Forgot your password?
typodupeerror

Comment: Re:not like megacorps don't control OSS already (Score 4, Insightful) 49

by Xest (#47916755) Attached to: Industry-Based ToDo Alliance Wants To Guide FOSS Development

That, and frankly, the FOSS community needs to make a choice - if they want the year of Linux on the desktop to stop being a joke and start being a reality there has to be a move towards this sort of professionality that gives corporations the confidence they need to roll it out onto the desktop - they need to know it will be supported for x number of years, they need to know there will be frequent updates to keep it competitive and so on.

There are a non-negligible number of participants in the FOSS community that want both global domination for their software, but just want to continue to developing in a laissez faire attitude - avoiding the responsibilities of creating something genuinely competitive and reliable enough to become dominant such as decent planning, professional efforts on documentation, design and UX stuff whilst also complaining that it doesn't get the attention they want.

IMO there are two choices developers have when working on FOSS projects:

1) Do it for fun and don't give a shit who or how many people use it, don't care how they use it, when they use it, just develop it because you like developing and because it's a good opportunity to showcase your skills and keep them sharp.

2) Care about how people use it, how successful it is, how widespread it is, but accept that there is a cost to this - the cost being that you have to expend effort doing the non-fun parts of software development like offering support, quashing bugs in a pre-agreed timeframe, providing documentation and so on and so forth.

Those who want the success and want to crush proprietary software without wanting to put in the effort of doing the boring stuff that makes proprietary software so successful in so many cases are living in fantasy land. Personally I have respect for people whatever decision they make above, what I don't have respect for are those who want to achieve the political agenda of 2) but simultaneously demand the lack of responsibilities of 1) - that's not realistic.

Comment: Re:hope for improvements (Score 1) 321

by Xest (#47916043) Attached to: Microsoft To Buy Minecraft Maker Mojang For $2.5 Billion

"Are you running it full screen at your maximum screen resolution with the highest graphical settings "

Yes absolutely, I think the laptop resolution is only 1366 x 768, but given the lower specs it shouldn't result in such a disparity. The laptop does struggle a little with Diablo 3 unless you turn the settings down a bit but certainly have never had any problems with Minecraft (it's actually the laptop my girlfriend uses to play - the system I play on is a 2.83ghz quad core with 8gb of RAM from 2008 with just the graphics card updated to a 650ti a year or two back and that easily gets 60fps at max settings on Minecraft and close enough on Diablo 3 too for reference).

"I'm pretty certain it's the game, not the drivers. It's the same regardless if it's on Linux or Windows."

Would it not be a universal issue if it was the game? This is genuinely the first I've heard of glaring Minecraft performance issues and I know of many people who run it at 60fps or more on far lower spec systems than you're running. I was going to suggest that maybe there's a problem with your Java install, but if it's effecting Linux and Windows equally that that seems unlikely (but not impossible of course) too. There must be something that Minecraft or Java/JNI doesn't like about your specific hardware configuration, but I really couldn't guess what.

Comment: Re:while... (Score 2) 112

by Xest (#47915827) Attached to: Indian Mars Mission Has Completed 95% of Its Journey Without a Hitch

The problem is you're taking that evidence and extrapolating it to "Asia" is a powerhouse. In doing so you're effectively feeding off the success of countries that have done exceptionally well like China when the reality is that different countries in Asia have had very different experiences. China has grown rapidly but Japan has stagnated, India has disappointed whilst in South America, Brasil has completely outflanked it.

Whatever India's policies, one thing is clear - it's not done very well compared to China and Brasil so those policies most definitely have not paid off. Starting from a low point with over a billion people India should've done much better than it has- it's policies have been largely a failure- not as big a failure as they could be sure, India is still growing well, but it's nowhere near it's potential, not even close, and it still has massive problems where little progress has been made despite neighbours like China making far greater progress on those issues (access to education, poverty).

Comment: Re:hope for improvements (Score 1) 321

by Xest (#47915809) Attached to: Microsoft To Buy Minecraft Maker Mojang For $2.5 Billion

Okay, let's make one thing clear - I am not saying that Java programs are always faster than C++ programs or any such thing. I am saying, correctly, that there are technical reasons why there are circumstances in which Java can outperform C++. Over the course of a large complex program these benefits will almost always be outweighed by the disadvantages. I'll address your points one by one:

1. "1. Java is faster than C/C++?! Citation fucking needed. Any time Java meets or exceeds C/C++ for speed is probably because the library that the Java program is leveraging is an extremely optimized, pre-compiled, native binary... probably written in C/C++."

