Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?
Back for a limited time - Get 15% off sitewide on Slashdot Deals with coupon code "BLACKFRIDAY" (some exclusions apply)". ×

Comment Re:This is really wierd (Score 1) 184

And as a result of Paris, there is a lot of racism directed at muslims the last couple days, or at least it suddenly feels like so. Nobody yet realizes that calling for the mass execution of muslims because they are evil and rape and murder sounds stunningly like Nazi rhetoric against the Jewish (just as untrue), and worse, it appears as though now it's culturally acceptable.

I have noticed the same thing. Whilst it does not appear to be universally acceptable, unfortunately it seems to me that it is the viewpoint of those who are most outspoken, particularly on social media, which turns it into an echo chamber. How do we combat this without simply getting into an argument with such people? There is no evidence that trying to reason with these people via the Internet achieves anything - in fact, there are countless forum posts and article comment sections which point to the exact opposite; arguing with anyone with such vehement beliefs in an environment without tone of voice, body language, and face-to-face contact only degrades into name-calling.

How do we *effectively* communicate that we do not wish such a thing to become culturally acceptable? Is it even possible, when mass media is intent on only ever reporting bad things?

Comment Re:Bottom line (Score 1) 86

That may be how it is, but it is not necessarily how it has to be. It is possible to build a system where the data is encrypted with per-user private keys, which never leave the user's device(s) - at least, not in the clear, and ideally only when being migrated/copied to other devices. Do all the crypto on the device, transmit & store it with private keys unknown to the owners of the infrastructure.

For all I know, this might in fact already be how iTunes & iCloud work already; that certainly seems to be the implication in the statement that data is "placed under the protection of your passcode ... [therefore] it's not technically feasible for us to respond to government warrants for the extraction of this data" (from your link). I'm pretty sure various online back-up services work this way.

Of course, there has to be a certain level of plaintext metadata: the fact that you have an account is not secret, nor are the amount of data stored, the access times, and the network addresses of devices used to access it. But the data itself? A system in which the service provider doesn't have centralised private keys is absolutely, completely feasible.

Comment Re: Bottom line (Score 1) 86

Could you clarify? References to Skype may or may not be relevant, but direct end-to-end communications is most certainly not impossible. It may be difficult in practice with contemporary IPv4 deployments (most devices are not directly addressable from the public Internet due to NAT), but of course it can be done: as long as it is possible for two devices to connect (which evidently it is, or we couldn't have an Internet at all), there is no "magic" which mandates that one or other of those devices be a corporate-controlled central server.

Central servers - effectively, brokers - do provide a lot of convenience: one place to publish & discover user presence, no need to bypass NAT at the endpoints because both connections are outbound, store & forward of messages for offline users, etc. But you *could* have a purely peer-to-peer network with offline exchange of contact details, or a central server used for nothing but storing details by which a user's device can be directly contacted.

Unless you count "routers" as "servers" - but with suitably randomised addressing and strong encryption, all that router logs will tell you is "device A sent some data to device B", nothing about the *meaning* of the data or the people behind it.

Comment Re:An APU with even 16GB of integrated memory (Score 1) 46

I'm confused. How would a GPU with 32GB of "integrated memory" be news, but a GPU with 32GB of [non-integrated*] memory is not news? I'm not sure what you mean when you say "integrated memory". This is not the graphics half of an APU, it is a discrete card, and nowhere do the summary or the press release state otherwise. The term "compute GPU" just means it's targeted at computing workloads, not graphics workloads.

What exactly is your complaint?

* Not even sure what this means, but you seem to be contrasting "integrated memory" and "memory".

Comment Re:Perl, bootstrap C++ (Score 2) 296

Are you trolling? For anyone not already intimately familiar with the process, the vertical learning curve of writing Perl bindings for C++ code will cause more pain, anguish, wailing and gnashing of teeth than writing in either pure Perl or pure C++. You will also gain nothing in portability: in fact you will lose, because portability will be the lowest common denominator of both Perl and C++ (I won't argue over which is lower to start with, both can be high with the right libraries), with the added headache of having to deal with two orthogonal sets of problems, in different languages.

Comment Re:Nostalgia sells. (Score 2) 79

I don't disagree that nostalgia sells, but I do disagree that what we are seeing here is purely nostalgia-driven. I, for one, prefer unrealistic "drift-style" racers to simulations - I get a lot of enjoyment from going as fast as possible, negotiating courses through a mixture of careful positioning and controlled drifts, with the height of skill being completing a lap without releasing the accelerator, without crashing.

Games which deliberately ape the looks & sounds produced by old systems may indeed rely heavily on nostalgia, but there are plenty of other games out there maintaining the old-fashioned arcade driving mechanics, whilst taking full advantage of modern hardware. Personally I would put Mario Kart 8 in this category (although it is debatable whether the Wii U can be called "modern" in the graphics department). In TFA itself, the Power Drive 2000 trailer may have retro music and a retro *feel* to the graphics, but the graphical fidelity itself is not artificially restricted. Elsewhere on Kickstarter, Formula Fusion [1] seeks to recreate the style and mechanics of the WipEout series, whilst not in any way pretending to be an old game - I for one am excited by the prospect of finally having what is essentially WipEout (in all but name) running on modern PC hardware, with all the bells, whistles and convenience that implies, but would probably be put off if they were to deliberately attempt an original-PlayStation aesthetic. The 90s Arcade Racer [2] is definitely playing heavily on nostalgia, littered with references to (as you may have guessed) various 90s arcade games, but again, it seeks to make the best of the underlying hardware.

