not all change is progress
November 10, 2014
Direct download links:
MP3 &
Ogg
LegacyOS – Trisquel 7
IntroIn a move that could be considered admirable or foolhardy depending upon your point of view, one of our number committed to running Fedora 21 beta as his daily driver for the next fortnight. Tune in next time to hear how the impetuous youth got on.
0:03:39 News
Enterprise
Suse enterprise Linux can take your system back in time
(other
interesting tidbits)
Fujitsu Developer Talks Up Btrfs File-System, Declares It
Ready To Use
Facts about Tumbleweed and Factory merging
Canonical rolls out home-grown Ubuntu OpenStack distro
(install
guide)
Ubuntu LXD: Not a Docker replacement, a Docker
enhancement (from the horse’s
mouth, and
some scepticism)
Ubuntu-ish Shorts
MintMenu
forked to become MateMenu
Ubuntu MATE community donations
An Intel-Based Ubuntu Touch Tablet Is Planning To Launch
Soon, but
will it be official?
Mobile & Small Form Factor
Mozilla hopes to challenge Raspbian as RPi OS of
choice
Nextbit Comes Out of Stealth, Demos Service That Lets a
Tablet Pick Up Where a Phone Left Off
Open Sores
Berliner Steuerbehörden wollen wieder MS Office
nutzen
Windows 10 Includes a Linux-Style Package Manager Named
“OneGet”
Security
Researcher Finds Tor Exit Node Adding Malware to
Binaries
GnuPG
2.1 released
Security scorecard finds messaging apps need more
development (and
some criticism)
What, No More World Cup Outlaws?!
Fab and Dan’s
announcement and explanation for the end of Linux
Outlaws
0:46:37 First Impressions
Jesse cast an intrigued eye over the Puppy-based LegacyOS.
And that’s that for First Impressions – there’s only so much that even your hosts can take. We do expect to bring you the occasional short review of other minor distros in the future, but only when we come across one that seems to offer something interesting and worthy of flagging up to a wider audience.
1:01:51 Feedback
A huge thank you to Joshua Krasnow and our regular Monthly Supporters for keeping the lights on. You guys rock :)
Thanks to David A Wheeler and TJ for their emails; allan, Will and Enzro Greenidge for their comments on our website; and to everyone on Twitter and Google Plus – it’s great to see so much engagement and a real sense of community continuing to build around the show.
Suitably chastened by David Wolski’s comment on G+, we spent less time talking about mobile devices this show. Getting the balance right is rather tricky, as mobile is quite clearly the only future of computing for the vast bulk of the world’s population. We’ll continue trying to walk this line.
Joel got in touch to suggest that Joe shouldn’t be judging the potential of the smartwatch category purely on his experiences with the Omate TrueSmart; whilst Sam gently took us to task over our perceived negativity towards Ubuntu.
We got a whole heap of feedback about our piece on GhostBSD last show, and the BSDs in general. Thanks to Brad Alexander, Florian, Nathan D Smith, Zen Floater, Floyd Wallace, Steven Rosenberg and SonOfNed for your thoughts, comments, and points of view. It’s definitely a topic we’ll be returning to in the future.
1:21:36 Trisquel 7
It may be 100% libre and endorsed by the FSF, but is Trisquel a practical option for regular users? We took a look at the recently released version 7 of this distro to find out. Whilst we found it somewhat of a curate’s egg – due to some downright odd decisions and oversights by the devs – it’s generally a tidy looking and functional distro, and would be a great place to start for anyone whose overriding concern is software freedom.
The episode’s reference to the new GNU License and Tutorial Guide (at Copyleft.org) was very timely for me. Having just attended the MeetBSD conference last week, I had been reflecting on the impact that the presence or absence of ‘copyleft’ requirements in the GPL vs. BSD licenses has had on Linux vs. BSD projects respectively.
I have long assumed that the copyleft requirements of the GPL licenses accounted for, in large part, the significantly larger amount of code that has flowed back upstream to the Linux project(s) over the past 20 years. I never fully understood the BSD perspective of, as one developer at MeetBSD put it, “take all the code, do what you want with it, just don’t sue us”.
