Look, I'm against piracy. Not only I wouldn't give them any updates, but I think pirates should swing from the mast, like in 1600. Or make them walk the plank.
But I also think there's a reason why there's "BS" in "BSA". Their statistics make me want to puke. They do such bullshit statistics as taking a pirated CD from Taiwan or China which includes some expensive piece of softwware, like 3D Studio Max, and say "see, there's 5000 USD worth of software on this CD. We think 10,000 chinese kids bought this CD for 5$, which robbed us of 50,000,000 USD."
That's utter bullshit. Most of those Asian and Eastern European pirates do not need 3D Studio Max and wouldn't buy it anyway, even if they could afford to. (Which they can't. As was said before, a chinese family would need to pay _all_ their income for _two_ _years_ to afford a license. Again: _all_ their income. That is, leaving them with no money for food, rent, clothes, etc.)
We're not talking 10,000 professional designers and architects who actually need it, we're talking mostly kids who much around with it a bit to make some skins for mods for old games. Maybe 1 of them will actually release an obscure mod, the rest just mucked a round a bit with it, uninstalled it and moved on to something else.
Would all 10,000 of them have bought 3D Studio Max if they couldn't pirate it? No. _I_ wouldn't buy it either, much as (1) I could easily afford it, and (2) I'm tempted to try modding "X2 - The Threat." (Which, sadly, only supports exporting stuff from 3DS MAX.) Now I don't pirate it either, but even I think it would be utterly retarded to pay $4000 on tools to mod a $40 game.
Yet the BSA would want me to believe that 10,000 dirt-poor kids from Taiwan would. That's so much bullshit, it could fertilize a few acres.
Ahh... Ok, I suppose O(n^2) can make one think of bubble sort.
Given enough lack of clue or lack of care for performance, though, one can squeeze an O(n^2) or worse in lots of other places. E.g., it's pretty easy to end up doing an O(n^2) even for searching through a table.
Or an example of clueless coding that's particularly dear to me, comes from some guys who basically tried to shaft me back in my freelancer days. Except they couldn't work with trees. At all.
They had data which was a simple binary tree. Or rather the perfect textbook example for something fit to be stored in a tree. (A big rectangle which could be cut horizontally or vertically into two. Each of which could then be cut in two. And so on.) They needed to save and load it.
But they couldn't work with trees. So they replaced the tree with an array, losing all the tree data in the process. (No, not a tree-inna-array, with indexes instead of pointers. Juts a flat array with no pointers.) And saved it as such. Which upon reading the file left them with the problem of which piece depends on which one. So they used backtracking to find out which pieces have already been cut.
The result? It took minutes to load even the smallest trees. For a piece of processing which shouldn't even have existed at all, if they only knew how to recursively traverse a tree.
Guess the wonder of one liners is that it just lets me wondering what you meant.
A binary search is O(log2(n)), which is to say, searching through 1 million records takes a mere 15 steps more than searching through 32 records. So there's nothing particularly fantasy about them. Such algorithms exist.
Other kinds of algorithms involve even less steps. E.g., the kinds of trees used in an English spell-checker (well, some spell checkers anyway), have 26 as a base for the logarithm. They also have other advantages over a binary search, such as being able to calculate a "distance" between two words.
Can they be beat? Sure. A hash table is O(1) for retrieving an element, so it executes faster than a binary search for most data sets. (I.e., given enough data so that calculating the hash isn't slowing you down. For string keys this pretty much means every time.)
But I'm guessing you knew such elementary things already, and I'm probably just boring you. So if you'd kindly clarify what you had in mind, I might even be able to give a less boring answer:)
No offense, but _why_ would you even need to think about such things as gender in an inanimate tool? Why do you need something like that to be comfortable around a piece of metal, silicon and epoxy? No, I'm really curious. Why?
My computer is an "it" and I'm very comfortable with that. Never felt any more need to assign it a gender than I needed to assign a gender to the desk it sits on, or to the chair in front of it.
I also can't help notice that the most successful interfaces amd devices (as opposed to gee-wow crap that's hyped by marketroids and then dies nevertheless) is stuff that's abstract and not over the top in simulating real life devices or beings. Humans are very capable of learning new and abstract things.
E.g., probably the most successful such device is the mouse. And there is no real life equivalent of it. But people nevertheless have no problem in learning to use it. E.g., the context sensitive popup-menu. There is nothing in real life which acts that way, and definitely nothing to anthropomorphise about it. But everyone learns very quickly to appreciate it. E.g., the hyperlink. Again, there is nothing in real life books, nor on real life humans, which acts that way. But it works so well, that it's used not only in browsers, but in other programs and games too. Etc.
By contrast, such hyped attempts at simulating the real world as Microsoft Bob, died a silent death and noone cared about them. Invariably such attempts waste more screen space for what a simple icon would have done just as well, and end up requiring more clicks to get to the same place. As a result, they end up requiring more time and effort to _discover_ where you really want to get, if you didn't know that already. ("Discoverability" is one of the main issues of UI design.)
So how would going even more extreme, and even assigning Bob a gender help UI design anyway? If you have actually put some thought into design and usability and ended up with a provable need for gender, I'd really like to hear about your findings.
How would a definitively feminine or masculine interface work, anyway? What would it need to have, that the current UIs don't, and which can't possibly be done without assigning it a gender? (Have a breast instead of a mouse?;)
And how would you convey the idea that your particular UI is definitively masculine or feminine to your users, without resorting to insulting over-exaggerated stereotypes?
I can just picture having a big animated 3D female head popping up instead of clippy, with a blank look on her face, giggling and twirling her hair. "Hi. I'm, like, Cindy. I'm, like, your UI helper. Isn't it rad? Hihihi. I, like, notice that you're, like, typing stuff. And it, like, looks like an URL. Do you, like, want me to, like, convert it to HTML for you?" *click* "Hihihi. Oopsie. I'll, like, leave it alone for now, then."
It's _not_ what I think of females, but I can see some lonely male geek using that stereotype. The same guys who think that tiny chainmail bikinis are good female armour, for example. Or a worse stereotype.
Ok, so maybe you'll avoid ending up this stereotyped. So how would you convey the idea that your GUI is definitively all female, to all your users, most of which never even thought that a GUI has or needs a gender?
I still don't think you should start doing every single silly trick in your code, like unrolling loops by hand, unless there's a provable need to do so. Write clearly, use comments, and use a profiler to see what needs to be optimized.
That is coming from someone who used to write assembly, btw.
But here's the other side of the coin: I don't think he included better algorithms in the "premature optimization". And the same goes for having some clue of your underlying machine and architecture. And there's where most of the problem lies nowadays.
E.g., there is no way in heck that an O(n * n) algorithm can beat an O(log(n)) algorithm for large data sets, and data sets _are_ getting larger. No matter how much loop unrolling you do, no matter how you cleverly replaced the loops to count downwards, it just won't. At best you'll manage to fool yourself that it runs fast enough on those 100 record test cases. Then it goes productive with a database with 350,000 records. (And that's a small one nowadays.) Poof, it needs two days to complete now.
And no hardware in the world will save you from that kind of a performance problem.
E.g., if most of the program's time is spent waiting for a database, there's no point in unrolling loops and such. You'll save... what? 100 CPU cycles, when you wait 100,000,000 cycles or more for a single SQL query? On the other hand, you'd be surprised how much of a difference can it make if you retrieve the data in a single SQL query, instead of causing a flurry of 1000 individual connect-query-close sequences.
(And you'd also be surprised how many clueless monkeys design their architecture without ever thinking of the database. They end up with a beautiful class architecture on paper, but a catastrophic flurry of querries when they actually have to read and write it.)
E.g., if you're using EJB, it's a pointless exercise to optimize 100 CPU cycles away, when the RMI/IIOP remote call's overhead is at least somewhere between 1,000,000 and 2,000,000 CPU cycles by itself. That is, assuming that you don't also have network latency adding to that RPC time. On the other hand, optimizing the very design of your application, so it only uses 1 or 2 RPC calls, instead of a flurry of 1000 remote calls to individual getters and setters... well, that might just make or break the performance.
(And again, you'd be surprised how many people don't even know that those overheads exist. Much less actually design with them in mind.)
So in a nutshell, what I'm saying is: Optimize the algorithm and design, before you jump in to do silly micro-level tricks. That's where the real money is.
Well, what sorta bugs me is that this kind of inflating scores has crept everywhere. E.g., let's take something fairly neutral for most people: game reviews.
Well, except for some fanboys who'll foam at the mouth if you gave their idol's game only a 99% rating. Bonus points if they haven't even played the fscking game yet, but they just _know_ it's a perfect 100%.