Incorrect. When Java is faster than C++ it's because it compiles with more context than a C++ application does when it is compiled. This allows for things like better virtual function inlining and subsequently things like better loop vectorisation. Again, this has to be taken in context, this does not mean that these optimisations that Java can make but C++ can not always make Java faster, that's absolutely false, but in practice it does mean that contrary to the common myth that Java performance is abysmal that performance has in the last decade rapidly approached (and again, in specific circumstances, surpassed) C++ performance.

2. "2. JIT compilation doesn't magically transform inherently slow semantics into super-fast-omg-native-speeds. Guess what? New'ing up a bunch of objects, destroying them, checking the bounds on every array, boxing and unboxing, pausing everything to run a garbage collect, using linked lists instead of arrays, and other similar abuses of memory and cache are fucking slow. Guess what Java does a lot of? You can pull the same stupidity in C or C++, but at least there's the option not to."

You are partly right here, and partly wrong. Some of the issues you cite are reasons why, in practice, large Java applications rarely perform as well as C++ counterparts, but you are also partly wrong - a number of the things you cite are either not issues, or are regularly optimised away by the compiler in practice. The garbage collector is indeed an issue that is the bane of obtaining deterministic performance with Java, but it's not a beast that is impossible to tame, which is why Java has had many successful applications to HPC tasks.

"3. "Well just write your Java program with memory stuff in mind." So... write it like C++ and STILL pay the penalty of the GC kicking in at random times and trashing your frame rate? No thanks. There's a reason engines like Unreal, Crytek, and id's are all written in C++: they write their own stuff to handle memory in an efficient fashion instead of a random stop-everything-while-I-clean-this-mess-up GC."

Sure, and there's a reason indies are almost entirely using JIT'd languages like Java and C# through tools and frameworks such as Unity and MonoGame - because when you're not going AAA then these types of frameworks perform perfectly well for gaming as proven by so many brilliant indie games in recent years. I absolutely agree that if you're making the latest and greatest cutting edge 3D engine that is going to blow away people with it's visuals that you'll still want to do it in C++, but let's be clear - this isn't most game development, far and away most game development in recent years is being done with managed languages - even on mobile there is a massive slant towards things like Unity. Even big studios like Blizzard are using it sometimes for games like Hearthstone.

"4. Why pay the penalty of waiting while the JIT converts stuff to optimized machine code? Why not START with optimized machine code?"

Because there is an absolutely massive development overhead in doing so.

There are a lot of folk stuck in the past, believing that C++ is still the god language, the be all and end all that should be used for everything, just as there is an even smaller subset of folk who see those C++ zealots even as heathens and believe everything ever should be written in C, and I'm sure it goes all the way down the chain to the one guy in a basement of some office block somewhere that insists everything should be written in machine code directly.

But that's not the real world, the fact is that lines have become blurred, it's not so clear cut any more that the lower level you go the better performance you get, hell, this isn't even a new thing - in the early 90s you could always write C/Assembly that performed better than C++, but by the end of the 90s and start of the 00s this was no longer true, there were few assembly programmers with the level of talent required to write more efficient assembly than a C++ compiler could churn out. The same is happening now with managed languages - as the last decade has progressed it's become ever less clear that there is any tangible performance benefit in using C++ for many applications and that's why even the game development world, one of the most high performance of all consumer facing worlds of development has even started to stray away from C++. There absolutely are still places where C++ matters, and professionally, I know this for a fact, because I use C#, Java, and C++ primarily and as a software architect it's been up to me to make the call as to when it's time to use C++ - a call that I most certainly do make when it makes sense. The reality is though that the time and place for C++ has consistently declined over the years relative to other technologies.

Now you might not like this, it might offend you greatly, you may take great issue with it, but it is reality, and no amount of bitterness and language fanboyism will change it, so you can really just make a choice - either sit bitter, posting anonymously (presumably because you're embarrassed to put your name to something you deep down know you don't fully understand) about something that wont change because your viewpoint is based on a few various fundamental lacks of understanding of the nuances of the technologies in question, or get over it and embrace it like everyone else that's actually being productive in this constantly changing world.

