|
|
Log in / Subscribe / Register

A Year of the Linux Desktop (KDE.News)

[Posted July 4, 2013 by jake]

KDE.News has an interview with Malcom Moore, who is the network manager for Westcliff High School for Girls Academy (WHSG) in the southeast of England. WHSG switched the students to Linux desktops (openSUSE 12.2 and KDE Plasma 4.10) a year ago, and Moore talks about the motivations for the switch, the transition, the reactions to Linux and Plasma, hurdles that were overcome, and so on. By the sounds, the conversion from Windows has been quite successful, though not without its challenges. "The Senior Leadership Team grilled me in two long meetings which was fun! Once you actually take a step back from the misconception that computers = Windows and actually seriously think about it, the pros clearly outweigh the cons. The world is changing very quickly. There is a survey that reports in 2000, 97% of computing devices had Windows installed, but now with tablets and phones, etc., Windows is only on 20% of computing devices, and in the world of big iron, Linux reigns supreme. We specialize in science and engineering and want our students to go on to do great things like start the next Google or collapse the universe at CERN. In those environments, they will certainly need to know Linux."
to post comments

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 12:22 UTC (Fri) by ibukanov (subscriber, #3942) [Link]

Interesting results about GUI:

> Secondly, during our testing, we encouraged students to try both KDE Plasma and GNOME. Plasma was by far the winner in terms of user acceptance. [The final software choice was openSUSE 12.2 and Plasma Desktop 4.10 - Ed].

This matches my family experience when a 7 year old child ended up using mostly KDE after having an option to play with Gnome Shell, Cinnamon, KDE, XFCE and LXDE on an old netbook.

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 13:05 UTC (Fri) by Cato (guest, #7643) [Link] (23 responses)

Great case study and well worth a read - cost savings appear to primarily be admin time, once the initial issues of needing NFS home directories were solved and network upgraded (can't see why LDAP would be very network intensive).

NFS home directories

Posted Jul 5, 2013 14:52 UTC (Fri) by david.a.wheeler (subscriber, #72896) [Link] (22 responses)

I suspect LDAP itself wasn't the primary issue, the issue is that they use "NFS home directories". Which means that when everyone logs in at the same time, all the computers are trying to download their respective home directories simultaneously.

NFS home directories

Posted Jul 5, 2013 15:01 UTC (Fri) by maxiaojun (guest, #91482) [Link] (14 responses)

How to do better?

"NFS home directories" seems to be a common setting.

NFS home directories

Posted Jul 5, 2013 15:43 UTC (Fri) by pboddie (guest, #50784) [Link] (1 responses)

Back in the early 1990s I used a campus-wide computing environment that had huge initial problems with NFS performance. The architecture of the environment apparently emphasised centrally-stored applications (and home directories, I suppose), but in the end I think the administrators realised that they were better off pushing some things onto local disks.

The workstations weren't diskless, as far as I know, and I think that the project was derived from the MIT Athena work (DEC was involved), but the institution concerned seems to have erased all historical knowledge from its Web pages, replacing it with the usual vain corporate look and generic brochure-level experience for "customers" of the institution, so I doubt that it's straightforward to piece together what people did and what people learned from the experience.

NFS home directories

Posted Jul 11, 2013 19:30 UTC (Thu) by cdmiller (guest, #2813) [Link]

NFS does not cause users to "download their home directory".

A poorly configured NFS can slow things down, but I'm guessing a server with a large cheap SATA drive or two trying to handle 60+ folks at once is the real culprit. That said Samba on a Ceph cluster of near line SAS works wonders.

In the 90's I ran a few setups with everything from NFS home dir's, /usr application directories and diskless X station clients. Worked pretty well when set up right. With the smaller software binaries of the time a lot less data was moving around. I keep around an old Pentium 120 with 16MB RAM and DOS/Windows 3.1 installed just for demonstrations. It will boot up and be into Microsoft Word before the students can even log into our Windows 7 lab image. Oh, and I walked uphill for a couple miles both ways to the public school.

NFS home directories

Posted Jul 6, 2013 1:49 UTC (Sat) by raven667 (guest, #5198) [Link] (6 responses)

pam_mkhomedirs and local disk is also a good way to handle a centrally managed roll-out. There is also some network filesystems which cache blocks locally, such as AFS, I think there are also extensions for NFS which do the same but I don't have any experience with them.

I'm not sure how well NFS works on the client these days, last I used it 5+ years ago it was still flaky and problematic with unfixable connection problems that would require the server or the clients to be rebooted. SMB was better behaved. NFS seems to have more usage in the server room with VM servers as the clients are more tightly controlled and fewer and exposing a file-level structure is better than a clustered file system, especially on log-journaled storage.