I've seen for example a review on a major review site, for an overhyped game that I truly hated. The reviewer went to great lengths to demolish every single aspect of it, and for good reason. The review told you just how much every single bit and detail and design decision sucked. The controls, the interface, the physics, the quests, etc. He sounded like he hated it all. It even ended with such damning words as saying that the game might be ok for someone who's never ever played another game before, and thus can't compare it to anything, but won't cut it for anyone else.
Then they give it some rating like 84%. Uh? Excuse me? Regardless of what _I_ thought about the game, that reviewer himself had wrote a hate-fest about the game. He had savaged it thoroughly. There was _nothing_ in his review that would, under a normal Gauss distribution, justify an 84% score. (Or for that matter, more than a 10-20% score. He didn't seem to mind the graphics, so I guess it wouldn't make it a perfect 0%.) But an 84% score it did get nevertheless.
I presume just not to hurt anyone's feelings too much.
And you know what bugs me about this kind of attitude? That the scale distorts more and more.
If in one year you gave a crap game a 40% instead of the 10% that it deserved, to spare some people's feelings... you've just distorted the scale some more. A couple more such events and now 40% means total crap. Now to spare feelings you start giving 50% to 60% for that. Oops. The scale got distorted some more.
Give it another decade and we'll have a scale of ratings that goes not from 0 to 100, but from 99% and 100%.
And some people will then foam at the mouth when their favourite game gets only 99.5%. Time to distort the scale some more.
Then what the heck was the purpose of this whole trip?
1. Java's main flaw is that it can't be swapped out (or it can, but the garbage collector brings it all right back in) and takes a shitload of memory. I.e., it's ok for small utilities, or small games like MegaMek, or a single IDE loaded on a 1 GB machine. But try loading even 1 MB more java stuff than the machine has RAM, and it'll crawl.
We actually have to develop server-side java stuff here, but the catch is that we also get to test them locally. Between a local copy of WebSphere, Eclipse, and a few other utilities, you'd be surprised how easy you can get a 1 GB machine to thrash through swap. (And that's more RAM than Joe User would have.) Most of us have learned to get around that by not keeping too much stuff loaded. (Already an exercise that with C programs would not be necessary.)
A coleague obviously hasn't, or rather can't. He also has to keep a Java based XML editor open, and stuff like that. His machine always feels slower than an old ZX Spectrum or Commodore 64. He and another coleague had, last I've heard, requested permission to buy their own computers with 2 GB RAM for work.
Yeah, that't gotta be the future of desktop environments to inflict upon Joe User.
2. Again: Sun's Java Desktop, is _not_ a Java based desktop. It's a bog standard Gnome desktop with a new theme and the exact same JVM that you'd get with any other Linux distro. It's not more geared towards Java than any other Linux distro you could buy.
Now I can see some advantages to it, but "geared towards Java" is _not_ one of them. "Java" is no more than a marketting buzzword there. It's just like Intel's bunnies telling you that a Pentium makes your Internet go faster.
Much as I'm both an AMD fan and living in Germany, I can't help wonder.
I doubt that any silicon fab involves little children crammed in sweatshops, sewing together CPUs for 14 hours a day with rusty dirty sewing machines. It's one big uber-purified and largely automated installation, because that's the only way it can possibly work at all. I.e., I wouldn't expect that much variation in the working conditions.
Plus, it's the first time I've heard that kind of argument to basically mean "let's keep the rich rich, and the poor poor." When people wanted to feel good about themselves about helping other people's condition, it used to mean giving to some charity or such. It's quite the change to see it mean "let's keep the poor unemployed and in poverty."
Now I'm not going to tell you what to do with your money. You can, for all I care, donate them directly to Bill Gates or start a "help the CEOs buy a new yacht" charity, if that's what makes you feel good about yourself.
I _think_ that what most people have in mind by "dynamic plot", really means "AI". The dream isn't to get a game lacking a plot at all, but one which more or less can addapt it to what you're doing.
Same as if you're playing a PnP (Pen-and-Paper) game with some buddies, the GM (Game Master) will accomodate some deviations from what he had in mind. If noone wants to go talk to the dwarves next, the GM will come up with something else.
Of course, that as such is impossible. But it maybe _some_ freedom wouldn't be that impossible.E.g., in Fallout 2, while you did generally follow the same plot, there were plenty of ways to solve each particular situation.
To take my favourite example: the Navarro base. You could do the dumb warrior thing and do a frontal assault, guns blazing, and hope you take out the turrets before they take you out. Or you could be a diplomat and bullshit the guards that you're a recruit, bullshit the drill sergeant, bullshit the quartermaster into giving you armour and a weapon, bullshit the officer into just giving you what you came for, bullshit the mechanics into giving you the vertibird plans, etc. Or you could sneak in through the back door. Or whatever. Or a combination of the above.
The GM also usually doesn't want to see you all dead. He'll adapt the enemies you meet to your level and skills better than any computer games. (That is, assuming the game even tries. Most don't.)
I've yet to hear of any PnP game where you had to spend the month running around in circles, doing random battles for xp, because otherwise you're not ready for the monsters in the next area. Yet in computer RPGs it happens all the time.
Heck, you could even decide to play a pure diplomat, and then the GM would probably let you talk the arch-villain into giving up instead of fighting him. Or solve the case and call the authorities, instead of fighting the villain yourself. Or solve some puzzle to blow up the villain's doomsday machine and lair, instead of a fight. Whatever.
I'll be the first one to aggree that most case mods look dumb and ugly. They just show a colour cacophony of green PCBs, red PCBs, blue PCBs, copper heatsink there, aluminum over there, coloured cables, and some added leds to make the colour cacophony complete. It looks as artistic as a clown after a tragic makeup accident, or like a bomb attack on a paint factory.
But... "fast"?? How in Odin's name do some ugly led fans and a blue lamp make a computer look fast? I'd really get depressed about some people's IQ if _that_ is their idea of a fast computer.
Well, maybe, but not even trying is IMHO the whole problem. There are plenty of "realistic" scenarios that aren't judgmental, nor preaching a particular ideology.
E.g.:
- SimCity was already mentioned in the article: it does not try to tell you that this ideology is better, or that other ideology is to blame.
- Capitalism II would sound like an obvious target to pick on, but here's the rub: it doesn't actually tell you that capitalism is better. It just is. You're a capitalist in an ideal capitalist world (i.e., something as inexistant as ideal communism), now go make some money. Nowhere does it say "well, see, if it was communism, everyone would starve now."
Nor does it try to spoon-feed you any particular consequence of your actions, to reinforce some ideology. It does not even attempt to judge social effects of your paying larger salaries, or being more environment friendly, or whatever. It doesn't tell you "greedy republicans have caused a recession" or "bleedin' heart socialists are costing the economy a fortune." It stays neutral.
- Steel Panthers is one of the most realistic turn-based WW2 simulations. It takes more factors into account (e.g., armour slope vs trajectory height) than the whole C&C series combined. You'd think it would be a logical choice for getting preachy, right?
But it doesn't preach anything. It just happens in WW2. We all know what happened, we all know who was evil and who was guilty of what. So the game doesn't need to keep stappling upon your forehead "you're an evil sonofabitch for playing a Wehrmacht officer" nor "you've chosen to play as one of Stalin's minions, you commie traitor". If you want to play with German Tigers or Soviet JS-3 tanks, there you go.
- I'll even say that Jet Set Radio isn't necessarily _that_ guilty. A future in which _someone_ is an oppressive dictator, isn't that hard to imagine. The problem would be if the game force-fed you ideological stuff like _who_ is going to be the evil dictator, or _why_ is your country going to the dogs.
Now contrast it with other games which feel a need to not only preach, but preach total unproven bullshit. They don't even feed you some neutral fact, like "by 2034, computers should be 1 million times faster", they feed you an _ideology_. A _dogma_.
Like that in the next 30 years we're all going to be flooded because we didn't stop global warming. Never mind that not even most meteorologists aggree on that, never mind that only 2% of greenhouse gasses are produced by humans, never mind that the last 20 years have brought a steady global _cooling_, never mind that even before that the grand total warming was 1 degree in a whole bloody century, never mind the evidence that it may just be a change in how much heat the sun sends this way, etc.
I.e., regardless of whether you want to believe in it or in the contrary, it's something still not proven and still not understood. But the game already feeds you one particular political view, and shows you some "consequences" like they're a proven scientiffic fact.