Nostalgia is certainly one aspect of all this, but don't underestimate the number of people who simply find these kind of games fun, and want to be able to play them easily & legally on contemporary hardware! I suspect I am not alone in finding that simulation-style games are not enjoyable without matching realistic controls, but have neither the space to dedicate to wheels, joysticks, throttles, pedals etc. - nor do I particularly want to spend the money or devote the time. For example, much as I am pleased that Elite: Dangerous and Star Citizen exist, I personally am holding out for No Man's Sky, simply because releasing on PS4 first means it is far more likely to have a simple control scheme which works on common controllers. Many will probably decry it as "dumbed down" or "retro"; I say it is just a different design decision.


Comment To those expecting lower requirements for the Vive (Score 1) 227

Do you honestly expect HTC and/or Valve to have invented some magic which somehow manages to render images at the same (or higher) resolution & framerate, with the same image quality and in-game graphics options, with any less beefy hardware to back it up? Or do you think Oculus are simply lying about what is needed for a good experience?

Comment Re:The GPL (Score 1) 469

In this case, they are all things that require knowledge of who is logged in - functions to do with actually tracking creation/switching/ending of sessions, or things where admins may wish to change policy based on who is logged in (e.g. non-superuser can't reboot a shared machine whilst anyone else is using it). I agree it does seem like a bit of a kitchen sink, but it represents things that need to be considered in tandem for this functionality to work well on desktop systems, which have not traditionally been considered in tandem - on the one hand, this is the kind of consolidation systemd opponents complain about; on the other hand, in my experience, all this stuff now works better than ever before.

IMHO, logind is not a good example if you want to demonstrate feature creep - it is a good example of providing a unified solution to a bunch of related problems which were not previously addressed in a satisfactory way. Better examples of feature creep are things like networkd (for dynamic configurations on desktops/laptops we already have NetworkManager; for static configurations on servers you only need to get the distro-specific network set-up right once then leave it alone), or timesyncd (what's wrong with existing NTP clients?).

Personally, I'm not strongly opposed to systemd, and have observed some benefits from it - but my usage of it is limited to my own desktop and laptop, I am not a sysadmin worried about having to re-learn how to administer an entire network. In the context of desktops & laptops, I would say systemd is a good thing; elsewhere, I don't consider myself qualified to have a strong opinion.

Comment Re:The GPL (Score 1) 469

To provide a consistent, reliable way of tracking who is logged in, and interfaces for doing various things related to user sessions - this includes providing controlled access to things logged-in users might want to do, such as suspend, reboot, power off, access input devices (separately from those in use by other users who may be logged in to the same machine), switching between different sessions, inhibiting suspend (because I'm watching a movie), and so on. A whole load of stuff which has traditionally been unreliable on desktops, or only worked for one user at a time, or worked differently for each distribution, or had no consistent mechanism for controlling access to the functions.

It has a man page: ... and a detailed description of the interfaces it provides, should you wish to provide an alternative implementation:

Desktop environments choose to depend on logind because it frees them of the responsibility to implement all this stuff themselves - which has traditionally been a mess, because the way these things are handled across different distributions has always been subtly different. Which group do I need to be in to be allowed access to reboot? What are the permissions on the device node for the keyboard? How does a generic video player app tell the system not to turn the screen off, without individual support for the methods used by various disparate desktop environments? If you have a common interface which all the DEs can use (and other apps with no specific DE affiliation), it becomes very tempting to, you know, *use* it.

Its access control goes via Polkit, which is itself a generic system to controlling access to privileged things. Polkit itself is not part of systemd.

Comment Supported != Secure (Score 4, Insightful) 137

TFA and the summary make it sound as if it is the lack of support contract which makes these systems insecure. This is complete and utter nonsense - it is the fact that they are running Windows XP which makes them insecure. It's not as if malicious hackers around the world were sitting there rubbing there hands in glee, waiting for the day the support contract expired to plunder the systems, having previously been completely and utterly thwarted in their evil plans by the exchange of funds between the UK government and Microsoft.

But at least a support contract would get them fixes for any newly discovered vulnerabilities, right? Well, maybe. No software is perfect, but the world - and Microsoft's practices - have moved on, and realistically it would take a *lot* of money for MS to spend a meaningful fraction of their resources securing an OS past the end of its useful commercial life.

Comment Re:Can't they just get it right? (Score 1) 88

"because there is less flickering in DirectX games". DirectX games played under Wine, or are your problems with AMD/ATI not actually directly related to Linux at all? I'm not sure what you mean by "flickering", but the problem anti-aliasing is designed to solve is, well, aliasing - that is, jagged edges on objects caused by the unavoidable fact that the on-screen image is composed of individual pixels, which becomes noticeable whenever different coloured objects don't line themselves up perfectly along pixel boundaries (i.e. most of the time).


If the problem is not strictly to do with jagged edges on objects, you may also want to read up on mipmapping and/or anisotropic filtering:

You might be misunderstanding the problem and exacerbating things through poor graphics options, or you might simply be abnormally sensitive to the limitations of interactive 3D graphics rendering. Alternatively, if by "flickering" you mean entire objects are actually disappearing/reappearing, that sounds like an application bug, or a hardware failure waiting to happen (e.g. video memory corruption resulting from overheating).

Quark! Quark! Beware the quantum duck!