NFS home directories

Posted Jul 8, 2013 19:33 UTC (Mon) by drag (guest, #31333) [Link]

NFSv4 is a different beast. NFSv3 still sucks, but I can't comment on the flakiness.

NFS home directories

Posted Jul 14, 2013 10:01 UTC (Sun) by thoeme (subscriber, #2871) [Link] (1 responses)

I am running NFS v3 from a lot of test machines to some central workstations (openSuSE12.3, up to date) acting as NFS servers, and so far I am really happy with the stability and robustness of the connections, surviving reboots of the clients and the server workstations.
This in contrast to CIFS (mount.cifs) where a non-responsive (Windows)-Server almost always requires a reboot of the client to get rid of the hanging client. Lazy umount did not work and I have found no option in mount.cifs man pages ("soft" is supposedly the default) which prevents this.

NFS home directories

Posted Jul 14, 2013 11:10 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]

I have a different story. I'm using Amazon EC2 and nodes change their IP address after a reboot. So it's impossible to restore NFS connections. Oh, and forced umount also doesn't work more often than it does.

NFS home directories

Posted Jul 18, 2013 17:21 UTC (Thu) by nix (subscriber, #2304) [Link] (2 responses)

I've had my home directory NFSv3-mounted from a headless server for the entire time I've been running Linux, since 1997. I've never had any real problems with it.

Periodically I consider going to NFSv4, but information on it is so conflicty that I haven't tried yet (e.g. some sites say that only one mountpoint is possible, which I *know* is wrong, but there is no information on what exactly differs these days between an NFSv3 and v4 setup, and if I get it wrong my home network falls apart: I should do some experimentation first really, if I can find the time).

I haven't yet found anything that's a good competitor to NFS. Either they require a complete re-mkfs just because you want to access your filesystem from another machine, or they're even more horrendously non-POSIXy than NFS (e.g. Samba).

... and meanwhile the desktop environments depend on inotify more and more: apparently anyone running NFS $HOME is just an ignorable edge case, while the kernel developers consider anyone using inotify an ignorable edge case, even when that includes major desktop environments: I do not see how this conflict can be resolved.

NFS home directories

Posted Jul 21, 2013 19:10 UTC (Sun) by kleptog (subscriber, #1183) [Link] (1 responses)

ISTM the solution is to add inotify to the NFS protocol. I understand there might be some technical issues in Linux to allow the filesystem to pass the message on, but it seems doable.

To me, being able to add notifications to inodes is a useful feature and don't particularly understand why network filesystems don't support it yet.

SMB with its opportunistic locking could emulate a part of it.

NFS home directories

Posted Jul 27, 2013 22:17 UTC (Sat) by nix (subscriber, #2304) [Link]

Quite. That's the only way to fix the problem. Unfortunately no network filesystem hackers appear interested in it (and, as usual, I haven't found the time, nor am I ever likely to).

Caches in NFS home directories

Posted Jul 8, 2013 4:55 UTC (Mon) by eru (subscriber, #2753) [Link] (4 responses)

By default, Linux DE:s put way too much things under home directories that should never be there. Like image caches and web browser caches. These should always be on local disk, since it is no problem if they disappear, and caching with a network mounted directory makes no sense at all. I wonder if using the NFS home directories for such "junk" was part of the problem.

I know many DE:s and other GUI programs can be configured to behave otherwise, but it is not the default! (One of my pet peeves with Linux programs, and seems unlikely to get fixed, because most open-source developers do not work in environments with NFS-mounted homes).

When used through text based sessions (like most Linux users in my organization), NFS home directories are no problem at all, as long as the server and network is reliable.

Caches in NFS home directories

Posted Jul 8, 2013 6:14 UTC (Mon) by niner (guest, #26151) [Link] (2 responses)

FWIW KDE puts those temp files in /tmp/kde-username and /var/tmp/kde-username. The home directory only contains symlinks to those. But many other programs misbehave in this regard like Firefox and on Windows Internet Explorer.

Caches in NFS home directories

Posted Jul 8, 2013 19:37 UTC (Mon) by drag (guest, #31333) [Link]

The browser is the only thing I really have issues with. Different browsers have different settings.

What I do is setup my user to use /nfs/home/$USER for the main home and then have /home/$USER for caching files.

Still it is pretty miserable for modern desktops. I wouldn't recommend NFS home directories on a large scale. The cheapness of PCs have rendered most uses for centralized storage and thin client stuff obsolete, although there is still a lot to be said for use that stuff for purposes of centralized management.