One insight I gained at the con was that right after the first BSDi BSD386 Release, based on a port of the UC Berkeley’s BSD Net/2 source, AT&T sued both BSDi and UC Berkeley. That lawsuit came very close to extinguishing the whole BSD Open Source movement at the moment of its inception. It also delayed the movement’s progress for years as the lawsuit crawled along.
Now I can certainly see how the BSD Open Source movement may have adopted a strong aversion to litigation once it was finally able to get started.
The Linux movement did have its own big litigation scares with the SCO vs. IBM and SCO vs. Novell lawsuits, but by the time of those lawsuits the Linux project had already been gaining steam for over a decade.
With the GPL ‘Guide’ weighing in at 142 pages of small font reading, I can see why there appears to be a rise _unlicensed_ open source projects. One suspects that the complexity and incompatibility of the GPL(s) and other copyleft licenses has gotten to the point that a growing number of developers feel unable to select one with any confidence, and many are now hoping to avoid legal quandaries by avoiding licenses altogether. If an Open Source contributor chooses a BSD license at a minimum, they at least get some indemnification from liability to help them stay out of court.
Personally, I haven’t lost all faith in the copyleft licenses, but I haven’t yet finished reading the new GPL Guide either. Thanks for mentioning it in the show.
New listener here, so this is something you might have talked about at an earlier stage.
Even if I don’t belive that Ubuntu Touch will be a success, I really do hope that I’m wrong. More alternatives in this area is a good thing and if it’s unix based and open, even better.
But a lot of the things your asking for are already available in Sailfish OS. It’s a full Linux system with a shiny GUI on top. It’s not as open source as one could wish. The GUI and some of the low level stuff is closed. Drivers and such are difficult to open up due to third party manufacturers, but the GUI is getting more and more open as time goes on. I don’t think Jolla have made any promises, but they say they want to be open, and they seems to be working in that direction.
The community on the other hand tends to be very open. Compared to the big two there’s not many apps available, but a lot (most?) of the apps are open source. The community efforts don’t stop at the software level, there’s also some rather serious hardware hacking going on. The biggest and probably best known project TOHKBD reached it’s €55k kickstarter goal in 7 hours.
For me it’s a system that is very much Linux and that felt more or less like home from the begining. You mentioned moving files. Since I like my command line I mostly tend to use scp, but nfs, smb, webdav etc are available. I can mount the phones file system to the the computer or the other way around.
… and it’s systemd based
The TOHKBD hardware slide-keyboard looks awesome, and may finally fulfil Joe’s wish for something like this on a modern smartphone.
Although I was fairly interested when I heard about it, I don’t think Sailfish OS is anywhere near as usable as Android. I know you can run some Android apps but I doubt it will be as stable (as if it could get less stable). Also it seems extremely expensive.
That said I hope it’s a huge success and the major manufacturers realise that there is a clear demand for high end phones with hardware keyboards.
It is true that it is very expensive – I certainly won’t spend this much money on a phone anytime soon. On the other hand, the diy kit for the keyboard was just 60$ in the kickstarter campaign.
The rate at which the campaign reached its goal could perhaps point to enough demand to drive both the price of Jolla phones down (by getting broader adoption), or other manufacturers to do something similar, as you said.
Can’t really comment on the android support. We have two Jollas in the household and none of them have that enabled. If your main concern is support for as many apps as possible there’s only two choices. None of the smaller players (sailfish/ubuntu/firefox/tizen/bb/windows/…) are going to cut it. It’s the same thing on the desktop, linux/macos is not even close windows then it comes to available applications.
I was at a conference recently where I was
introduced to the whole Sailfish thing. It is
EXTREMELY sweet.
Of course, this does not magic up developers
for apps. It was extremely snappy when I
looked at it, but of course, they do not have
it on weak hardware like my Android
phone…
(It’s like this Newell guy is living in a
world where Android does not exist. Oh,
wait…)
There’s been discussion if BTRFS is ready for prime time.
Chris Mason, the main developer of BTRFS says, it is so
(http://www.heise.de/open/meldung/Btrfs-Erfinder-stuft-sein-Linux-Dateisystem-als-stabil-ein-2437356.html?view=zoom;zoom=2).
With a couple of exceptions, though. Like RAID 5 and 6.