Comment: Re:NSA scorecard on on truth? (Score 1) 191

by Xest (#47915721) Attached to: New Details About NSA's Exhaustive Search of Edward Snowden's Emails

You pay the NSA to spy arbitrarily on civilian photos, files, and text messages in breach of the Universal Declaration of Human Rights that the US has ratified into law?

No wonder things like the Boston bombings and 9/11 happened if you're more interesting in paying your security services to arbitrarily breach the right of privacy of everyone and anyone rather than you know, real actual potential threats to your country.

If you think the solution to finding those threats is, rather than making better tools for finding the needle in the haystack, to simply make the haystack bigger and the needle harder to find than ever, then you're part of the problem. You obviously really don't understand what security services are for or how they can act more efficiently, your opinion of what they should be doing is the anti-thesis of effective security services so it's not surprising that they commit transgressions internally when you openly support committing of transgressions and ineffectiveness externally.

You apparently see these as two distinct problems, naively oblivious to the fact that they're one and the same- you can't openly support illegality and ineffectiveness when it suits and then bitch and moan about it when you suddenly take issue with it.

Comment: Re:suppliers tested? (Score 1) 201

by Xest (#47915693) Attached to: Sapphire Glass Didn't Pass iPhone Drop Test According to Reports

That's not inconsistent, it's testing an untried proposition. What you're suggesting is taking a loved proposition and completely getting rid of it.

Obviously it's not me that doesn't understand about designing for HCI because I get something fundamental that you don't - I get that if you have something that works for users, and that is loved by users, then you don't throw it out the window and start all over again - minor 1mm adjustments to perfect an already great design are enough.

Comment: Re:hope for improvements (Score 4, Informative) 321

by Xest (#47908747) Attached to: Microsoft To Buy Minecraft Maker Mojang For $2.5 Billion

"They need to drop Java or figure out a way to compile Java to actual machine code so the game runs well."

Yeah, why has no one thought of this?

You realise the way modern Java (since like 1999) works is that you write an application in Java, that Java code gets compiled to Java bytecode, which you can think of as a cross platform version of assembly, and then that Java Virtual Machine on which you run that bytecode (i.e. the compiled Java application) does in fact convert it into actual machine code right? Not just machine code, but machine code optimised for the exact machine the JVM is executing on? This allows the JVM to reach C++ levels of performance and some cases go beyond, because C++ is generally only compiled for a specific architecture, whilst the JVM optimises for a specific machine.

This does mean slow first time execution of modules as each module is optimised to that executing machine's native machine code the first time it is used, but after that first execution of the program or library you're basically getting native performance.

For what it's worth though, the console version of Minecraft (360, PS3, PS4, Xbox One) is apparently written in C++ because some of those platforms - i.e. the Xbox - don't have a Java Virtual Machine on which the Java version could be executed on.

I have a relatively low end laptop, it cost like £300 a year ago, and it runs Minecraft absolutely fine. What spec are your machines if they can't even run Minecraft?

Comment: Re:An end to XBox? (Score 3, Insightful) 321

by Xest (#47908491) Attached to: Microsoft To Buy Minecraft Maker Mojang For $2.5 Billion

I doubt Microsoft cares how it does in Japan nowadays, Japan stopped being a relevant indicator of the health of a video game industry entrant about 10 years ago. Since then both the US and subsequently Europe became bigger markets by far, and even markets like Brasil and China are arguably more worth spending your time on now than Japan if you're in that industry. Japan's two decades of economic stagnation have really hit it's relevance to the industry hard in this respect - the struggling Wii U and Sony's precarious overall financials (The PS4 is doing well though thankfully) have only exacerbated the problem.

Despite their mis-steps this generation they actually did well last generation in the end in large part because they were pulling in over $1bn of pure profit from Xbox Live subscriptions alone within a few years of the launch of the 360. This couple with the highest attach rate by a decent margin coupled with higher profits-per-game than the Wii last generation allowed them to be more profitable despite not shifting anywhere near as many consoles as the Wii did.

Whether they'll keep doing well is anyone's guess, but the XBox division is currently a massively different beast compared to how it started last generation with it's RROD writeoffs and massive initial R&D expenses on the system.

There were rumours of them selling it off and such but I can't see them getting rid of it now that it's finally been a healthy net profit centre for a good few years now - it would seem odd to invest 10 years on profitably making your way into a key target area for Microsoft - the living room - only to then give up when you've achieved your goals of decent market penetration and real actual profit, still, stranger things have happened so I guess we'll see.

Comment: Re:NSA scorecard on on truth? (Score 4, Insightful) 191

by Xest (#47908375) Attached to: New Details About NSA's Exhaustive Search of Edward Snowden's Emails

I've no idea because personally I'm not American and hence not affected by the IRS' actions so nor do I particularly give a shit what they have or haven't lied about.

The IRS may be a bigger threat to liberty in the US than the NSA, but it's certainly a non-entity in terms of dangers to liberty for the whole of the rest of the world compared to the NSA which is a real genuine threat for those of us not living in America yet still having our data stolen and our privacy invaded.

Comment: Re:"console shooter" (Score 1) 93

by Xest (#47907375) Attached to: Early Reviews of Destiny: Unfulfilled Potential

It's not so much the success of the franchises year on year I was referring to but ultimately the size of the demographic in question.

Ultimately these sorts of games and most PC games (educational games aside) are decide to fulfil a simple purpose - to entertain, to provide people with an avenue to relax and have fun.

So my point is, ultimately, that if franchises like CoD, BF, Halo and so forth are routinely shifting 5 - 10 million units year on year whilst PC games only ever achieve these sorts of figures with one title every 3 years then my argument is that clearly these franchises are successful in their goal - that a majority of the populace looking for such entertainment find it happily in these games.

My point therefore, is that if a majority are happy with these titles as an option for their enjoyment that one cannot realistically objectively define these titles as "rubbish" but only subjectively do so as a minority viewpoint. Fundamentally, as a large majority of gamers prefer the console medium for even their shooters it seems a bit of a stretch to say objectively that they are rubbish else if they were then people would instead just game on PCs (much like they used to in the Quake/Doom days when PCs were vastly more popular for shooters).

It was only with the advent of things like Halo on the Xbox and Goldeneye 64 on the N64 that console shooters really started to take off and I'd argue that there was a turning point there at which gamers started turning towards consoles for more than just mario-esque platformers and top down shooters and the like and started gradually to prefer these platforms for shooters in general.

I don't think there is a mutually exclusive PC or console gamer demographic, most console players I play with are also PC gamers and I myself am, but I find more and more that the games I play on my PC are games like RTS games and harder to categorise and very innovative (e.g. Minecraft) indie games more and more and shooters less and less - this despite cutting my teeth on PC shooters, and, as I say, still having the fondest memories from Quake above all else with Doom and Wolf coming in after that.

Comment: Re:NSA scorecard on on truth? (Score 5, Insightful) 191

by Xest (#47907297) Attached to: New Details About NSA's Exhaustive Search of Edward Snowden's Emails

But this is a sworn declaration, and if it's a sworn declaration then it must be true because it's not like anyone has been caught lying under oath on this topic is it!

Honestly, sworn declarations on this topic and the lack of punishment for breaching their oath when swearing the truth means you might as well read "Sworn declaration" as "In a conversation with his mate Dave down the pub".

Comment: Re:suppliers tested? (Score 1) 201

by Xest (#47907007) Attached to: Sapphire Glass Didn't Pass iPhone Drop Test According to Reports

Um, I don't know if you read a different article than the one you linked but it doesn't back up your rant. You're claiming they didn't innovate and tried just adjusting the controller by 1mm. They stuck to the design because the 360 design was popular and widely praised as one of the best console controllers of all time. Why would they want to stray too far from that? even the Wii U's pro controllers and the PS4 controller has converged towards the 360's design.

As for innovation the article points out they tried smells, projectors and all sorts, how is that not innovating? If their R&D tells them people like the existing controller with only minor adjustments then that's valid R&D, why go with something completely alien just for the sake of being able to argue "Hey look guys we innovated! it's shitter than the previous design and everyone hates it, but we innovated!".

Innovation in R&D isn't about changing shit wildly for the sake of it, there has to be a product that people love at the end of it else it's pointless.

APL hackers do it in the quad.

Working...