Caches in NFS home directories

Posted Jul 9, 2013 5:08 UTC (Tue) by Mook (subscriber, #71173) [Link]

And interestingly enough, Firefox can do things correctly - it does so on Windows already (the cache goes to local disk, the actually interesting stuff goes into the roaming profile). It looks like it also does the right thing on Linux these days, using $XDG_CACHE_HOME and falling back to ~/.cache. (I didn't know that before starting this - I was trying to find out who's spec'd ~/.cache!)

So... I guess it's sort of a configuration issue, because the defaults were not doing the right thing? :)

I'm fairly sure Internet Explorer on Windows also does the right thing, here. Nobody should be putting ~/AppData/Local on NFS.

Caches in NFS home directories

Posted Jul 8, 2013 6:33 UTC (Mon) by krake (guest, #55996) [Link]

"I know many DE:s and other GUI programs can be configured to behave otherwise, but it is not the default!"

Well, I guess the default to use a sub directory of $HOME is used because it is a safe assumption for a user writable directory that is not writable by another user (thus safe from symlink attacks or similar).

Anyone setting up a system with NFS mounted directories should have enough administrative knowledge to set an environment variable.

NFS home directories

Posted Jul 5, 2013 15:28 UTC (Fri) by njwhite (guest, #51848) [Link] (5 responses)

Wonderful article :)

So why the previous Windows setup didn't have as high network demands upon login? I don't really know anything about corporate(ish) Windows setups.

NFS home directories

Posted Jul 5, 2013 17:30 UTC (Fri) by jwakely (subscriber, #60262) [Link]

> So why the previous Windows setup didn't have as high network demands upon login?

IME it takes so damn long to boot up and pauses for such random periods at random times that you probably don't get a thundering herd all hitting the server at once ... on many school networks you're probably lucky it you manage to login at all ;)

NFS home directories

Posted Jul 5, 2013 20:31 UTC (Fri) by zlynx (guest, #2285) [Link]

I have certainly seen Windows login take a very long time.

The usual culprit is the user putting gigabytes of data files in their Roaming Profile, which is very easy to do by mistake.

Another common problem is login scripts which need access to some kind of shared resource like a database. These scripts are often badly written by sysadmins who never test on laptops running a VPN on a airport WiFi connection.

NFS home directories

Posted Jul 6, 2013 1:52 UTC (Sat) by raven667 (guest, #5198) [Link]

Roaming home directories in Windows are cached client side as well so the only traffic to the file server should be to sync state, no need to transmit the full file contents of everything read during login.

NFS home directories

Posted Jul 8, 2013 19:39 UTC (Mon) by drag (guest, #31333) [Link] (1 responses)

Unlike Linux, Windows has been widely used for excessively large deployments of users for many decades and much of the problems associated with managing users and desktops have been ironed out years ago.

Linux still hasn't caught up with Windows 2000 in many regards.

NFS home directories

Posted Jul 8, 2013 19:41 UTC (Mon) by drag (guest, #31333) [Link]

Oh.

And many Orgs simply ban roaming profiles. It's not really needed anyways for most situations. It's a sexy concept to have one configuration available on all desktops, but in practice it's very 'Meh'.

NFS home directories

Posted Jul 8, 2013 11:56 UTC (Mon) by kiko (subscriber, #69905) [Link]

Well, download may be too strong a word -- the desktop environment does require reading some files in the home directory, but it's not that many files and certainly not a sizeable amount of data.

For comparison, we run a 20-person diskless environment off a single server, with everything NFS-mounted, and for any operation but unlinks it's comparable to a local disk.

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 13:13 UTC (Fri) by maxiaojun (guest, #91482) [Link] (36 responses)

Except enthusiasm, how can one gain required skill to manage a large deployment of Linux desktop?

There are certification program, etc. to cover equivalent skill in alternative proprietary desktop OS.

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 15:46 UTC (Fri) by pboddie (guest, #50784) [Link] (35 responses)

Red Hat offers certifications, and I guess there are others doing so too. May I suggest an Internet search for "Linux certification"?

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 17:16 UTC (Fri) by maxiaojun (guest, #91482) [Link] (34 responses)

I don't the Linux certifications out there cover enough stuff for real desktop deployment.

After all, Red Hat is still shipping KDE 4.3, what can you expect from their certification program?

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 17:37 UTC (Fri) by rahulsundaram (subscriber, #21946) [Link] (33 responses)

Why is KDE 4.3 a problem for the kind of audience that would be helped by a certification program? Certification takes time to develop and it is never going to against the very latest open source versions since they aren't included in any commercially supported product anyway.

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 18:02 UTC (Fri) by maxiaojun (guest, #91482) [Link] (32 responses)

But the deployment mentioned uses KDE 4.10

Do you think it will be more successful if they use some "Enterprise Linux"?