Or like this or that economic approach _will_ bring us all into poverty and oppression. Sorry, as the saying goes, even if you put all the world's economists end to end, they still wouldn't reach a conclusion. There are people who've studied all their life how the economy works, and they _still_ don't fully understand it. If it was that simple, obvious and fully understood, everyone would know exactly what to do to get perpetual growth and never a recession.
But no, a game designer with no clue of the economy, feels like he's more qualified than thousands of economists. Not only he knows what't the One True Way (TM) the economy works, he also can extrapolate and show you an accurate result of what it will be like in 100 years from now. And exactly why.
That's funny, 'cause I've actually worked with Oracle on both Sun and Intel hardware. The whole "but Suns are better at IO" or "but Suns are more scalable under load" myths, are just that: myths.
Not surprisingly, a PC with DDR RAM actually beats a Sun with SDR RAM every time.
" Clearly IBM has decided that time wasted on Linux is more profitable (financially or otherwise) than time spent with Windows."
Actually, nope. There never was a version of Windows for IBM mainframes. If you know of one, please do provide a link.
IBM still doesn't do much to push Linux on the desktop.
Where IBM did see Linux as damn good, was as an alternative to their _own_ crap on those mainframes. Which, honestly, given IBM's skill at making software, isn't even much of an achievement.
The whole point about that "made by people who have nothing to do with their time, for people who have nothing to do with their time" phrase was about usability. You do know what usability is, right?
So again: IBM's making fast computers... what does that have to do with usability? Does software become blindingly obvious when you have more than x MFLOPS? No, seriously, I want to know.
And let me tell you something else about IBM. IBM has a long fine tradition of selling crap dysfunctional software cheap, and ultra-expensive consultants to make it work.
E.g., WebSphere.
So I wouldn't be too flattered about Linux fitting into _that_ strategy. IBM is going to be _delighted_ if you have trouble getting your new Linux server to rum, and have to pay for a man-month of their consultants.
IBM, predictably, also doesn't even try to have any usability.
A bit of historical trivia: much of the reason why IBM and Microsoft broke up, years ago, had to do with having different visions for OS/2. IBM thought that Microsoft is wasting their money on crap like coding a GUI. Who the heck would ever want a GUI on their desktop?
So your bringing IBM into it just confirms my point, if anything. Yes, I can see how IBM would be the last one to mind lack of usability.
Now let me ammend that a bit. Linux is nowhere near as bad as IBM's software, and particularly not on the server side. Thank goodness for that. But that's pretty much in spite of IBM.
So, alas, far from getting me in a mindless "whoa, Linux must be good because IBM likes it" zealot trance, it only makes me hope that maybe IBM can learn a little usability from Linux. Not that Linux has that much of it, but even that's several orders of magnitude better than anything IBM ever made. (Less buggy than IBM's stuff too, but that's off topic for usability.)
However, since most uses for that server are 100% based on integer speed (e.g., a web server, application server or database server), the Sun is also more than 50% slower. Unless you're doing 3D rendering on it, that FPU speed is 100% irelevant.
I.e., let me put this otherwise for you: a _two_ way Xeon server would aready run circles around it. And those are even cheaper. In fact, taking a Dell Poweredge 2650 and outfitting it with the most expensive Xeons (3.2 GHz, 2 MB cache), 8 GB DDR, DVD, 2x36 GB hard drives, like in your example, costs... $11,225. Gee wiz... it's still cheaper than the v440 (and would be even cheaper with 3.06 GHz 1MB Xeons), and again: it's still faster than the Sun.
I.e., to repeat what I've said before, you'd need at least an 8 way UltraSparc to come even close to a 4 way Xeon.
OK, so basically in the end you install pre-compiled binaries, without aggressive optimizations. Then, in the constext of what the parent post was saying... the advantage of using Gentoo is... what?
No, honestly. Please do explain the major advantage of using precompiled Gentoo binaries, compiled with GCC, as opposed to installing precompiled Mandrake packages, which are compiled with Intel's compiler? Mandrake's are faster, they're tested, and they have tech support. Yours are... what? L33t by default, just because they bear the Gentoo name? Or?
Are you naturally retarded, or do you have to try hard to look that way?
IBM's making fast computers, and the time _I_ waste configuring various crap on Linux... what do those have to do with each other? Is my time measured on CPU cycles, or what the fsck? Does that fast IBM computer reduce the time it takes me to read through a gazillion messages in a mailing list, hoping I'll find exactly where this and that config files are stored?
You know, it's retards like you which give PHBs a bad name. People who think that an Intel CPU makes their internet go faster, just because some Intel ad told them so. Or in your case, that IBM makes their install go faster. Get a clue.
"It's rock-solid stable, and its the speediest of any distro I have tried (no doubt due to all your applications being optimized for your specific system)."
Riight... So you just need to tweak a few flags on GCC and it will handily beat Intel's compiler. Now why didn't Intel think of that?
And you only need to do it 2-3 times, until you find the right flags that are faster than Intel's compiler, but don't make that program randomly core-dump. Is that a great use of my time, or what?;)
Hey, can I have some of those funny cigarettes too?
"If they came out with an "enterprise" version I think I would give it a whirl, I can see it easily being a great fit in my server room."
Rriigghhtt. No doubt, any self-respecting business just has to love a system which takes 2 days of compiling to get running. You know, just in case that hard drive fails. They'll love the whole goddamn site being down while you recompile with portage.
For that matter, no doubt they'll love every single patch and upgrade involving hours of compiling on their production machine. Yeah, noone needed those CPU cycles anyway.
Oh, and I'm sure they'll be delighted to run their server software compiled with your custom flag mix, which occasionally core-dumps, rather than something tested and stable. Server uptime is for lusers, anyway. If you can squeeze 1ms out of the 500ms taken to serve a page (mostly database time), surely that's worth running an unstable and untested home-brewn compiler flag mix.
I guess some geeks just have no business sense, and that's that.
"I cannot use Windows with a clear conscience because of IE's and Outlook's persistent security failures. [...] The difference is that if someone finds a bug in Mozilla that puts me or my network at risk, I can wipe it clean from my hard drive and fall back on alternative software packages."
To quote Red-vs-Blue: "Whop-de-fucking-do." So under Windows surely you can't possibly use Mozilla or Opera instead of IE. Boo hoo. The bad evil Microsoft men in black are holding a gun to your head and forcing you to only use IE.
Funny how I never had that problem. Whatever sites work well in Opera or Mozilla under Linux, blimey, seem to work just as well in Opera or Mozilla under Windows. And those who don't (and that's the whole problem), don't under Linux either.
"Add in IIS for Windows incarnations with IIS installed an running."
Oh yes, presumably the same evil Microsoft men-in-black are also preventing you from activating the firewall;)
"Cost is another obvious difference, but one that I think will eventually catch up to Microsoft more than any antitrust case or business practice."
On/. I'll probably get modded down as a troll, but here goes anyway: cost is exactly why I like Windows. Because my time is more valuable. If paying a couple hundred to MS saves me from weeks of tracking down dependencies and recompiling everything just to get a damn game to run in WineX, then that copy of Windows has already paid for itself.
"There are a growing number of open source software projects that meet or even exceed their commercial competitors capabilities. OpenOffice, Mozilla, and Apache to name a few."
As long as you don't mind that OpenOffice lacks about 80% of MS Office's functionality, and makes a bloody mess out of half the MS Word documents you might have to use. Yeah, I'm sure every business is in a position to be high and mighty and tell all their clients and suppliers "ha ha, very funny, now stop sending me Word documents."
And as long as you don't mind that Mozilla makes a mess out of about 20% of the sites you'd ever want to view with it. Yeah, I'm sure Mom will be delighted to give up playing backgammon on the MSN site, just to support a stupid browser war. I'm also sure she'll be thrilled to move her money to another bank, just because her bank doesn't support anything but IE. Sure thing. I'll uninstall her IE right away;)
For that matter, I'm sure she'll be thrilled to throw away the PaintShop Pro she just learned to use for her digital camera, and spend more time learning Gimp's interface. (Which is probably a matter of preferences, but it makes _me_ cringe.) Or going through the same fun that ESR ranted about recently to get those pics printed over the network, on the printer on Dad's computer.
Yes, you could argue that none of those are necessarily a fault of Mozilla or OOo itself. And I'd even aggree. But the fact remains: for most people it's more hassle than it's worth. They don't care if that site's getting rendered all wrong is the fault of Mozilla or of the web designer. It just doesn't show right on Mozilla.
Actually, while 256 MB is indeed a joke for KDE and OpenOffice nowadays, you don't have to stick to the most primitive stuff to get the work done.