With SLES 12 and openSUSE 13.2 these feaures are
de-activated by default. A true luddite would question
this. But how stable is Ext4, XFS or even NTFS? If you
want a rock solid filesystem, go with ZFS. That’d be
another argument for going FreeBSD.
If you’ve not seen it, would recommend a watch of this video I linked to back in show #18: https://www.youtube.com/watch?v=DxZzSifuV4Q
Hello and thank you for an other interesting show
I intended to comment on your views of the GPL back when I listened to episodes 7/8. I can’t say I quite agree with your criticism of it, in two ways: First, I think seeing freedom, as a philosophical concept or a political ideal is not entirely well understood as just the absence of any restrictions, at least not in a world of social interactions. In such a world, it is essential to consider the effects any arrangements has on others. And when placing some restrictions on one agent’s actions allows others to be more free – and that is the idea of the GPL – such restrictions can be justified from the standpoint of freedom.
Now, in a few ways, in an ideal world, such restrictions would not be necessary. First, because anyone using freely available code would recognise their obligations of fairness to contribute back (it is painfully obvious how far that is from our world). And second, because, as in the case of OS X and BSD, the freely licensed code is available to people. But since in a capitalist world it matters whether code is *available*, or whether it is packaged, sold, marketed and lobbyed for by a huge corporation makes all the difference, and will mean that most people will use the proprietary system. Now, you could say that it is their choice not to be free, and you would right to some extend. But since money and market power makes a difference in usability and interoperability with the rest of the capitalist world, the choice isn’t entirely straightforward and fair.
In short, I think in a world dominated by proprietary software, the GPL is essential in preserving software freedom, and that if Linux hadn’t been copylefted, the consequences would have been dire. This, by the way, is where I agree with what you said in the earlier show: the system should be changed.
As for Trisquel, I really like the fact that it is there, and aren’t the shortcomings in hardware support clearly just due to lacking openness of hardware standards?
Hi Daniel – you bring up some interesting points on licensing. As we talked through the topic, the penny finally dropped for me (show #9) why the two sides seem irreconcilable; and it’s for the reason that you articulate – a difference in worldview between those who favour negative and those who favour positive liberties (freedom from vs freedom to). And the two views genuinely are philosophically irreconcilable, which is why there will never be a consensus reached on the licensing topic (since the two sides disagree over what ‘freedom’ is).
It’s simply a matter of OED freedom n. distinct meaning I.1 – to do with liberty vs. distinct meaning I.9 – lack of encumbrance (e.g. restrictions). There’s no positive or negative freedom about the four freedoms – they’re akin to an assertion of human rights. Just that it follows from those ‘rights’ there are things which shouldn’t be done to others.
Although it’s a rather extreme example, calling them negative is rather like saying the right to life is negative because you’ve not got the freedom (I.9) to kill people (yeah, at least some anarchists have wanted that).
But you’re right the difference means the two sides will never be reconciled, it’s just a question of how many change camp over time.
Anyway, as a Trisquel user thanks for the review. I had a good laugh at being called a computing vegan, a fair enough analogy within the limits of how you went about the review. The amusement was because if you hang around on the forum for long enough you’ll discover how many of David Sedaris’ ‘flexatarians’ use Trisquel (i.e. they don’t eat meat when other people are looking) :-).
Actually, Paddy was saying that the GPL way of thinking follows a way of thinking that interprets freedom as positive, i.e. given, freedom to do (the things specified in the four freedoms), whereas the BSD interpretation takes freedom to mean unrestricted choice, including taking code into a proprietary project.
As someone who does philosophy, I can say it would be nice if you could get clear about a concept like freedom by looking up the oed definitions, but it ain’t so.
I am less sure that the equation of these conceptions works like this though. First, these two ways of thinking about freedom don’t exhaust the possibilities and there are other ideas about it that incorporate elements from both – it’s there much harder to say which notion of software freedom actualises e.g. freedom as ¬domination.
And also, the difference between the BSD, GPL and public domain style freedom lies not so much in whether they explicitly allow freedoms to do certain things or remove restrictions, but what their domain is – whether or not they apply recursively so to say. The GPL is a hack of the legal system to extend the freedom generated by open software to users of code that was made available by its developers, whether or not other agents stand in the line of distributing it.
ps. thanks to the Lazarus Firefox extension for saving me from having to rewrite this after a wayward click.