A Year of the Linux Desktop (KDE.News)

Posted Jul 5, 2013 19:57 UTC (Fri) by rahulsundaram (subscriber, #21946) [Link] (31 responses)

There aren't enough details provided to conclude one way or the other. For most large deployments where certifications will add any value, maintenance cycle of the distribution is a much bigger concern than the incremental changes between 4.3 and 4.10.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 2:03 UTC (Sat) by rsidd (subscriber, #2582) [Link] (26 responses)

One can't argue with the success of RHEL but I don't think it is a good advertisement for KDE, which has always been a second-tier DE for RH. The changes from 4.3 to 4.10 were hardly 'incremental'. Many users started to find KDE4 usable only around 4.5 or 4.6.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 4:15 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link] (25 responses)

It is incremental when you consider what the large customers typically care about (ie) their custom applications running without any issues when they upgrade. It is not like your average desktop user.

And FYI, RHEL has several major customers using KDE (it is not quite 50/50 with GNOME but is pretty substantial) and continues to hire KDE developers who are upstream contributors on a regular basis. Within Red Hat, the developers tend to quite diverse in their preferred DE or WM usage as well.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 5:56 UTC (Sat) by maxiaojun (guest, #91482) [Link] (1 responses)

I'm not saying KDE there is broken.

The problem, what's the value of students learning in an environment where the software are obsolete and patched.

For example, should a Python newbie start with Python 2.6 today?

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 13:47 UTC (Sat) by pboddie (guest, #50784) [Link]

Python 2.6 is not obsolete. There are relatively few differences between 2.6 and 2.7, the latter being what the core developers of Python would have you use if you insist on using CPython 2.x at all.

And just like the story of KDE 4.x, anyone eventually adopting Python 3.x will have saved themselves a lot of work by not jumping on board immediately as soon as everybody was told that Python 3 was the way to go. Meanwhile, PyPy only currently supports Python 2.x, as do most of the other Python implementations, so it is only "obsolete" in the minds of those who cannot accept that there are people out there who are happy enough with what they already have.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 14:00 UTC (Sat) by pboddie (guest, #50784) [Link] (22 responses)

In this case, the customer didn't go with Red Hat, so it probably reinforces the view that Red Hat perhaps don't prioritise KDE as much as they should. Maybe Red Hat will learn from this and other experiences and increase their commitment, not only to the Linux desktop but also in other areas beyond basic services, cloud infrastructure and JBoss, which is the impression of their portfolio one gets from redhat.com. Offering stuff like coherent groupware solutions, for example, would be a much more convincing way of responding to the aggressive vertical integration practised by their competitors.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 14:35 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link] (19 responses)

Drawing conclusions from one deployment would be premature. You need to look at it more broadly.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 16:59 UTC (Sat) by pboddie (guest, #50784) [Link] (18 responses)

Well, I do give Red Hat credit for making progress. Around the time of Bluecurve, Red Hat was already getting justified criticism for shipping dated/broken KDE packages, and early Fedora versions had a KDE experience that I can only describe as "a joke", but in the past couple of RHEL versions the KDE support reached a point where it is as usable as my ancient Kubuntu desktop.

It's a shame that the progress made only applies to KDE 3, however, but I'm sure Red Hat recognises the need to get on top of things a bit more with KDE 4, especially since KDE 4, unlike KDE 3, is still a moving target. If more support were given to the applications, maybe Red Hat might start taking market share, but that actually means being visible to end-users and offering them something that replaces what they currently use, as opposed to covering a specific functional area in the server room and waiting for decision makers to point at it in a fit of "harmonisation" and have it kicked out in favour of a "total solution" from a competitor.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 17:19 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link] (17 responses)

As a KDE user then, I always thought the criticisms against Bluecurve were mostly bogus but in any case, Red Hat shifted away from the consumer desktop market a long time ago and their current customers don't care about desktop environments much as much as the applications they run on them and while they are targeting new markets on a regular basis (openshift, openstack etc), I doubt they are ever going to target the desktop market. Noone is making money off it. Canonical tried and has basically failed and changed their focus to mobile recently.

A Year of the Linux Desktop (KDE.News)

Posted Jul 7, 2013 14:44 UTC (Sun) by pboddie (guest, #50784) [Link] (16 responses)

There's a company in Washington state making plenty of money from desktop computing. Of course, a lot of that money is made by making people take those products when they buy computers, but even in the enterprise the desktop is a means to an end to that company and is seen as part of a total solution where you start making the big money once the customer is locked in.

Saying that nobody makes money off something that is an essential strategic instrument is a bit like the former CEO of HP wanting to sell the personal computing division because his personal dogma couldn't accommodate the realisation that without a foot in the door customers would go elsewhere out of convenience. Maybe Red Hat makes a decent enough amount of money from RHEL workstation licences (or whatever the client stuff is called these days), but it leaves the company vulnerable to "rationalisation" efforts where people start questioning why they pay for this extra thing when a competitor's site licence covers the same stuff (on paper) plus extra stuff, hateful though that stuff might be.