E.g., check out XFCE 4. Yes, the default look is like CDE, and I can see how that would put many people off. But if you take some time to rearrage it to a more comfortable layout (e.g., icon bar on the side, and configured so the windows never overlap it), it pretty much does all I've ever wanted from a window and desktop manager. Even in Windows.
Or you could run a combination of ICEwm and DFM. DFM gives you icons on the desktop and a file manager, while ICEwm gives you a task bar and a start menu. Together they take a mere couple of MB, and give you most of the Windows desktop functionality. (I.e., no HTML on the desktop or other useless bells and whistles which I always disabled on Windows anyway.)
Either alternative looks and acts modern enough, starts up _way_ faster than either KDE or WinXP, and will leave you with plenty of memory left to run OpenOffice even on that machine.
I guess all I'm saying is that the Linux choices aren't as polar as "either the most bloated stuff ever made, or rock bottom 1980's functionality." There is plenty of stuff out there which falls somewhere in the middle.
Dunno about him, but _I_ have better things to do with my time than compiling and rolling together my own Linux distro. Gentoo (since it's also in the news today) and its _primitive_ stone-age install was already enough of a waste of my time. I really _don't_ want to get any more extreme than that, thank ye very much.
And in fact, I'll say that this is _the_ problem with Linux. It's made by people who have nothing to do with their time, for people who have nothing to do with their time.
To get back on topic: It may well eventually overtake Windows, but that's when a whole different lot of people get into the act. People who don't thing "whoa, this is sooo cool... I dug through their sources and for a week, and read the newsgroups for 4 hours a day, and I figured it out. I'm sooo l333t." Instead it will take people who think "fsck it, I don't have time for this crap. I just want to press a button or two and have this configured, tested and running. I want it to do the repetitive menial tasks (like selecting the initial mirror in Gentoo) automatically, not make me do that through a text mode browser and command line. And if it knows that I'll also need to configure XYZ next, then it should jolly well do that for me, not expect me to manually launch yet another command line utility. And I want it to bloody remember my choices, so it doesn't make me configure the exact same DSL connection _again_ half an hour later."
I.e., people with the exact opposite mentality than whoever came with the Gentoo install. _Then_ Linux will be ready for Joe Average.
Which is just more FUD and BS to make PHBs feel good about buying Sun's crap. Sorry, software failure with Windows at least makes some more believable FUD, but for hardware failures Sun is no better than anyone else.
Seein' as, for example, the SCSI or FC hard drives in a Sun are the _exact_ same drives, from the _exact_ same manufacturers. Or that the RAM is made of the _exact_ same RAM sticks you can have in your mom's old K6. Or have you seen the heatsinks on Sun's CPUs? Well, then fsck off, because the aluminum heatsink on my Athlon 64 is actually smaller.
And you know, it's funny how our Linux servers at the previous company held just fine under heavy load. Or our main client's computers. They had a system which supported hundreds of users concurrently, and it ran on 3 cheap commodity PCs running Linux.
The 3 computer cluster was purely "just in case" one of them fails, because one of them was perfectly capable of taking that load without breaking a sweat. We actually tried. (Of course, there was some good programming involved, not taking the most baroque framework possible and a farm of unskilled monkeys.) And you know what? None of them actually failed under load.
Which also takes care of the "but Suns have hot swap!" BS. You could pretty much hot-swap a whole server in that config. Just put Tomcat on another PC, spend 5 minutes copying our app on it, and there you go: you have a replacement for a whole computer. Or take one off line for repairs, if needed, and the users wouldn't even notice.
They could be routed to another server in the middle of anything, without losing their session or any data. Heck, without even noticing that their previous server is being messed with. (Did I mention good programming? I think so.)
Sorry, it's not just that PCs are cheap. IMHO Sun has forgot how to design a CPU. Or a chipset. Sorry, 1.2 GHz just doesn't cut it, no matter what IPC you have. Doubly so when the best you can offer with that CPU consists of:
- SDR RAM in an age when everyone's moved on to DDR
- 32 bit memory bus, when even the original Pentium in the 90's had a 64 bit bus
- crap outdated components and cards at ludicrious prices (E.g., is that an ancient ATI Rage that they're selling for almost $500? Well, gee, in the PC world you can already get a GeForce 6800 for that kind of money.)
Sorry, Suns just don't cut it. You'd need somewhere between 8 and 16 of the latest UltraSparcs in a box, to even touch a cheap 4 way Xeon for a server. And you can check out for yourself what the Sun would cost in that configuration.
Charging that kind of ludicrious prices was justifiable when they at least had the bang to claim _some_ advantage over a PC. Not when a Pentium 4 or K8 or G5 (pick your favourite there) is running circles around the UltraSparc.
Sun has fallen so far behind the technology curve, it's not even funny. And firing most of their R&D stuff doesn't give me any confidence that they'll spring back to having a competitive computer any time soon either.
Honestly, I just can't recommend a Sun with a straight face any more. Do you want a unix-y workstation? Get a cheap PC, install your favourite linux distro on it, and there you go. It'll run circles around a Sun. Do you want a RISC Unix workstation? Get a Mac. (And I'm not even a Mac fan.) Do you want a Unix server? Same thing.
Will they die? Strictly speaking: probably not. They can always turn into yet another Dell, packing together PCs from components made by others.
But it sure won't be the same Sun. In a sense, the old Sun as we knew it, _will_ die. To be mean: and in a sense, good riddance.
Well, to be honest, IMHO the browser should have never been used for half the stuff it's used today for.
From a programmer viewpoint, it's a throw back to the dumb terminals of the 70's. Only with more overhead, and without a permanent connection.
The stateless nature of HTTP meant that essentially if you needed state, you had to do your own buggy approximation of a stateful connection. (E.g., the various "HTTP session" implementation in various languages or frameworks.)
The problem however, is that I've yet to see a project which didn't have session-related problems. Ranging from mild usability problems (e.g., you lost all your data when your session timed out), to waste of resources (someone stored an object in the session, which in turn pointed to a whole 100 MB tree), to identity theft (e.g., when enabling one to revive an older session id), all the way to such catastrophes as having users marked "new product" and products marked "new user" when someone used two windows at the same time.
And a lot of other stuff was literally based on leaps of faith. Like having faith that the user will never open a link in a second window, in the catastrophe implementation mentioned above. Or having faith that when you've opened a second window for your dialog, the user won't do other stuff with the previous window in the meantime. Or various other versions of assuming that the user will always mindlessly stick to the exact workflow you had in mind.
Of course, you think that such problems could and will never happen to your project. Think again. They even appear in such high profile projects as IBM WebSphere's admin console.
So IMHO if spammers and ad provider idiots manage to finally kill the idea that HTTP is _the_ only medium for a program to run over, good riddance.
I don't think many people have anything against advertising as a whole. But, yes, being obnoxious and annoying to your potential buyers is what's the real problem here.
As I've said it before, in the early days of the web the sites would have maybe one banner on the main site, and that was it. No animations, no huge flash movies, no pop-ups and no pop-unders. And most definitely no trying to install spyware crap on your computer.
And noone wanted to block those. Heck, some of us even clicked on them. In fact, I'll say that if back then someone would have come and said that they made an ad blocker, most people would have just said "you're lame, and you have too much time on your hands."
But in the meantime things have changed. Sorry, no matter how I much want to turn it around in my haead, and to find an excuse, I can't find any valid excuse for today's sorry state.
Now I'm bombarded with a barrage of pop-up, pop-unders, 500k flash movies, and even downright redirects to some vendor's site, before I even see what the site has to offer. Do I even want that site's "free services"? No idea, I'll first have to close all these windows and say "no" to their generous offer to install a dialer on my computer.
And in the end, I'll be mean and say that 99% of those sites didn't offer anything I wanted anyway. The mentality seems to be that any retard who can't even string together coherent sentences, has a right to make an income with ads. Which I'll disaggree.
Doing fast Google searches is one thing, but you may notice how Google doesn't need lame pop-unders to stay in business. You may also notice how Google not only had a service to offer, but also a _business_ _model_ for it. (Yes, using that search engine to offer targetted relevant text ads.)
And that's one site that deserves to stay in business. That's what capitalism is about.
On the other hand, let's look at the sites that do need obnoxious ads to stay in business. Yes, the millions of whiny blogs, or "look ma, I too can copy news off slashdot" news sites. They seem to assume that just having a stupid web site makes them so invaluable, that they have a _right_ to annoy me for a buck. Sorry, no. It doesn't work that way.
