At the end of the line, as you guys said, you find yourself in the middle of an unmitigated train wreck.
Which brings me to the new packaging technologies, and the concerns I share with at least some of you that “there be dragons there”. Containerization can have huge benefits, whether on the server or desktop, but as we’ve seen in the server world it can (and due to human nature, does) become a workaround for issues that should instead be worked through. More importantly, containers can (and I think will) become a serious attack surface as security updates and other bug fixes don’t get applied to them as they do to the underlying system through normal distro and repository maintenance. On the server side sysadmins should already be keeping a careful watch for that. I’m not convinced that most desktop users will, or can, do the same — especially with the (I think overemphasized) focus on beginning or non-technical users of the Linux desktop.
Phil — if your comment had arrived before we recorded #82 I’d have used your words on agile; as it was, I ended up making the same point (and no doubt less eloquently ;) myself. There is a lot of guff that passes for effective process out there in the dev world.
Comments are now closed.
The content of this website, and that of the podcasts produced by the website owners, is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License.