A Year of the Linux Desktop (KDE.News)

Posted Jul 7, 2013 23:33 UTC (Sun) by rahulsundaram (subscriber, #21946) [Link] (5 responses)

When I specifically referred to Linux on the consumer desktop and you respond with Microsoft and workstation subscriptions...

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 10:51 UTC (Mon) by pboddie (guest, #50784) [Link] (4 responses)

You brought up the consumer desktop in a comment on an article about deploying Linux in an organisation, and in response to a comment about deploying Linux in organisations. Where are the goalposts now?

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 11:40 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link] (3 responses)

Yes but you did read what you replied to

A Year of the Linux Desktop (KDE.News)

Posted Jul 9, 2013 11:07 UTC (Tue) by pboddie (guest, #50784) [Link] (2 responses)

Yes. What does the consumer desktop have to do with the article?

A Year of the Linux Desktop (KDE.News)

Posted Jul 9, 2013 15:21 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] (1 responses)

I was responding to your comments on Bluecurve and expanding the market and you bought up Windows in your reply which has nothing to do with the article either.

A Year of the Linux Desktop (KDE.News)

Posted Jul 10, 2013 12:29 UTC (Wed) by pboddie (guest, #50784) [Link]

Is Windows not a competitor for the enterprise desktop? But anyway, just to clarify something, I wrote...

Around the time of Bluecurve, Red Hat was already getting justified criticism for shipping dated/broken KDE packages...

By which I meant that on Red Hat 7.3 which didn't ship with Bluecurve, people were already wanting to install newer third-party KDE packages because they felt that Red Hat had not kept up. Don't feel too bad about this: Kubuntu got this kind of criticism later on.

I also doubt that Bluecurve, even though it is tangential to this discussion (as I was merely using it as a reference point), was a consumer-only strategy. There are good reasons for doing something like Bluecurve, and it arguably shows that the focus at Red Hat has been on the non-consumer desktop for a long time. Contrary to various assertions, all I'm saying is that the company might want to strengthen its commitment to the enterprise desktop.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 5:03 UTC (Mon) by rsidd (subscriber, #2582) [Link] (9 responses)

It's amusing to see someone lecture Red Hat, the only company ever to make a profitable business from Linux, on how to run their business. As I noted above, one can't argue with RH's success. I was only arguing with the claim that KDE 4.10 is an incremental improvement over the 4.3 shipped by RHEL, and observing that KDE is not a first-tier desktop for RHEL so it is not a good example (those who want to use KDE will go elsewhere, I expect, or maybe install a newer, RH-unsupported KDE on top of RHEL).

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 5:36 UTC (Mon) by neilbrown (subscriber, #359) [Link] (7 responses)

> the only company ever to make a profitable business from Linux

Are you sure about that? The *only* one?

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 5:40 UTC (Mon) by rsidd (subscriber, #2582) [Link] (6 responses)

Name another? As in, sustained profitability, with Linux software as the core business (ie, not HPC hardware sellers etc). I'm also excluding Android here.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 6:01 UTC (Mon) by neilbrown (subscriber, #359) [Link] (2 responses)

You've introduced that word "core" here. I'm not sure what that means.

Redhat uses Linux to sell services.
IBM uses Linux (And other things) to sell services and hardware.
LWN uses Linux to sell news
Google uses Linux to sell advertising.
Samsung uses Linux to sell phones (oh wait, you ruled out Android - not sure why).
etc etc.

The only way I can read your statement is as suggesting that Redhat sells Linux as their business model, but as you can get the same software for free that hardly seems likely. I think they sell something else (support, development, easy-access, a-name-that-instills-confidence, maybe some stylish headwear?) too.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 18:31 UTC (Mon) by raven667 (guest, #5198) [Link]

They sell the binaries that they compile, you are free to do what you want with the source.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 22:09 UTC (Mon) by mathstuf (subscriber, #69389) [Link]

The name also comes with "the government has vetted this to run on their machines" (provided you've run through their lockdown procedure). I'm pretty sure CentOS doesn't have that.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 6:19 UTC (Mon) by niner (guest, #26151) [Link] (2 responses)

SUSE does pretty good as well and runs profitable.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 15:17 UTC (Mon) by Wol (subscriber, #4433) [Link] (1 responses)

Yes but Novell bought SuSE because they were (apparently) on the verge of going bust.