Slashdot Mirror
Look, I'm against piracy. Not only I wouldn't give them any updates, but I think pirates should swing from the mast, like in 1600. Or make them walk the plank.
But I also think there's a reason why there's "BS" in "BSA". Their statistics make me want to puke. They do such bullshit statistics as taking a pirated CD from Taiwan or China which includes some expensive piece of softwware, like 3D Studio Max, and say "see, there's 5000 USD worth of software on this CD. We think 10,000 chinese kids bought this CD for 5$, which robbed us of 50,000,000 USD."
That's utter bullshit. Most of those Asian and Eastern European pirates do not need 3D Studio Max and wouldn't buy it anyway, even if they could afford to. (Which they can't. As was said before, a chinese family would need to pay _all_ their income for _two_ _years_ to afford a license. Again: _all_ their income. That is, leaving them with no money for food, rent, clothes, etc.)
We're not talking 10,000 professional designers and architects who actually need it, we're talking mostly kids who much around with it a bit to make some skins for mods for old games. Maybe 1 of them will actually release an obscure mod, the rest just mucked a round a bit with it, uninstalled it and moved on to something else.
Would all 10,000 of them have bought 3D Studio Max if they couldn't pirate it? No. _I_ wouldn't buy it either, much as (1) I could easily afford it, and (2) I'm tempted to try modding "X2 - The Threat." (Which, sadly, only supports exporting stuff from 3DS MAX.) Now I don't pirate it either, but even I think it would be utterly retarded to pay $4000 on tools to mod a $40 game.
Yet the BSA would want me to believe that 10,000 dirt-poor kids from Taiwan would. That's so much bullshit, it could fertilize a few acres.
Ahh... Ok, I suppose O(n^2) can make one think of bubble sort.
Given enough lack of clue or lack of care for performance, though, one can squeeze an O(n^2) or worse in lots of other places. E.g., it's pretty easy to end up doing an O(n^2) even for searching through a table.
Or an example of clueless coding that's particularly dear to me, comes from some guys who basically tried to shaft me back in my freelancer days. Except they couldn't work with trees. At all.
They had data which was a simple binary tree. Or rather the perfect textbook example for something fit to be stored in a tree. (A big rectangle which could be cut horizontally or vertically into two. Each of which could then be cut in two. And so on.) They needed to save and load it.
But they couldn't work with trees. So they replaced the tree with an array, losing all the tree data in the process. (No, not a tree-inna-array, with indexes instead of pointers. Juts a flat array with no pointers.) And saved it as such. Which upon reading the file left them with the problem of which piece depends on which one. So they used backtracking to find out which pieces have already been cut.
The result? It took minutes to load even the smallest trees. For a piece of processing which shouldn't even have existed at all, if they only knew how to recursively traverse a tree.
Sad but true.
Guess the wonder of one liners is that it just lets me wondering what you meant.
:)
A binary search is O(log2(n)), which is to say, searching through 1 million records takes a mere 15 steps more than searching through 32 records. So there's nothing particularly fantasy about them. Such algorithms exist.
Other kinds of algorithms involve even less steps. E.g., the kinds of trees used in an English spell-checker (well, some spell checkers anyway), have 26 as a base for the logarithm. They also have other advantages over a binary search, such as being able to calculate a "distance" between two words.
Can they be beat? Sure. A hash table is O(1) for retrieving an element, so it executes faster than a binary search for most data sets. (I.e., given enough data so that calculating the hash isn't slowing you down. For string keys this pretty much means every time.)
But I'm guessing you knew such elementary things already, and I'm probably just boring you. So if you'd kindly clarify what you had in mind, I might even be able to give a less boring answer
No offense, but _why_ would you even need to think about such things as gender in an inanimate tool? Why do you need something like that to be comfortable around a piece of metal, silicon and epoxy? No, I'm really curious. Why?
My computer is an "it" and I'm very comfortable with that. Never felt any more need to assign it a gender than I needed to assign a gender to the desk it sits on, or to the chair in front of it.
I also can't help notice that the most successful interfaces amd devices (as opposed to gee-wow crap that's hyped by marketroids and then dies nevertheless) is stuff that's abstract and not over the top in simulating real life devices or beings. Humans are very capable of learning new and abstract things.
E.g., probably the most successful such device is the mouse. And there is no real life equivalent of it. But people nevertheless have no problem in learning to use it. E.g., the context sensitive popup-menu. There is nothing in real life which acts that way, and definitely nothing to anthropomorphise about it. But everyone learns very quickly to appreciate it. E.g., the hyperlink. Again, there is nothing in real life books, nor on real life humans, which acts that way. But it works so well, that it's used not only in browsers, but in other programs and games too. Etc.
By contrast, such hyped attempts at simulating the real world as Microsoft Bob, died a silent death and noone cared about them. Invariably such attempts waste more screen space for what a simple icon would have done just as well, and end up requiring more clicks to get to the same place. As a result, they end up requiring more time and effort to _discover_ where you really want to get, if you didn't know that already. ("Discoverability" is one of the main issues of UI design.)
So how would going even more extreme, and even assigning Bob a gender help UI design anyway? If you have actually put some thought into design and usability and ended up with a provable need for gender, I'd really like to hear about your findings.
How would a definitively feminine or masculine interface work, anyway? What would it need to have, that the current UIs don't, and which can't possibly be done without assigning it a gender? (Have a breast instead of a mouse?;)
And how would you convey the idea that your particular UI is definitively masculine or feminine to your users, without resorting to insulting over-exaggerated stereotypes?
I can just picture having a big animated 3D female head popping up instead of clippy, with a blank look on her face, giggling and twirling her hair. "Hi. I'm, like, Cindy. I'm, like, your UI helper. Isn't it rad? Hihihi. I, like, notice that you're, like, typing stuff. And it, like, looks like an URL. Do you, like, want me to, like, convert it to HTML for you?" *click* "Hihihi. Oopsie. I'll, like, leave it alone for now, then."
It's _not_ what I think of females, but I can see some lonely male geek using that stereotype. The same guys who think that tiny chainmail bikinis are good female armour, for example. Or a worse stereotype.
Ok, so maybe you'll avoid ending up this stereotyped. So how would you convey the idea that your GUI is definitively all female, to all your users, most of which never even thought that a GUI has or needs a gender?
Well, yes and no.
I still don't think you should start doing every single silly trick in your code, like unrolling loops by hand, unless there's a provable need to do so. Write clearly, use comments, and use a profiler to see what needs to be optimized.
That is coming from someone who used to write assembly, btw.
But here's the other side of the coin: I don't think he included better algorithms in the "premature optimization". And the same goes for having some clue of your underlying machine and architecture. And there's where most of the problem lies nowadays.
E.g., there is no way in heck that an O(n * n) algorithm can beat an O(log(n)) algorithm for large data sets, and data sets _are_ getting larger. No matter how much loop unrolling you do, no matter how you cleverly replaced the loops to count downwards, it just won't. At best you'll manage to fool yourself that it runs fast enough on those 100 record test cases. Then it goes productive with a database with 350,000 records. (And that's a small one nowadays.) Poof, it needs two days to complete now.
And no hardware in the world will save you from that kind of a performance problem.
E.g., if most of the program's time is spent waiting for a database, there's no point in unrolling loops and such. You'll save... what? 100 CPU cycles, when you wait 100,000,000 cycles or more for a single SQL query? On the other hand, you'd be surprised how much of a difference can it make if you retrieve the data in a single SQL query, instead of causing a flurry of 1000 individual connect-query-close sequences.
(And you'd also be surprised how many clueless monkeys design their architecture without ever thinking of the database. They end up with a beautiful class architecture on paper, but a catastrophic flurry of querries when they actually have to read and write it.)
E.g., if you're using EJB, it's a pointless exercise to optimize 100 CPU cycles away, when the RMI/IIOP remote call's overhead is at least somewhere between 1,000,000 and 2,000,000 CPU cycles by itself. That is, assuming that you don't also have network latency adding to that RPC time. On the other hand, optimizing the very design of your application, so it only uses 1 or 2 RPC calls, instead of a flurry of 1000 remote calls to individual getters and setters... well, that might just make or break the performance.
(And again, you'd be surprised how many people don't even know that those overheads exist. Much less actually design with them in mind.)
So in a nutshell, what I'm saying is: Optimize the algorithm and design, before you jump in to do silly micro-level tricks. That's where the real money is.
Well, what sorta bugs me is that this kind of inflating scores has crept everywhere. E.g., let's take something fairly neutral for most people: game reviews.
Well, except for some fanboys who'll foam at the mouth if you gave their idol's game only a 99% rating. Bonus points if they haven't even played the fscking game yet, but they just _know_ it's a perfect 100%.