I think you’ve been misled by my mistaken assumption Paddy was rehearsing the same attributions of positive and negative which are a perennial on the Trisquel forum and elsewhere.
The central point was ‘There’s no positive or negative freedom about the four freedoms.’ Nor by extension is there positive or negative freedom in the ‘BSD view.’ Neither is freedom from or freedom to. The notion is merely an artefact of analysis (which is a podcasters’ role) of those two particular schools of thought. When unpacked this is largely in keeping with your later points expanding on the complexities of ‘software freedom.’
Similarly the OED definitions were only supposed to be sufficient to characterising those two positions, not defining them. Much less anything to do with a definition adequate to the rigours of philosophy.
You’re right about the ‘GPL hack.’ It seems to me to have the benefit you describe earlier because copyleft circumvents a ‘tragedy of the commons’ with development time on free software being the common resource. When a company has to give back any changes to free software it redistributes then it is forced to contribute to the commons.
To close I’ll tease you I’m the one here who is doing philosophy. I’m a Trisquel user putting up with all the less convenient methods of doing the things the podcasters mistakenly thought you couldn’t do with free software. Thus I have made my ambulando argument. You, from what you say, read or teach philosophy in some capacity which is not the same thing :-).
Great show! As a self-confessed distro-hopper, I have loved the first-impressions section. Nevertheless, I agree that things were getting a bit bland (the distros, not the podcast!). Given the replication and duplication of work that plagues the linux development community, many of the minor distributions are nothing more than variations on a theme. I look forward to the occasional distro review when something really interesting crops up, or when a new release of a distro is announced.
A minor suggestion. One of the great things about your show (IMHO) is how much I have learnt about the linux world. I think there is great educational value to what you do. Perhaps you could make this a bit more formal with the odd tutorial, glossary, or ‘how does that work’ segment. A recent episode of mintcast explained how init systems worked and this was great. Anyway, just an idea – and thanks for a great show!
the random distro reviews made me think about how productive the work put into the interesting parts in minor distros is. Perhaps you could reflect on the incorporation of neat ideas in small projects in more widely used systems some time.
I agree with Daniel in that some of the innovative aspects of the lessor known distros made for very interesting discoveries. Given that every review takes such a big time commitment, perhaps a modification to the ‘random distro’ selection process could help protect the hosts from sinking a lot of time into evaluating distros with little that is really new.
The ‘random distro’ selection could be used to point you at a distro that you may not have otherwise looked at, but give the evaluator the option of skipping a full blown review of the selected distro if they can’t find anything on the internet that they feel justifies their time to do an installation of it.
Great show guys. Another thought provoking episode.
I’d take issue though with the concept of software freedom. There are not many vegetarians, and even fewer vegans, because that sort of way of living your life restricts your freedom. The freedom to do what you want when you want. And that sort of freedom, the idea of common sense, rule of thumb, “it works good enough” is what I think is essential in Linux and accounts for its success in adoption as the third major Desktop Operating System.
Trisquel sounds like my experiences with NetBSD, and FreeBSD a number of years ago. Sure there are some neat things, but there is enough difficulty in making things work good enough for the computer to be a productive tool that I did not stick with them very long. If I had to boil down the user’s philosophy of Linux for most people it would that it works, good enough, to liberate the computer as a useful tool. That certainly would describe my journey from Mac/Windows user to mostly Linux user.
Its my computer. I paid for it. I want the freedom to run whatever software I like on it. Trisquel therefore is a non-starter because it limits my freedom of action. Do I care if a certain low-level driver is closed source? Not particularly. My computer is a tool, I use for both pleasure and making money.
I do think that with Microsoft chasing the phone platform with Windows 8 and 8.1, Mac OS X doing the same, and now Canonical all hoping to grab a piece of the global adoption of smartphones likely numbering in the billions, the desktop has been neglected.
I’d disagree with Joe that BSD’s orientation has been security only. That’s true of OpenBSD, surely, and they don’t exactly have a lot of low-level software drivers for things like wifi cards for that reason. But FreeBSD has for years used NDIS Wrappers aka “Project Evil” to wrap Windows drivers (currently XP only I think) into FreeBSD. Its how I got my aging Dell Laptop onto my Wifi network. The limitations are as I understand them, no WPA/WPA2, only WEP or unecrypted. But they are there.
And you can already use the compatibility layer to run most software, and my understanding is that FreeBSD recently switched to emulate CentOS instead of Fedora. So that capacity is already there, its just ugly and kludgy. With the addition of Jails (a mini-VM, more desktop friendly in resource usage than server-side Docker) and ZFS snapshots, PC-BSD/FreeBSD would seem to have a “killer app” for people wanting a robust, traditional desktop to create things, enjoy media, and use their computer for intensive tasks requiring hours of focused concentration.
Smart phones are great for, well making and receiving calls, texts, taking photos, browsing the web or some app optimized site. But you would not want to do your taxes on it, or file the monthly books, or balance your checking account, or edit a podcast. Heh.
Leaving aside System D, growing like Kudzu, as more and more distros chase smart phones, I think there is room for a FreeBSD based system to chase the early lessons of Canonical — making your computer easy to use the way you want it, with the software you choose and with a more robust, less error prone OS could win a sizeable devotion from people who need a computer not a phone for intensive work: independent writers, consultants, engineers, scientists, and so on. I certainly don’t put anything sensitive on the cloud, and back up my sensitive data old-school style with rsync and USB hard drives. Not particularly cutting edge but it works, and I not Google or Dropbox, have control over my data.
In short for me, software freedom rests on controlling my own computer, not being locked down into a walled garden where I only get what a corporation decides I get. Its my computer, so I should be able to choose what I want.
Sorry for the rant guys, again I love your show and learned a lot from listening.
I took this course a while ago – the Coursera “Internet
History, Technology, and Security” course at https://www.coursera.org/learn/insidetheinternet
–
This course is on the history of technology and how
technology and networks have developed. This is great for
learning the fundamentals of computer technology.
Coursera now changed it so you can enroll and take it at
any time, The teacher is an open source contributor and
enthusiast, so you get that as well. Strongly recommended
even for people like me with an interest in technology,
but not working in the field professionally.
I would be interested to hear more about what Linux and computing/technology podcasts you guys enjoy. I found Linux Luddites from the promos played on mintCast and Going Linux shortly after I started listening to those shows (the first Linux ones I tried). I stopped listening to Going Linux because it was a little too basic for me. I tried the Ubuntu UK podcast as well but it’s general topics overlapped too much with Mintcast and Linux Luddites and its Ubuntu specific content was not relevant enough to me (I don’t use Ubuntu). I had heard of Linux Outlaws but never listened to it because I had never heard it praised the way you guys did in this episode. I will try out some of the old episodes even though it has ended.
For tangentially related shows, I like Programming Throwdown. There are several other open source, programming and security shows that I listen to on occasion but not every episode (Crypto-gram, FLOSS Weekly, Security Now, ShopTalk, Software Engineering Radio, Sophos Security). Also, since Paddy mentioned BSD podcasts in the last show during the Ghost BSD review I have tried out BSD Now and BSDtalk and found some of the content interesting so far though I have never tried BSD myself.
Hi Will,
My personal linux podcast list looks like this (only
in addition to those you’ve already listed):
Everyday Linux (a new one to me, and more of the
off-topic chatty style than more focussed podcasts,
but good fun)
Linux for the rest of us (with the door-to-door
geek)
Linux Action Show (abbv to LAS, and a big show, I
mean a real big show [watch a few and you’ll get the
joke])
Linux Unplugged (by Chris off off LAS above)
Linux Voice (from the guys who make the magazine –
another British cast)
Sunday Morning Linux Review (a good mix of knowledge
and even some BSD)
And general tech shows;
Android App Addicts
Click (BBC news sub-show – the podcast is different
to the TV version)
Hak5 (video podcast only mainly)
Tech Talk Today (by Chris off of LAS)
Bad Voltage (a personal favourite of mine and which
should probably be mainly in the linux camp, but
isn’t solely linux)
I hope this helps you add some to your listening repertoire. And feel free to share back any others you come across and enjoy.
Thanks, Jesse. I will look into those.
I have been listening to more of the BSD Now
podcasts — very well made show. I found out from
there that one of the hosts also does a security
focused podcast called Tech Snap that I am
interested to try out (it seems a little
technical though).
FLOSS Weekly will occasionally be interesting – I like that they do one project per show, so they can get a bit deeper.
Regarding Paddy’s comment that OpenStack is “flavor of the month” – what other technologies have you called the same, and what was their fate? I know a lot of people called hypervisors a “toy”, for example. =P
The genius of OpenStack (or any, and I shudder to type this, “platform-as-a-service”) is that it places the platform (compute, storage, network) behind an API so that applications can be portable and scalable and resilient by default (and on commodity hardware). I enjoyed this “manifesto” which lays out some of the basic design principles in play here: http://12factor.net/
Can you see how this paradigm is beneficial for information systems as a whole, and not just a buzzword?
Allow me to offer my thoughts on containers from a couple different perspectives.
As a sysadmin, I have a somewhat adversarial relationships with the applications I host. They are not packaged by my distro, or even a respectable upstream. They cannot be expected to behave properly or respect boundaries. They don’t target a common userspace platform. Therefore I cannot trust applications to run together on the same host. In this day and age it is too expensive and ineffecient to dedicate a server to each application, hence the rise of virtualization. I can make the most of my hardware and keep applications separated (as well as a number of other nice features of VMs).
However, if I am virtualizing the full stack, e.g. virtual hardware and a kernel for each application, that is overkill. After all it is not the kernel or hardware which need to be unique to the application, but the userspace (library versions, installed kit, config, etc.). Containers allow me to only virtualize the absolute minimum and thereby remove a bunch of the overhead which comes from administering full-fledged VMs. If my containers are based on snapshots (BTRFS; ZFS), I can even do this in a space-efficient manner.
So I see it totally different than Paddy: rather than being a lazy replacement for shoring up the system, containers are how I keep the system itself strong by means of efficiently segregating applications. BSD Jails did it, Docker is doing it, and there are a number of “thin app” technologies in the Windows market which are doing the same thing.
As a programmer, you have probably read about the various benefits. Containers guarantee runtime behaviors will be identical in development and production because those two environments (and any in between) have an identical userland and compatible kernel (remember Linus’ “we don’t break userspace”). Also, containers lend themselves very nicely to continuos integration, refreshes, and branching (I’ll take buzzword bingo for 200, Alex).
Yeah if you are running your own server using only stock apps (maybe LAMP or a BSD running BIND) containers seem lazy. But when you go outside the lines a bit, they start to look pretty enticing.
No arguments here with the ‘common API interface’ paradigm (to enable application portability across Cloud Platform vendors). It seems to me that OpenStack is more then a ‘flavor of the month’, I see a lot of parallel with the POSIX standard(s) and multi-vendor consortium politics.
I’d like to see OpenStack succeed in providing Cloud Platform customers insurance against vendor lock-in, but only time will tell. I don’t work in the cloud application industry, but I spoke this week with a friend who does work on OpenStack and the situation sounded a lot like the early days of POSIX to me. I hope that OpenStack can prove to be more effective then POSIX has.
“It is critical to note that no vendor I’ve ever
spoken to thinks that OpenStack interoperability
means that you should be able to easily switch
between distributions or OpenStack-based service
providers. Rather, the desire is to ensure that
there’s enough of an interoperability construct
that there can be a viable OpenStack ecosystem –
it’s about the ability of ecosystem vendors to
interoperate with a variety of OpenStack-based
vendors, far more than it is about the user’s
ability to interoperate between OpenStack-based
solutions. To reiterate another point from my
previous research notes: Customers should expect
to be no less locked into an OpenStack-based
vendor/provider than they would into any other
CMP or cloud IaaS provider.”
http://blogs.gartner.com/lydia_leong/2014/05/14/reflections-on-the-openstack-atlanta-summit/
Of course, one quote doesn’t necessarily reflect the true situation. And you could make a case that those firmly in the traditional Linux camp will seek to keep things ‘open’, and simply try to make revenue through services. I guess my cynicism (of which, sadly, more on display in the show we’ve just recorded) is born of long experience of an industry that thrives on introducing vendor lock-in.
Agreed Paddy on the point about vendor lock-in being a constant business strategy in tech. Disappointing when it becomes blatant in open source endeavors.