Cheers,
Wol

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 16:04 UTC (Mon) by niner (guest, #26151) [Link]

SUSE may have had financial troubles a decade back, but remember that at this time, Redhat just made the first profits in its history. Nowadays both earn a healthy amount of money.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 10:57 UTC (Mon) by pboddie (guest, #50784) [Link]

Well, I'm glad that I provide entertainment. Maybe the point being made was that it's not necessarily enough to sell Linux into organisations on its own, or at least there are plenty of opportunities being missed in confining the products offered to just Linux and basic services. Given that Red Hat got into JBoss, maybe the company's leadership sees some value in a diverse portfolio as well, but Linux plus JBoss does not really cover all the bases, does it?

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 16:10 UTC (Sat) by sciurus (guest, #58832) [Link] (1 responses)

At least they're starting to compete with Active Directory via FreeIPA.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 20:10 UTC (Mon) by drag (guest, #31333) [Link]

FreeIPA is the only thing in Open Source-land that comes close to Active Directory.

Samba4 can be considered, but you'd still need Windows stuff to manage users and such I believe.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 2:46 UTC (Sat) by maxiaojun (guest, #91482) [Link] (3 responses)

BTW, is the change between GNOME 2.28 and GNOME 3.x incremental?

For maintenance cycle, yes, "Enterprise Linux" has much longer cycle. Google Chrome, etc., can be show stoppers, though. Supporting GTK 2.18 in a GTK3 world isn't fun at all.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 4:06 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link] (2 responses)

GNOME 3.x will be part of RHEL 7 and certifications would cover that then but I don't see the relevance to this discussion. It is hardly a GTK3 world when all major ISV applications are using GTK2 and anyway, those libs are parallel installable.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 5:49 UTC (Sat) by maxiaojun (guest, #91482) [Link] (1 responses)

> GNOME 3.x will be part of RHEL 7

The software is not released yet, not to mention the delay of certifications and CentOS

In a school environment, students are expected to learn by using latest and greatest open source software, right? Even if there are some hidden ISVs for Linux, it's irrelevant to students.

A Year of the Linux Desktop (KDE.News)

Posted Jul 6, 2013 14:38 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link]

Students don't benefit much from learning the latest desktop environments. School curriculum typically covers more fundamental technology like programming languages

A Year of the Linux Desktop (KDE.News)

Posted Jul 7, 2013 3:44 UTC (Sun) by fedtroll (guest, #91489) [Link] (3 responses)

OK, KDE and Gnome, let's be real. There's no "Linux Desktop" in 2013. Due to 10+ years of technical incompetence and infighting, KDE and Gnome have 1% market share. The Ubuntu "Unity" desktop is so freakish only paid Ubuntu employees hype it. Stop the lies of the "Linux Desktop". The truth is, Linux dominates Supercomputing and Web Servers with 99% market share.

A Year of the Linux Desktop (KDE.News)

Posted Jul 7, 2013 12:16 UTC (Sun) by wertigon (guest, #42963) [Link]

Actually Unity is not *that* bad once you get used to it - it's Gnome 3 with a tad bit different layout of everything, that's all. It's better than the Modern UI in Windows, that's for sure.

But yea, it's *different*, and it's certainly not what Gnome 2 was. Gnome 2 was rather polished at the end there, shame they had to ruin it. Right now I'm waiting for the next LTS before giving Ubuntu another chance thou. It's UI isn't terrific but it isn't terrible, but all the crap with MIR, Amazon etc has put me on hold.

A Year of the Linux Desktop (KDE.News)

Posted Jul 8, 2013 14:14 UTC (Mon) by ovitters (guest, #27950) [Link]

This is not slashdot.

A Year of the Linux Desktop (KDE.News)

Posted Jul 31, 2013 3:22 UTC (Wed) by zenaan (guest, #3778) [Link]

Dear fedTROLL,

looks like I missed this post, and I apologise for the late reply. It seems that you have more recently distilled your key points in your following post:
https://lwn.net/Articles/559615/

to which I endeavored sincerely to provide comprehensive answers.

There is no need to not keep our spirits high here on LWN, and as another poster suggested, yes, this is not /.

So to remind ourselves, every problem is an opportunity, it is just a matter of cognising that opportunity :)

High School for Girls?

Posted Jul 8, 2013 7:18 UTC (Mon) by m.alessandrini (guest, #36991) [Link] (20 responses)

Why in a civilized country should be there a "school for girls"? Especially in the domain of science and engineering, it really reminds of something ancient and bad.

High School for Girls?

Posted Jul 8, 2013 7:39 UTC (Mon) by mpr22 (subscriber, #60784) [Link]

Because the Tories wouldn't want to offend the "party faithful", and anyone who really cares was never going to vote Tory in the first place.

High School for Girls?