I've seen for example a review on a major review site, for an overhyped game that I truly hated. The reviewer went to great lengths to demolish every single aspect of it, and for good reason. The review told you just how much every single bit and detail and design decision sucked. The controls, the interface, the physics, the quests, etc. He sounded like he hated it all. It even ended with such damning words as saying that the game might be ok for someone who's never ever played another game before, and thus can't compare it to anything, but won't cut it for anyone else.
Then they give it some rating like 84%. Uh? Excuse me? Regardless of what _I_ thought about the game, that reviewer himself had wrote a hate-fest about the game. He had savaged it thoroughly. There was _nothing_ in his review that would, under a normal Gauss distribution, justify an 84% score. (Or for that matter, more than a 10-20% score. He didn't seem to mind the graphics, so I guess it wouldn't make it a perfect 0%.) But an 84% score it did get nevertheless.
I presume just not to hurt anyone's feelings too much.
And you know what bugs me about this kind of attitude? That the scale distorts more and more.
If in one year you gave a crap game a 40% instead of the 10% that it deserved, to spare some people's feelings... you've just distorted the scale some more. A couple more such events and now 40% means total crap. Now to spare feelings you start giving 50% to 60% for that. Oops. The scale got distorted some more.
Give it another decade and we'll have a scale of ratings that goes not from 0 to 100, but from 99% and 100%.
And some people will then foam at the mouth when their favourite game gets only 99.5%. Time to distort the scale some more.
Then what the heck was the purpose of this whole trip?
Let me just say two things:
1. Java's main flaw is that it can't be swapped out (or it can, but the garbage collector brings it all right back in) and takes a shitload of memory. I.e., it's ok for small utilities, or small games like MegaMek, or a single IDE loaded on a 1 GB machine. But try loading even 1 MB more java stuff than the machine has RAM, and it'll crawl.
We actually have to develop server-side java stuff here, but the catch is that we also get to test them locally. Between a local copy of WebSphere, Eclipse, and a few other utilities, you'd be surprised how easy you can get a 1 GB machine to thrash through swap. (And that's more RAM than Joe User would have.) Most of us have learned to get around that by not keeping too much stuff loaded. (Already an exercise that with C programs would not be necessary.)
A coleague obviously hasn't, or rather can't. He also has to keep a Java based XML editor open, and stuff like that. His machine always feels slower than an old ZX Spectrum or Commodore 64. He and another coleague had, last I've heard, requested permission to buy their own computers with 2 GB RAM for work.
Yeah, that't gotta be the future of desktop environments to inflict upon Joe User.
2. Again: Sun's Java Desktop, is _not_ a Java based desktop. It's a bog standard Gnome desktop with a new theme and the exact same JVM that you'd get with any other Linux distro. It's not more geared towards Java than any other Linux distro you could buy.
Now I can see some advantages to it, but "geared towards Java" is _not_ one of them. "Java" is no more than a marketting buzzword there. It's just like Intel's bunnies telling you that a Pentium makes your Internet go faster.
Much as I'm both an AMD fan and living in Germany, I can't help wonder.
I doubt that any silicon fab involves little children crammed in sweatshops, sewing together CPUs for 14 hours a day with rusty dirty sewing machines. It's one big uber-purified and largely automated installation, because that's the only way it can possibly work at all. I.e., I wouldn't expect that much variation in the working conditions.
Plus, it's the first time I've heard that kind of argument to basically mean "let's keep the rich rich, and the poor poor." When people wanted to feel good about themselves about helping other people's condition, it used to mean giving to some charity or such. It's quite the change to see it mean "let's keep the poor unemployed and in poverty."
Now I'm not going to tell you what to do with your money. You can, for all I care, donate them directly to Bill Gates or start a "help the CEOs buy a new yacht" charity, if that's what makes you feel good about yourself.
But I just have to wonder...
I _think_ that what most people have in mind by "dynamic plot", really means "AI". The dream isn't to get a game lacking a plot at all, but one which more or less can addapt it to what you're doing.
Same as if you're playing a PnP (Pen-and-Paper) game with some buddies, the GM (Game Master) will accomodate some deviations from what he had in mind. If noone wants to go talk to the dwarves next, the GM will come up with something else.
Of course, that as such is impossible. But it maybe _some_ freedom wouldn't be that impossible.E.g., in Fallout 2, while you did generally follow the same plot, there were plenty of ways to solve each particular situation.
To take my favourite example: the Navarro base. You could do the dumb warrior thing and do a frontal assault, guns blazing, and hope you take out the turrets before they take you out. Or you could be a diplomat and bullshit the guards that you're a recruit, bullshit the drill sergeant, bullshit the quartermaster into giving you armour and a weapon, bullshit the officer into just giving you what you came for, bullshit the mechanics into giving you the vertibird plans, etc. Or you could sneak in through the back door. Or whatever. Or a combination of the above.
The GM also usually doesn't want to see you all dead. He'll adapt the enemies you meet to your level and skills better than any computer games. (That is, assuming the game even tries. Most don't.)
I've yet to hear of any PnP game where you had to spend the month running around in circles, doing random battles for xp, because otherwise you're not ready for the monsters in the next area. Yet in computer RPGs it happens all the time.
Heck, you could even decide to play a pure diplomat, and then the GM would probably let you talk the arch-villain into giving up instead of fighting him. Or solve the case and call the authorities, instead of fighting the villain yourself. Or solve some puzzle to blow up the villain's doomsday machine and lair, instead of a fight. Whatever.
I'll be the first one to aggree that most case mods look dumb and ugly. They just show a colour cacophony of green PCBs, red PCBs, blue PCBs, copper heatsink there, aluminum over there, coloured cables, and some added leds to make the colour cacophony complete. It looks as artistic as a clown after a tragic makeup accident, or like a bomb attack on a paint factory.
But... "fast"?? How in Odin's name do some ugly led fans and a blue lamp make a computer look fast? I'd really get depressed about some people's IQ if _that_ is their idea of a fast computer.
Well, maybe, but not even trying is IMHO the whole problem. There are plenty of "realistic" scenarios that aren't judgmental, nor preaching a particular ideology.
E.g.:
- SimCity was already mentioned in the article: it does not try to tell you that this ideology is better, or that other ideology is to blame.
- Capitalism II would sound like an obvious target to pick on, but here's the rub: it doesn't actually tell you that capitalism is better. It just is. You're a capitalist in an ideal capitalist world (i.e., something as inexistant as ideal communism), now go make some money. Nowhere does it say "well, see, if it was communism, everyone would starve now."
Nor does it try to spoon-feed you any particular consequence of your actions, to reinforce some ideology. It does not even attempt to judge social effects of your paying larger salaries, or being more environment friendly, or whatever. It doesn't tell you "greedy republicans have caused a recession" or "bleedin' heart socialists are costing the economy a fortune." It stays neutral.
- Steel Panthers is one of the most realistic turn-based WW2 simulations. It takes more factors into account (e.g., armour slope vs trajectory height) than the whole C&C series combined. You'd think it would be a logical choice for getting preachy, right?
But it doesn't preach anything. It just happens in WW2. We all know what happened, we all know who was evil and who was guilty of what. So the game doesn't need to keep stappling upon your forehead "you're an evil sonofabitch for playing a Wehrmacht officer" nor "you've chosen to play as one of Stalin's minions, you commie traitor". If you want to play with German Tigers or Soviet JS-3 tanks, there you go.
- I'll even say that Jet Set Radio isn't necessarily _that_ guilty. A future in which _someone_ is an oppressive dictator, isn't that hard to imagine. The problem would be if the game force-fed you ideological stuff like _who_ is going to be the evil dictator, or _why_ is your country going to the dogs.
Now contrast it with other games which feel a need to not only preach, but preach total unproven bullshit. They don't even feed you some neutral fact, like "by 2034, computers should be 1 million times faster", they feed you an _ideology_. A _dogma_.
Like that in the next 30 years we're all going to be flooded because we didn't stop global warming. Never mind that not even most meteorologists aggree on that, never mind that only 2% of greenhouse gasses are produced by humans, never mind that the last 20 years have brought a steady global _cooling_, never mind that even before that the grand total warming was 1 degree in a whole bloody century, never mind the evidence that it may just be a change in how much heat the sun sends this way, etc.
I.e., regardless of whether you want to believe in it or in the contrary, it's something still not proven and still not understood. But the game already feeds you one particular political view, and shows you some "consequences" like they're a proven scientiffic fact.