Thanks for the pointer to the report from Lydia, I’ve generally found the analysis from Gartner to be pretty good over the years. It piqued my interest in OpenStack enough that I did some more surfing. There still appears to be a lot more different constituencies in the OpenStack realm then I was expecting at this point in time. Perhaps inevitable given the complexity of what is being implemented and the challenges of resolving differences through the open source process that is in place for OpenStack.
Some are still expecting OpenStack
application portability across Cloud
Providers, as this proposed presentation for
this year’s OpenStack Summit reflects
(Application Portability Within
the OpenStack Marketplace
) by some folks from Rackspace.
However, I found this panel discussion to be insightful about the reality of the current situation. It seems that portability for OpenStack applications is a lot farther away then I had assumed :-(
I’m not sure exactly which ‘platform’ problems Paddy was implying, I realize that he was just making an off-hand comment during the show. I’m sure that Paddy has some valid issues in mind.
I’d enjoy hearing more about Paddy’s thoughts. Like Nathan I can see some real value from container technology in addressing application dependency and versioning conflicts on hosts, particularly in environments with high application density or rapid application turn-over.
One issue that I have seen repeatedly is that many articles on containers state or imply that they offer security sandboxing for applications, but that’s more an educational problem then a technical one IMHO.
I have been continuing to learn more about systemd and BSD in recent weeks, partly because I have heard BSD mentioned as an escape plan for avoiding systemd. Would the point of such a move just be to avoid Red Hat? From what I can tell so far, BSD is developed more monolithically than Linux, so in a way systemd moves Linux into being more like BSD any way. Also, from some talks I have listened to, it seems like the BSD developers are interested in tackling some of the same issues that systemd does (more centralized control over services, more portable configuration and usage of services, etc.). So a move from systemd/Linux to BSD seems to like ditching Red Hat for the various BSD foundations but does not seem to be a big change philosophically (meaning the Unix philosophy). Am I off the mark with that assessment? I am still very much an init system and BSD noivce.
Will, you ask some good questions. I’ve been looking into them for myself over the last year. It seems that you may have also heard Jordan Hubbard’s from the recent MeetBSD conference. Some of Jordan’s talk did cover service management features found in systemd that were lacking in the BSDs. I was able to attend that conference in person and met Jordan and many other BSD developers while there.
This is how things appear to me at this point; systemd implements vastly more functionality then the set of features that Jordan was advocating for in his talk. In the early days of systemd, Lennart was even promoting systemd as http://www.youtube.com/watch?v=_2aa34Uzr3c.
While systemd does implement some very useful functionality, it appears to me that it will be difficult to use part of systemd without dragging along all the systemd functionality due to module inter-dependencies. I remain skeptical of Lennart’s claims of the independence of the systemd modules, as well as the ‘simple’ replace-ability of (most) of them.
You make a good point Will that the major BSDs release teams bundle a lot more of the core OS along with the kernel in their releases (compared to the Linux release process), the BSD releases are more like a Linux distro. However, I’d be really surprised if in their pursuit of some of the service management features that Jordan was advocating for, if any of the BSDs were to introduce the massive amount of functional change (and module interdependence) that systemd seems to drag along. Minimizing the module and utility interdependence is a big part of the “Unix Philosophy” as I interpret it, although I concede that its applicability is greatly reduced in the modern computing world.
The developers that I met at MeetBSD struck me as a very mature and level headed group who were very proud of the quality and reliability of FreeBSD. They recognize that the BSDs need to continue to evolve to stay relevant, but I get the impression that they are comfortable with a more conservative and reflective approach to change. Some would argue that the evolution of the BSDs is too slow, and I’d agree that it handicaps their adoption in some contexts (e.g. mobile and Desktops), but it does make their future trajectory a lot more predictable then that of Linux. A comforting thing for someone with a Luddite tendencies :-)
it appears that one of the URL links from my post was mangled, should read:
… In the early days of systemd, Lennart was even promoting systemd as “The Core OS”: http://www.youtube.com/watch?v=_2aa34Uzr3c
I just want to add that Linux Outlaws brought me to Linux Luddites this past spring. While it is sad to see any podcaster call it quits, I am glad I found you guys, and I think you are the best Linux podcast in production. Cha-ching. I will help spread the word.
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.