Posted Jul 8, 2013 8:02 UTC (Mon) by dlang (guest, #313) [Link] (17 responses)

because some people believe that removing sexual based distractions from the classroom (including the "I want to play the princess rather than show up the guy I'm interested in" behaviour) can improve the education that girls end up with.

Like everything else in education, there are studies that indicate this is true, and others that indicate that it doesn't help as much as just having teachers, staff, and parents who really care (one advantage that just about all private schools have is that parents have to go out of their way to get the kids in the school, so they are almost always more interested in the outcome than average)

High School for Girls?

Posted Jul 8, 2013 8:12 UTC (Mon) by m.alessandrini (guest, #36991) [Link] (16 responses)

I'm a bit afraid of these studies, I mean, whether it's "girls are more intelligent than boys and need no distractions" or "girls are more stupid than boys and need special schools"... In any case it's something slowing down the process of sexual parity.
One good reason would be: women have always been discriminated, now they show they can do things equally or better than men... But if you look at the article, it seems that the school managers are still mainly men.

High School for Girls?

Posted Jul 8, 2013 8:51 UTC (Mon) by dlang (guest, #313) [Link] (2 responses)

it doesn't have to be anything about discrimination (past or present), just reducing distractions in the classrooms can be a good thing

now, it can be argued that this isn't preparing them for the 'real world', but there are times when the primary focus needs to be on the technical material to be learned, and times where the social aspect needs to intrude into the academic space.

However, I do fall into the camp that says that school is only a portion of the day (let alone the year), and that the primary focus of a school should be on teaching the subjects, leave the social learning for the rest of the time (which can include after-hours school sponsored activities)

I am not sure that I believe that an all girls school is a great win, but there has been enough success shown from this model that I'm willing to defend it as a valid option for some people.

High School for Girls?

Posted Jul 18, 2013 17:31 UTC (Thu) by nix (subscriber, #2304) [Link] (1 responses)

I do fall into the camp that says that school is only a portion of the day (let alone the year)
I wish that were true. When I was in secondary school I had ~3hrs homework a day, ~10hrs at the weekend and several huge big projects forced on us for the gap between school years. This was mostly thanks to teachers never coordinating at all or finding out whether other teachers had scheduled a huge pile of homework on the same day, and each assuming that of course all the time in the evening would be available for us to do work for them.

And that was two decades ago! The volume of homework in UK private schools has, I am told, gone up substantially since. (I can't imagine how anyone can find the time to do it.)

High School for Girls?

Posted Jul 19, 2013 0:17 UTC (Fri) by mathstuf (subscriber, #69389) [Link]

> This was mostly thanks to teachers never coordinating at all or finding out whether other teachers had scheduled a huge pile of homework on the same day

Well, with my n=1 sample size, my grade school teachers had a general idea of when big projects were due. Of course, this falls apart if some people aren't on the same track as everyone else (e.g., a year ahead or behind in math/science), but on the whole, there was little overlap between large projects with relatively narrow time windows (month-long projects are going to overlap no matter what you do, but then again, there's a month of time to spread between them).

> …in UK private schools…

And my sample is a US public school.

High School for Girls?

Posted Jul 8, 2013 9:00 UTC (Mon) by neilbrown (subscriber, #359) [Link] (5 responses)

My son went to a boys school that was just next door to a girls school. That sounds like perfect gender parity to me.

On the other hand, the anecdotal evidence I've heard suggests that girls do better at single-sex schools and boys do better a co-ed schools. So there doesn't seem to be much gender parity there.

Maybe all girls should go to single-sex schools, and all boys to co-ed schools.

And while one girls' school may have a lot of male teachers, that doesn't mean they all do. My two sons went to two different single-sex schools - one with a male Principal and one with a female Principal. Both have plenty of male and female teachers.

And to be honest, if gendered mix were the main issue of concern in our schooling, I think we would be doing very well indeed.

High School for Girls?

Posted Jul 8, 2013 9:34 UTC (Mon) by m.alessandrini (guest, #36991) [Link]

"Maybe all girls should go to single-sex schools, and all boys to co-ed schools."

A bit difficult, don't you think so? :-)

High School for Girls?

Posted Jul 8, 2013 18:41 UTC (Mon) by raven667 (guest, #5198) [Link] (3 responses)

Maybe the best mix would be to gender-segregate 4-5 years between the ages of 10-16 or so, before puberty I think it would be irrelevant and after the worst of puberty is over it is probably unnecessary.

High School for Girls?

Posted Jul 9, 2013 10:41 UTC (Tue) by sorpigal (subscriber, #36106) [Link] (2 responses)

If the goal is to remove that kind of distraction then perhaps it would be better to segregate by sexual orientation rather than gender.

High School for Girls?