Or like this or that economic approach _will_ bring us all into poverty and oppression. Sorry, as the saying goes, even if you put all the world's economists end to end, they still wouldn't reach a conclusion. There are people who've studied all their life how the economy works, and they _still_ don't fully understand it. If it was that simple, obvious and fully understood, everyone would know exactly what to do to get perpetual growth and never a recession.
But no, a game designer with no clue of the economy, feels like he's more qualified than thousands of economists. Not only he knows what't the One True Way (TM) the economy works, he also can extrapolate and show you an accurate result of what it will be like in 100 years from now. And exactly why.
Laughable.
That's funny, 'cause I've actually worked with Oracle on both Sun and Intel hardware. The whole "but Suns are better at IO" or "but Suns are more scalable under load" myths, are just that: myths.
Not surprisingly, a PC with DDR RAM actually beats a Sun with SDR RAM every time.
" Clearly IBM has decided that time wasted on Linux is more profitable (financially or otherwise) than time spent with Windows."
Actually, nope. There never was a version of Windows for IBM mainframes. If you know of one, please do provide a link.
IBM still doesn't do much to push Linux on the desktop.
Where IBM did see Linux as damn good, was as an alternative to their _own_ crap on those mainframes. Which, honestly, given IBM's skill at making software, isn't even much of an achievement.
The whole point about that "made by people who have nothing to do with their time, for people who have nothing to do with their time" phrase was about usability. You do know what usability is, right?
So again: IBM's making fast computers... what does that have to do with usability? Does software become blindingly obvious when you have more than x MFLOPS? No, seriously, I want to know.
And let me tell you something else about IBM. IBM has a long fine tradition of selling crap dysfunctional software cheap, and ultra-expensive consultants to make it work.
E.g., WebSphere.
So I wouldn't be too flattered about Linux fitting into _that_ strategy. IBM is going to be _delighted_ if you have trouble getting your new Linux server to rum, and have to pay for a man-month of their consultants.
IBM, predictably, also doesn't even try to have any usability.
A bit of historical trivia: much of the reason why IBM and Microsoft broke up, years ago, had to do with having different visions for OS/2. IBM thought that Microsoft is wasting their money on crap like coding a GUI. Who the heck would ever want a GUI on their desktop?
So your bringing IBM into it just confirms my point, if anything. Yes, I can see how IBM would be the last one to mind lack of usability.
Now let me ammend that a bit. Linux is nowhere near as bad as IBM's software, and particularly not on the server side. Thank goodness for that. But that's pretty much in spite of IBM.
So, alas, far from getting me in a mindless "whoa, Linux must be good because IBM likes it" zealot trance, it only makes me hope that maybe IBM can learn a little usability from Linux. Not that Linux has that much of it, but even that's several orders of magnitude better than anything IBM ever made. (Less buggy than IBM's stuff too, but that's off topic for usability.)
However, since most uses for that server are 100% based on integer speed (e.g., a web server, application server or database server), the Sun is also more than 50% slower. Unless you're doing 3D rendering on it, that FPU speed is 100% irelevant.
I.e., let me put this otherwise for you: a _two_ way Xeon server would aready run circles around it. And those are even cheaper. In fact, taking a Dell Poweredge 2650 and outfitting it with the most expensive Xeons (3.2 GHz, 2 MB cache), 8 GB DDR, DVD, 2x36 GB hard drives, like in your example, costs... $11,225. Gee wiz... it's still cheaper than the v440 (and would be even cheaper with 3.06 GHz 1MB Xeons), and again: it's still faster than the Sun.
I.e., to repeat what I've said before, you'd need at least an 8 way UltraSparc to come even close to a 4 way Xeon.
OK, so basically in the end you install pre-compiled binaries, without aggressive optimizations. Then, in the constext of what the parent post was saying... the advantage of using Gentoo is... what?
No, honestly. Please do explain the major advantage of using precompiled Gentoo binaries, compiled with GCC, as opposed to installing precompiled Mandrake packages, which are compiled with Intel's compiler? Mandrake's are faster, they're tested, and they have tech support. Yours are... what? L33t by default, just because they bear the Gentoo name? Or?
Are you naturally retarded, or do you have to try hard to look that way?
IBM's making fast computers, and the time _I_ waste configuring various crap on Linux... what do those have to do with each other? Is my time measured on CPU cycles, or what the fsck? Does that fast IBM computer reduce the time it takes me to read through a gazillion messages in a mailing list, hoping I'll find exactly where this and that config files are stored?
You know, it's retards like you which give PHBs a bad name. People who think that an Intel CPU makes their internet go faster, just because some Intel ad told them so. Or in your case, that IBM makes their install go faster. Get a clue.
"It's rock-solid stable, and its the speediest of any distro I have tried (no doubt due to all your applications being optimized for your specific system)."
;)
Riight... So you just need to tweak a few flags on GCC and it will handily beat Intel's compiler. Now why didn't Intel think of that?
And you only need to do it 2-3 times, until you find the right flags that are faster than Intel's compiler, but don't make that program randomly core-dump. Is that a great use of my time, or what?
Hey, can I have some of those funny cigarettes too?
"If they came out with an "enterprise" version I think I would give it a whirl, I can see it easily being a great fit in my server room."
Rriigghhtt. No doubt, any self-respecting business just has to love a system which takes 2 days of compiling to get running. You know, just in case that hard drive fails. They'll love the whole goddamn site being down while you recompile with portage.
For that matter, no doubt they'll love every single patch and upgrade involving hours of compiling on their production machine. Yeah, noone needed those CPU cycles anyway.
Oh, and I'm sure they'll be delighted to run their server software compiled with your custom flag mix, which occasionally core-dumps, rather than something tested and stable. Server uptime is for lusers, anyway. If you can squeeze 1ms out of the 500ms taken to serve a page (mostly database time), surely that's worth running an unstable and untested home-brewn compiler flag mix.
I guess some geeks just have no business sense, and that's that.
"I cannot use Windows with a clear conscience because of IE's and Outlook's persistent security failures. [...] The difference is that if someone finds a bug in Mozilla that puts me or my network at risk, I can wipe it clean from my hard drive and fall back on alternative software packages."
;)
/. I'll probably get modded down as a troll, but here goes anyway: cost is exactly why I like Windows. Because my time is more valuable. If paying a couple hundred to MS saves me from weeks of tracking down dependencies and recompiling everything just to get a damn game to run in WineX, then that copy of Windows has already paid for itself.
;)
To quote Red-vs-Blue: "Whop-de-fucking-do." So under Windows surely you can't possibly use Mozilla or Opera instead of IE. Boo hoo. The bad evil Microsoft men in black are holding a gun to your head and forcing you to only use IE.
Funny how I never had that problem. Whatever sites work well in Opera or Mozilla under Linux, blimey, seem to work just as well in Opera or Mozilla under Windows. And those who don't (and that's the whole problem), don't under Linux either.
"Add in IIS for Windows incarnations with IIS installed an running."
Oh yes, presumably the same evil Microsoft men-in-black are also preventing you from activating the firewall
"Cost is another obvious difference, but one that I think will eventually catch up to Microsoft more than any antitrust case or business practice."
On
"There are a growing number of open source software projects that meet or even exceed their commercial competitors capabilities. OpenOffice, Mozilla, and Apache to name a few."
As long as you don't mind that OpenOffice lacks about 80% of MS Office's functionality, and makes a bloody mess out of half the MS Word documents you might have to use. Yeah, I'm sure every business is in a position to be high and mighty and tell all their clients and suppliers "ha ha, very funny, now stop sending me Word documents."
And as long as you don't mind that Mozilla makes a mess out of about 20% of the sites you'd ever want to view with it. Yeah, I'm sure Mom will be delighted to give up playing backgammon on the MSN site, just to support a stupid browser war. I'm also sure she'll be thrilled to move her money to another bank, just because her bank doesn't support anything but IE. Sure thing. I'll uninstall her IE right away
For that matter, I'm sure she'll be thrilled to throw away the PaintShop Pro she just learned to use for her digital camera, and spend more time learning Gimp's interface. (Which is probably a matter of preferences, but it makes _me_ cringe.) Or going through the same fun that ESR ranted about recently to get those pics printed over the network, on the printer on Dad's computer.
Yes, you could argue that none of those are necessarily a fault of Mozilla or OOo itself. And I'd even aggree. But the fact remains: for most people it's more hassle than it's worth. They don't care if that site's getting rendered all wrong is the fault of Mozilla or of the web designer. It just doesn't show right on Mozilla.
Actually, while 256 MB is indeed a joke for KDE and OpenOffice nowadays, you don't have to stick to the most primitive stuff to get the work done.
E.g., check out XFCE 4. Yes, the default look is like CDE, and I can see how that would put many people off. But if you take some time to rearrage it to a more comfortable layout (e.g., icon bar on the side, and configured so the windows never overlap it), it pretty much does all I've ever wanted from a window and desktop manager. Even in Windows.
Or you could run a combination of ICEwm and DFM. DFM gives you icons on the desktop and a file manager, while ICEwm gives you a task bar and a start menu. Together they take a mere couple of MB, and give you most of the Windows desktop functionality. (I.e., no HTML on the desktop or other useless bells and whistles which I always disabled on Windows anyway.)
Either alternative looks and acts modern enough, starts up _way_ faster than either KDE or WinXP, and will leave you with plenty of memory left to run OpenOffice even on that machine.
I guess all I'm saying is that the Linux choices aren't as polar as "either the most bloated stuff ever made, or rock bottom 1980's functionality." There is plenty of stuff out there which falls somewhere in the middle.
Dunno about him, but _I_ have better things to do with my time than compiling and rolling together my own Linux distro. Gentoo (since it's also in the news today) and its _primitive_ stone-age install was already enough of a waste of my time. I really _don't_ want to get any more extreme than that, thank ye very much.
And in fact, I'll say that this is _the_ problem with Linux. It's made by people who have nothing to do with their time, for people who have nothing to do with their time.
To get back on topic: It may well eventually overtake Windows, but that's when a whole different lot of people get into the act. People who don't thing "whoa, this is sooo cool... I dug through their sources and for a week, and read the newsgroups for 4 hours a day, and I figured it out. I'm sooo l333t." Instead it will take people who think "fsck it, I don't have time for this crap. I just want to press a button or two and have this configured, tested and running. I want it to do the repetitive menial tasks (like selecting the initial mirror in Gentoo) automatically, not make me do that through a text mode browser and command line. And if it knows that I'll also need to configure XYZ next, then it should jolly well do that for me, not expect me to manually launch yet another command line utility. And I want it to bloody remember my choices, so it doesn't make me configure the exact same DSL connection _again_ half an hour later."
I.e., people with the exact opposite mentality than whoever came with the Gentoo install. _Then_ Linux will be ready for Joe Average.
Which is just more FUD and BS to make PHBs feel good about buying Sun's crap. Sorry, software failure with Windows at least makes some more believable FUD, but for hardware failures Sun is no better than anyone else.
Seein' as, for example, the SCSI or FC hard drives in a Sun are the _exact_ same drives, from the _exact_ same manufacturers. Or that the RAM is made of the _exact_ same RAM sticks you can have in your mom's old K6. Or have you seen the heatsinks on Sun's CPUs? Well, then fsck off, because the aluminum heatsink on my Athlon 64 is actually smaller.
And you know, it's funny how our Linux servers at the previous company held just fine under heavy load. Or our main client's computers. They had a system which supported hundreds of users concurrently, and it ran on 3 cheap commodity PCs running Linux.
The 3 computer cluster was purely "just in case" one of them fails, because one of them was perfectly capable of taking that load without breaking a sweat. We actually tried. (Of course, there was some good programming involved, not taking the most baroque framework possible and a farm of unskilled monkeys.) And you know what? None of them actually failed under load.
Which also takes care of the "but Suns have hot swap!" BS. You could pretty much hot-swap a whole server in that config. Just put Tomcat on another PC, spend 5 minutes copying our app on it, and there you go: you have a replacement for a whole computer. Or take one off line for repairs, if needed, and the users wouldn't even notice.
They could be routed to another server in the middle of anything, without losing their session or any data. Heck, without even noticing that their previous server is being messed with. (Did I mention good programming? I think so.)
Sorry, it's not just that PCs are cheap. IMHO Sun has forgot how to design a CPU. Or a chipset. Sorry, 1.2 GHz just doesn't cut it, no matter what IPC you have. Doubly so when the best you can offer with that CPU consists of:
- SDR RAM in an age when everyone's moved on to DDR
- 32 bit memory bus, when even the original Pentium in the 90's had a 64 bit bus
- crap outdated components and cards at ludicrious prices (E.g., is that an ancient ATI Rage that they're selling for almost $500? Well, gee, in the PC world you can already get a GeForce 6800 for that kind of money.)
Sorry, Suns just don't cut it. You'd need somewhere between 8 and 16 of the latest UltraSparcs in a box, to even touch a cheap 4 way Xeon for a server. And you can check out for yourself what the Sun would cost in that configuration.
Charging that kind of ludicrious prices was justifiable when they at least had the bang to claim _some_ advantage over a PC. Not when a Pentium 4 or K8 or G5 (pick your favourite there) is running circles around the UltraSparc.
Sun has fallen so far behind the technology curve, it's not even funny. And firing most of their R&D stuff doesn't give me any confidence that they'll spring back to having a competitive computer any time soon either.
Honestly, I just can't recommend a Sun with a straight face any more. Do you want a unix-y workstation? Get a cheap PC, install your favourite linux distro on it, and there you go. It'll run circles around a Sun. Do you want a RISC Unix workstation? Get a Mac. (And I'm not even a Mac fan.) Do you want a Unix server? Same thing.
Will they die? Strictly speaking: probably not. They can always turn into yet another Dell, packing together PCs from components made by others.
But it sure won't be the same Sun. In a sense, the old Sun as we knew it, _will_ die. To be mean: and in a sense, good riddance.
Well, to be honest, IMHO the browser should have never been used for half the stuff it's used today for.
From a programmer viewpoint, it's a throw back to the dumb terminals of the 70's. Only with more overhead, and without a permanent connection.
The stateless nature of HTTP meant that essentially if you needed state, you had to do your own buggy approximation of a stateful connection. (E.g., the various "HTTP session" implementation in various languages or frameworks.)
The problem however, is that I've yet to see a project which didn't have session-related problems. Ranging from mild usability problems (e.g., you lost all your data when your session timed out), to waste of resources (someone stored an object in the session, which in turn pointed to a whole 100 MB tree), to identity theft (e.g., when enabling one to revive an older session id), all the way to such catastrophes as having users marked "new product" and products marked "new user" when someone used two windows at the same time.
And a lot of other stuff was literally based on leaps of faith. Like having faith that the user will never open a link in a second window, in the catastrophe implementation mentioned above. Or having faith that when you've opened a second window for your dialog, the user won't do other stuff with the previous window in the meantime. Or various other versions of assuming that the user will always mindlessly stick to the exact workflow you had in mind.
Of course, you think that such problems could and will never happen to your project. Think again. They even appear in such high profile projects as IBM WebSphere's admin console.
So IMHO if spammers and ad provider idiots manage to finally kill the idea that HTTP is _the_ only medium for a program to run over, good riddance.
I don't think many people have anything against advertising as a whole. But, yes, being obnoxious and annoying to your potential buyers is what's the real problem here.
As I've said it before, in the early days of the web the sites would have maybe one banner on the main site, and that was it. No animations, no huge flash movies, no pop-ups and no pop-unders. And most definitely no trying to install spyware crap on your computer.
And noone wanted to block those. Heck, some of us even clicked on them. In fact, I'll say that if back then someone would have come and said that they made an ad blocker, most people would have just said "you're lame, and you have too much time on your hands."
But in the meantime things have changed. Sorry, no matter how I much want to turn it around in my haead, and to find an excuse, I can't find any valid excuse for today's sorry state.
Now I'm bombarded with a barrage of pop-up, pop-unders, 500k flash movies, and even downright redirects to some vendor's site, before I even see what the site has to offer. Do I even want that site's "free services"? No idea, I'll first have to close all these windows and say "no" to their generous offer to install a dialer on my computer.
And in the end, I'll be mean and say that 99% of those sites didn't offer anything I wanted anyway. The mentality seems to be that any retard who can't even string together coherent sentences, has a right to make an income with ads. Which I'll disaggree.
Doing fast Google searches is one thing, but you may notice how Google doesn't need lame pop-unders to stay in business. You may also notice how Google not only had a service to offer, but also a _business_ _model_ for it. (Yes, using that search engine to offer targetted relevant text ads.)
And that's one site that deserves to stay in business. That's what capitalism is about.
On the other hand, let's look at the sites that do need obnoxious ads to stay in business. Yes, the millions of whiny blogs, or "look ma, I too can copy news off slashdot" news sites. They seem to assume that just having a stupid web site makes them so invaluable, that they have a _right_ to annoy me for a buck. Sorry, no. It doesn't work that way.