Posted Jul 9, 2013 15:33 UTC (Tue) by raven667 (guest, #5198) [Link] (1 responses)

That's a good point and I have no idea what the right answer is. For people who are aware of their sexual preferences at that age there can certainly be some tension in a same-sex environment although it is unlikely to be reciprocated except for the small community who share that orientation. That has to be balanced against the stress caused by singling people out and putting them in an opposite-sex environment, a situation where they are very obviously outsiders. And it doesn't actually fix the issue of sexual tension as you are by definition putting people who are attracted to _each_other_ together, and then putting them in a situation where it may be more difficult to form bonds with the larger group.

So I think that in a sex-segregated school environment, the incidence of homo-sexuality will just need to be tolerated and sexual tension will be of much less prevalent than in a 50/50 mixed-gender environment just due to numbers alone. In any event sexual activity among homosexual teens will not result in unwanted pregnancy so the societal risks are fairly low and I don't advocate for remaining segregated throughout the teen-age years, only for the first half, as everyone needs to learn how to live with their fellow human of whatever gender and orientation before they can be considered an adult.

High School for Girls?

Posted Jul 10, 2013 10:19 UTC (Wed) by sorpigal (subscriber, #36106) [Link]

I don't doubt anything that you say, but the whole idea of separating people based on some arbitrarily chosen dividing line seems unhealthy to me. You could also separate the introverts from the extroverts and the jocks from the nerds and maybe find some measurable benefit in grades, but I hardly think it's justified even so. Sex, either meaning gender or sexual orientation, is just one of the many things that you have to learn to deal with when interacting with other people and it seems overall counter-productive in the production of functional adults to attempt to remove it from the equation.

High School for Girls?

Posted Jul 8, 2013 15:21 UTC (Mon) by Wol (subscriber, #4433) [Link] (6 responses)

Just look at the school league tables, then!

I believe, iirc, that in maths the top ten schools by exam results are almost entirely single-sex, girls schools.

Cheers,
Wol

High School for Girls?

Posted Jul 8, 2013 17:45 UTC (Mon) by m.alessandrini (guest, #36991) [Link] (5 responses)

Wow, that's amazing. I'm wondering if girls are really smarter than boys, or the school system has problems in handling psychological and hormonal differences between them.

High School for Girls?

Posted Jul 8, 2013 20:27 UTC (Mon) by drag (guest, #31333) [Link] (1 responses)

Most schools systems generally suck at teaching anything.

The worst of all is the USA public school system. Some places feel the need to drug no less then 20% of the male population because they are unwilling to deal with them.

High School for Girls?

Posted Jul 18, 2013 17:33 UTC (Thu) by nix (subscriber, #2304) [Link]

This is irrelevant: the league tables he's talking about are private schools (generally very expensive ones). They definitely do *not* suck at teaching.

High School for Girls?

Posted Jul 8, 2013 20:54 UTC (Mon) by neilbrown (subscriber, #359) [Link]

> I'm wondering if girls are really smarter than boys,

As boys and girls tend to mature at different times (girls earlier), it is entirely possible that if you compare boys and girls in their mid teens and control for age, then you may well find that girls are "smarter" than boys. Five years earlier or five years later you would probably get a different result.

High School for Girls?

Posted Jul 9, 2013 11:20 UTC (Tue) by paulj (subscriber, #341) [Link] (1 responses)

I doubt it has anything to do with smarter. Speculating, I think it has to do with emotional and other brain-structure differences that impact learning. E.g. boys perhaps are less likely to listen and be told, and learn better through doing. I don't know much about pædagogy, but I think it's down to differences of that order.

I doubt there's a strict division. Some boys may actually do better with the methods that generally suit girls better, and vice versa.

High School for Girls?

Posted Jul 9, 2013 15:45 UTC (Tue) by raven667 (guest, #5198) [Link]

> boys perhaps are less likely to listen and be told, and learn better through doing

I'm sure there are some scientific studies and data to be had but it's very difficult to tease out any correlation and even more fraught to tie that to any causation when we are still lost in nature vs. nurture and trying to even identify the many and subtle environmental influences. Even in your example if you look at the opposite it betrays the basic cultural biases that affect all of our thinking, Eg. girls are more likely to do what they are told, followers and not leaders, and less likely to learn through experience by doing.

Everyone has these kinds of cultural biases, including the researchers, which makes any process of tugging out nuggets of real knowledge so very difficult. Also, as with most social studies, the experimental participants are generally young, middle class, college students which also skews any results in unknown ways.

High School for Girls?

Posted Jul 9, 2013 1:27 UTC (Tue) by bojan (subscriber, #14302) [Link]

Maybe because in civilised countries people are allowed to think differently, like to disagree at times, like to have a choice and all those other "terrible" things.


Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds