The short answer is "I'm not". No one party fits me well. So I look for the candidate whose stated goals match mine, or at least is the lesser of evils compared to the rest of the bunch.
In 2000 I would have happily supported Colin Powell. Unfortunately, he didn't run. In 2004, it was a vote against more than a vote for. I've come to the belief that the more fractured government is, the better off for us. Splitting the legislature and the executive so that no one party holds sway seems the best course from my perspective, and I try to vote accordingly.
PE's are next to useless for many industries. Where they aren't, they bestow legal liability and responsibility for things you sign. Because you can sign construction drawings, for instance, you become legally liable should it fail due to design issues. Because of this liability, however, you also increase your pay, and your insurance.
You missed my point entirely. Yes, we can accomplish feats of engineering the Romans could only dream about (lack of materials and computer models are merely two reasons) But will our feats last 2000 years and be serviceable? I doubt it.
The ability of an individual to recall and apply engineering principals on an examination only proves that an individual can take an examination. Even 8 hours is no where near enough time to do anything in depth. I regularly had single problems for homework that took more than 8 hours each to solve, and that would only prove I knew how to solve a single problem. Covering multiple topics on an 8 hour exam is proving nothing more than literate knowledge of the topic. That's why PEs require 5 years of experience and a sponsor.
You are correct. The original Architect's drawing was almost impossible to build. (Architects don't get much respect from engineers either, at least where I went to school. They're viewed more as artists than engineers, with all the negatives that entails. Architects frequently support those views by proposing impossible to build designs that merely look pretty).
The rod was supposed to thread through 2 walkways in one piece, so each walkway's brackets would only support 1 walkway's weight. The contractor split the rod, causing the upper walkway's bracket to take on the full load of both walkways. Even so, it might have held due to factors of safety usually being quite conservative in construction, but with the opening ceremony, the walkways were heavily crowded with viewers watching the ceremony below, and the brackets on the upper walkway gave way dropping both walkways to the ground.
And therein lies another reason to not get a PE. They're your drawings, and you're liable, even if someone changes them to be realistically able to build them. (A junior engineer I believe approved the change btw, or did the change, it's been way too long ago) Either way, this incident finished more than one career.
You know - the "family first, anti-drugs, small government" republicans? That same party where both the president and VP have DWIs? Where every single candidate in the 2008 race has been divorced at least once? The party that over the past 6 years has increased the size of the government and budget to the largest ever?
Disclaimer: I was a Republican. The above facts are just some of the many reasons I no longer am. The hypocrisy of that party boggles the mind.
PE's used to be very important. It was the only way engineers could make serious money instead of merely working for a PE in a position very much like a journeyman.
This all started changing when the majority of jobs available to engineers started being for large corporations, where PEs became all but irrelevant. We're not talking about CEs here, as they by and large still carry PEs if they're working in the construction industry.
Perhaps the animosity against PEs is because the GP has worked for one of the anointed gifted morons with a PE? I have rarely seen someone as arrogant as the PE a couple of my friends worked under. That guy convinced me to go look elsewhere after a mere 5 minutes of "discussion" (him basically stating how great he was and the great things he did). There were also the pedantic spewings of a couple of professors I had, one of which I believe lost his professional practice due to the Hyatt walkway collapse in Chicago. At least the level of intense animosity he displayed when discussing the particulars of that event and why there'd be no partial credit ever in his class certainly led me to believe that's why he returned to academia. He certainly didn't seem to enjoy teaching or talking with students.
Macs are a Lazy man's Linux. Its true, I am lazy and want XCODE and my other web IDEs. I fight my Linux system a bit too much and settled on the Mac. The hallmark of a good engineer, software or otherwise... I haven't really gotten into XCode yet, but it looks suitable for a dev IDE. I still run Eclipse for most everything.
C# and visual studios + SQL Server pretty Incredible. But to put MS in its place, they are better for Intranets, but can't touch web dev. And notice that AAPL is embracing their developer community. Just got a mac pro quad as a dev box. I personally haven't touched C# as a production project yet. From several friends and my admittedly limited experience, I don't think I want to either, at least not until it's gone through another revision or two and adds some basic enterprise level features.
MS SQL is unsuitable for anything of medium load production quality. Why? Because of escalating locks and the overall poor performance per CPU. I happen to know a company that couldn't be more MS oriented that's converting to a non MS DB because their MS SQL DB, partitioned to the point of diminishing returns, can no longer be scaled (they've got 800+ DB servers).
MS just never got big business. They developed a small almost toy-like OS that they leveraged (illegally) into a monopoly, and they've been struggling ever since with making portions of it scale. You can go back to the first versions of NT for the painfully obvious shortcomings of vision. The maximum effective networked set of systems on early NT was 1 network segment (however you masked it, provided you had good network equipment). Routers caused problems with the NT network stack in communicating with clients. Each time they fixed something, they broke something else. They added WINS to get around the router issue, then ran into scalability issues because each client pings the PDC every 5 or 10 seconds or something silly like that. It was also impossible to create a multi-geographic site with WAN links in a single domain when you scaled up to just 5K users that were split among only 4 major sites. (These were big WAN pipes btw, they weren't the problem, the latency and increasing delays of PDC response is what killed it) They "fixed" this by adding ADS. They effectively broke Exchange when they did, because try sending a 250K user mailing list with ADS. (It takes upwards of 3 hours to process just 1K users on the last system I benchmarked, FYI, and it took less than 1 min before ADS).
As for Web Dev, they've just never understood it either, because it contradicts their "prime directive". Web apps almost by definition remove OS lockin, and that's something that's anathema to MS.
well, I happen to have the full set of MSDN disks for 96-99. I'm sure that as I don't have an Alpha machine (I'm a little green;) that I could part with those disks if you can't find them elsewhere. Actually, I'm positive I can part with them, as I was planning to bin 13 them.
Glad you liked. And oh boy, you don't want to know what I do on ripping music. But here goes.
I use EAC (Exact Audio Copy). First thing you should know - it allows for configuring threads, and 2 threads is the bare minimum - it allows for ripping a song and encoding a song simultaneously. Since I have a dual-core system, I use 3 threads. I rip and encode a CD in about 4 minutes.
FLAC is the apparent lossless compression scheme of choice. I used to use MonkeyAudio, but not lately. I just store the uncompressed WAVs on a couple of 300GB drives. I also compress to MP3 using Lame. I can bring them into iTunes by mounting the folders and AAC compress them that way. Note that I have a combination of Windows PCs and Macs, and haven't fully decided how to convert to Mac only in this case.
I don't know about album art on iTunes, other than you can set a setting to retrieve album art even post rip. I didn't really care, so never investigated it. (I know - shock of shockers, I really only care about the music;)
Still haven't had a chance to look through my collection yet, so no other bands come forward at this time.
Since the topic was how case-insensitivity was hurting Java, the question is why isn't the Java runtime passing that flag to get the case-sensitivity it wants? Actually, it does. The problem comes into focus whenever you start using any windows based tools that do not necessarily honor it, or their windows variants don't honor it.
Windows NT had some awful scaling problems as recently as NT4, when it used to struggle to scale past 4 CPUs executing heavy IO. Linux used the global-kernel-lock arch in 2.0, and wasn't preemptible until 2.6. I'm just saying that last I checked, OSX isn't ready for >4 heavy SMP yet.
For comparison, it would be hard to test OSX with a lot of CPUs because the most you can get with Apple hardware is what, 4? Why doesn't Apple have a high-end server model with a lot more? You can get a Mac Pro, and drop 2 Kentfields in it. 8 CPUs. As for servers, that's a different story.
BTW, when's the last time you saw more than 4 cores on an NT based system? Seriously? Any time I've seen a need for over 4 cores, it's Sun first, and then IBM, depending upon SLAs and the like.
Note that none of this really addresses your comment about OSX. I agree that it's not ready for server functions in that range. Hopefully they'll fix a lot of those issues soon, like with the Leopard release.
Copy the disk!:) Or, better yet, ask some friends. I'm sure I'm not the only one that still has his 3.51, 4.0 and Win2K disks in a set of sleeves. Find someone with a packrat penchant and a long MSDN subscription.:) Although unless they have the MSDN disks for around 1998/9, you'll most likely be installing SP6, I think it was. I'd have to check mine, I just fished them out of the closet last weekend.
Feeding the troll?
I guess I must be imagining that Azureus is running on my Windows box as we speak. Hmm. You are aware that the bulk of Java coders and programs are enterprise? You know, those things that run businesses and bring in money?
I've never had Windows fail to come out of hibernation or sleep (with the exception of once when the battery died halfway through the hibernation process). Besides, in the event you do need to boot, doesn't Windows boot faster than Mac/Linux? Obviously you're blessed and the Dalai Lama should bow down to you. You've never had it fail except for once because you've only done hibernate/sleep once. The reboot time is about the same on my pared down (tweaked) desktop vs my laptop, I think. I've only rebooted the laptop once in the past 3 months for the 10.4.9 update, so I'm a little hazy on how long a non system update reboot takes.
As opposed to user friendly MS applications that do what they want? Ever hear of Visual Studio, you know, that program that anyone who develops for Windows uses? The IDE that everyone else emulates? I wouldn't know where to begin on this one. Let's start with MS apps. They only think they want those because, like people in the amazon who'd love a Yugo, they just haven't seen anything better.
Now to Visual Studio. If there ever were a bigger POS out there, I don't know what it is. Notepad, maybe? It's the most rudimentary of IDEs in existence, falls woefully short of real productivity enhancing features, and most horridly produces lots of useless boilerplate code that you wind up having to rewrite anyways. The GUI portion works, but that's about it. Since I don't write thick client, that aspect is useless to me. And god help you if you need to do heavy refactoring of code with VS. (BTW, I own 2 copies of VS, although not the latest. A friend I trust owns that one and stated its usability had not improved).
Maybe should invest in a laptop that isn't a piece of crap. This argument has been outdated since Win2000 came out; I thought you guys had finally given up trying to get people to keep believing it. Is around a $5K top end laptop with Windows XP Pro good enough for you? I'll get the exact model from another friend of mine who's about to buy a MBP.
Another draw is multi-media editing software. The software on OSX just works better and easier than anything I've seen on an MS system.
And another argument from 1998. You really need to pull your head out. Even software like Photoshop just plain works better on a Mac. (Especially now that CS3 is out) You should also look at the other packages available. Heck, the out of the box Mac comes with better multimedia editing software than most people buy for their PCs. I'm not even sure how expensive I'd have to go to match a PC package with the Mac packages I've got. I know the $100+ packages won't do.
I love how Apple fanboys always seem to cover up that Apple participates in more vendor lock-in that Microsoft, and instead explain that you can just "hack it". Oh, and breaking the licensing agreement and giving Apple the option of revoking your software license is always a great idea too. Do check Vista's licensing agreement and get back to me on that. Especially the OEM ones.
Let's see, Intel processor, Intel chipset... Probably about the same as a 16 or 32 core PC! I think you missed the sly aside there. I was indirectly referencing AMD. But that's no surprise. You appear to be living in a constant state of self-righteous denial, so don't let me disuade you from joining your PAL Ballme on stage.
I can only say I'm happy I'm not paying your salary. C++ sucks for web development. Way too slow to develop and costly to maintain. Pearl[sic] sucks in the maintenance arena. It's a great little scripting language and for single use cases, but for rapidly changing web sites in production - been there, done that, converted to Java and had a good night's sleep..NET just can't support the load. Try doing 35K concurrent pricing requests with yield management sometime. By the time you get that running, you'll be sucking a block's worth of power. I suppose next you'll tell me that MS SQL is the best DB server ever? (This is a huge joke in case you're too dense to catch it. MS SQL has so many issues no self-respecting architect would allow it in their systems) BTW, those 35K users are served off of 6 physical machines in the app server layer (where your C# code would run) and a single DB machine. There's more to the system, but that's all I'll say.
You sound like a junior programmer. You haven't yet learned enough about large systems to talk realistically about them, and it shows. Just ask anyone who's been supporting a large dynamic web site w/ Perl. It's been done, I know one group, and they don't sleep much. They're also about 8 revisions behind the Java version that went live 3 years ago. For some reason they haven't rolled it over yet.
As for Macs, they seem to be selling just fine, shipping almost 5% of the worldwide laptop market last year. That's 5% of the units, think about what that means in $'s.
Egads - you must have used NT 4.0 post SP1. The initial release was guaranteed to BSOD if you kept it up and running, and didn't run on laptops worth a crap. Why was it guaranteed to BSOD you ask? Because MS, in its godlike power, didn't think a thing about synchronizing a 20 bit kernel memory page counter and a 32 bit user space page counter... (when the kernel's counter rolled over, it was merely a matter of time.)
The rest of windows systems don't hibernate/sleep for shit. Win2K flip a coin. Half the time it BSOD'd on a roughly $5k at the time laptop. WinXP? Unless I strip it down it's unstable as hell. Then again, the frequent windows update forced reboots probably kick the system enough that you'll almost never run into a system issue caused by long living processes. We won't even talk about non-NT based systems.
You should get the native Windows compiled Unix tools. It really helps when you're popping back and forth. ls works. Oh, and cd \ works just fine too, taking you to the root of the drive you're on. If you configure your disks so that they mount in directories, you'll only have a single drive letter and no issues utilizing \ for root.:)
I also vaguely recall something somewhere about how to get the command line to recognize / for \. Why a visionary like Bill couldn't make this configurable is beyond me, since '/' is reserved as far as filenames go. But wait, Bill's like the l337 h4x0rz! It'll only take him 10000010000010001 days to figure it out.
NTFS is perfectly case sensitive. The Win32 interface to it generally isn't, but can be if you ask for it. AFAICT, Java is the one that decides to use filesystems in a case-insensitive manner, because that's what it asks for when it calls functions like CreateFile. Interesting then, that nothing deals with case sensitivity and that "MyFile", "myfile", and "MYFILE" all resolve to the same file. Try it sometime. I vaguely remember there's some esoteric thing you can set, but you'll break 99% of all windows programs if you use it. (Something about the POSIX subsystem dropped since XP came out, IIRC.) BTW, I'm talking absolute edge case software like Notepad (ships with the OS!) and Word.
Windows has a good BSD style "*nix subsystem" too. Yep, that's one bang up set of tools. Not even SSH. Telnet. Yep, keeping up with the times. Real UNIX integration there.
OK, so that was perhaps a bit harsh. But who runs this? I've never seen it in the wild. I've seen Cygwin everywhere, which is a much richer *nix shell environment, but I don't like it. (BTW, this is how I found out how shitty Windows file system case support is. I had to rename the files in Cygwin to see them. I believe it defaulted to all uppercase first.)
It's too bad that OSX has such kernel scaling problems, what with very coarse locking (somewhat improved in Tiger) and the necessity to use slow BSD user threads (as opposed to Mach kernel threads). Those things are going to need to be fixed before 16 or 32 cores are worthwhile, and I hope they don't have to break too much compatibility to do it. It'd be interesting to see what those numbers are on Intel cores. That paper is almost 2 years old. I'm thinking it may have improved some. It is true that the coarse locking was ameliorated in Tiger but not removed. Either way, that article is virtually irrelevant to today's systems other than to point out the pThread issue, and that MySQL does/did have issues on OSX. It's fine for my dev usage, but I've never heavily loaded it, so can't say I've ever run into this particular set of bottlenecks.
As great as Parallels is, I don't think I'd want to run OSX in a VM on Linux or Windows. I don't even have a real reason to run Parallels yet, it's just cool to be able to run XP in a window.:) Most of my current software needs are met by OSX software, and those that aren't do not currently run in a VM.
I had the same stance. But those features didn't matter as much for a laptop. I bought a powerbook 2 years ago, and upgraded to a MBP this year and haven't regretted it for a moment. I'm even considering a mini.
Both you and the GP have it wrong. Neither of those matter. How many of the machines out there play games vs how many are used just to browse the web, word process, and play with those nifty digital pictures all those cameras pump out?
But let's put aside the consumer market for the moment. There's something much more important afoot. Remember Ballmer's little prancing mantra? "Developers. Developers! DEVELOPERS!!!"? Guess what? There's a ton of Java developers out there. Their code does not run on MS OSes in general in production. Their tools are generally OS agnostic as well. In fact, in general, their tools run better on non-MS OSes. (Something about case-sensitive file systems)
Enter a nice, lightweight, reliable laptop (MBP) which about half the Java programmers I know have moved to over the last 6 months. Everyone that has loves them. There's the additional advantage that it's a *nix subsystem, which happens to mesh nicely with our targeted deploy environments. Add to that the hugely user friendly user features, and a bunch of us have come to realize that with Macs we get work done, we're not working on our systems. We don't have to do maintenance, configure them just so, or wait for them them to boot forever when they've failed to recover from a sleep/hibernate situation.
Now let's tie this back to Ballmer's rant, even though I'm discussing Java developers here. The developers I'm talking about are your top end developers, the ones who have to design and document architectures and give presentations. They start using user friendly non-MS applications that do what they want. They don't suffer embarassing BSODs on waking from sleep, or during their presentations. Lastly, they're also generally multiple language programmers, including C/C++ (and now add Objective C to their arsenal) and all of a sudden, there's more developers for Apple than MS.
Another draw is multi-media editing software. The software on OSX just works better and easier than anything I've seen on an MS system. Even software that covers both systems, Photoshop and Capture One, the Mac version either runs better or there is an equivalent Mac version that's just plain better.
As for making your own system with your own hardware, Apple sells their software already and if you're willing to hack it, you can run it on other hardware. It's just unsupported, which reduces Apple's liability. They're very successful at what they do, and it should be interesting to see where they go. I personally hope to see them use the best hardware and improve their threading issues in the OS in the future. I'd love to see a 16 or 32 core Mac Pro in the near future - imagine the processing ability of such a system.:)
Yes, I'm saying it doesn't. If you'll read the rest of the thread and your linked article, you'll note that the conclusion of anandtech's articles states:
When only running one or two CPU intensive threads, Quad FX ends up being slower than an identically clocked dual core system, and when running more threads it's no faster than Intel's Core 2 Extreme QX6700. But it's more expensive than the alternatives and consumes as much power as both, combined. And it's no coincidence that your conclusions and page link both refer to single threaded performance. And even there the difference isn't all that great. (BTW, that'd be 1 $500 CPU vs 1 $500 CPU, the second CPU is superfluous for single threaded gaming. Heck, you can make that 1 $250 Opteron 152 core and it will probably come as close.)
As for multitasking, only the power consumption is in question.
But the point I was making wasn't whether it was the most power conservative solution, but that 3+ year old tech can equal brand new tech from Intel.
So go ahead and pick nits all you want. Considering that AMD should get at least 10% improvement from the move to 65nm alone (increased clock speeds @ same or lower price) plus lower TDPs, and not counting on any of the myriad of other improvements they're adding to K8L, the fact that Intel can't trounce AMD's current old tech across the board would worry me if I were Intel.
Considering the laptop is brand new tech, and AMD is 3+ year old tech, I don't really consider it much of a win. My laptop, btw, is hooked up with RAID0 FW 800 drives. That certainly helps with performance tests that test more than the CPU/RAM, especially considering that the desktop has no RAID drives.
Intel's FSB has been the major bottleneck for years now. They're running a Bugotti on a potholed dirt road with a small smooth section in the middle. Once they dump the FSB which sounds like it may happen with Nehalma, they may be back in the running with multi-processor machines. Why may? Because the FSB is only one major bottleneck in multiprocessor machines. NUMA is another approach to better utilize resources, especially in the world of virtualized machines. AMD has both today, which is why they dominate the server space.
And speaking of innovations, AMDs 64bit extensions, built in memory controller, NUMA support, and hardware virtualization weren't worthy? Yes, Intel "innovated" the superscalar deep pipeline (P4) and the disaster that is the Itanium (could they have picked a more appropriate name?) but bad designs generally don't make for good longevity in the business world and I'm not sure they'll be listed as innovations, but more like "mistakes".
Lastly, AMD already licenses the HTT to others, so it is conceivable that someone could build a co-processor and plug it into a multi processor board and work with an AMD CPU. Pluggable cores implies multi-processor boards, and currently Intel isn't even in the parking lot. I think it funny that Intel announces a GPU core on chip after everyone already stated that was a driving reason for AMD to buy ATI. That goes right back to innovations. I think Intel has lost their way. They were a progressive company at one time, only because they had to be because of competition. They thought they'd "won", and sat on their laurels. AMD acquired a bunch of really smart folk and IP, and now their kicking their butt technologically, Core 2 not withstanding. The impression I get is that they're truly struggling to stay in the ball-game. That's based on previous AMD iterations, so we'll have to wait and see how far the K8L leaps, and if it's enough to surpass the C2D in single threaded performance. I have no doubt it's going to eat their lunch in 3+ threaded applications, and in 2+P systems they'll be untouchable for years yet.
Consumer Reports and others have not ranked Dyson anything more than average. I used to have other links, but that was 2 years ago when I was in the market for a new vacuum cleaner. I needed one that would clean dog hair out of the carpet without jamming the power head (3 chows produce lots of hair). The top rated vacuum for pet hair was the $500 Kenmore model. I bought it and I've been happy with everything but robustness of the powerhead. Fortunately, the 5 year extended warranty has already proven worth the $60.:)
In the reviews I read at the time, Dyson was at or near the bottom of every review set. Rainbows above $1200 were also considered, so it wasn't price that was a driver here. That link above lists someone saying something along the lines of "OMG - look how much crap came out of my carpet!" and somehow psychologically that states that the clear cylinder Dyson "cleans" the carpet better than, say, a bagged vacuum. Rainbow owners I know said the same thing, since you get to see the dirt every time you vacuum as you have to dump the water out. I wonder if these folks have ever looked into the vacuum bag after a single vacuuming? During shedding season, that's about how long a bag lasts.:-/
In any case, the reviews that convinced me Dyson stunk were the ones that measured exhaust particulates. Dyson came in dead last in the class of vacuums claiming to have hepa filters etc. I was also turned off by having to dump the dirt in a messy operation into the trash. Same issue with the Rainbows, although they performed near the top for particulates. Something about sucking air through water really filters out a lot of airborne crap.
Slashdot Mirror
The short answer is "I'm not". No one party fits me well. So I look for the candidate whose stated goals match mine, or at least is the lesser of evils compared to the rest of the bunch.
In 2000 I would have happily supported Colin Powell. Unfortunately, he didn't run. In 2004, it was a vote against more than a vote for. I've come to the belief that the more fractured government is, the better off for us. Splitting the legislature and the executive so that no one party holds sway seems the best course from my perspective, and I try to vote accordingly.
PE's are next to useless for many industries. Where they aren't, they bestow legal liability and responsibility for things you sign. Because you can sign construction drawings, for instance, you become legally liable should it fail due to design issues. Because of this liability, however, you also increase your pay, and your insurance.
You missed my point entirely. Yes, we can accomplish feats of engineering the Romans could only dream about (lack of materials and computer models are merely two reasons) But will our feats last 2000 years and be serviceable? I doubt it.
The ability of an individual to recall and apply engineering principals on an examination only proves that an individual can take an examination. Even 8 hours is no where near enough time to do anything in depth. I regularly had single problems for homework that took more than 8 hours each to solve, and that would only prove I knew how to solve a single problem. Covering multiple topics on an 8 hour exam is proving nothing more than literate knowledge of the topic. That's why PEs require 5 years of experience and a sponsor.
You are correct. The original Architect's drawing was almost impossible to build. (Architects don't get much respect from engineers either, at least where I went to school. They're viewed more as artists than engineers, with all the negatives that entails. Architects frequently support those views by proposing impossible to build designs that merely look pretty).
The rod was supposed to thread through 2 walkways in one piece, so each walkway's brackets would only support 1 walkway's weight. The contractor split the rod, causing the upper walkway's bracket to take on the full load of both walkways. Even so, it might have held due to factors of safety usually being quite conservative in construction, but with the opening ceremony, the walkways were heavily crowded with viewers watching the ceremony below, and the brackets on the upper walkway gave way dropping both walkways to the ground.
And therein lies another reason to not get a PE. They're your drawings, and you're liable, even if someone changes them to be realistically able to build them. (A junior engineer I believe approved the change btw, or did the change, it's been way too long ago) Either way, this incident finished more than one career.
Republicans were in power.
You know - the "family first, anti-drugs, small government" republicans? That same party where both the president and VP have DWIs? Where every single candidate in the 2008 race has been divorced at least once? The party that over the past 6 years has increased the size of the government and budget to the largest ever?
Disclaimer: I was a Republican. The above facts are just some of the many reasons I no longer am. The hypocrisy of that party boggles the mind.
Yep, those Romans and their PEs. I sure am glad they had them. 2000 years and still serviceable!
PE's used to be very important. It was the only way engineers could make serious money instead of merely working for a PE in a position very much like a journeyman.
This all started changing when the majority of jobs available to engineers started being for large corporations, where PEs became all but irrelevant. We're not talking about CEs here, as they by and large still carry PEs if they're working in the construction industry.
Perhaps the animosity against PEs is because the GP has worked for one of the anointed gifted morons with a PE? I have rarely seen someone as arrogant as the PE a couple of my friends worked under. That guy convinced me to go look elsewhere after a mere 5 minutes of "discussion" (him basically stating how great he was and the great things he did). There were also the pedantic spewings of a couple of professors I had, one of which I believe lost his professional practice due to the Hyatt walkway collapse in Chicago. At least the level of intense animosity he displayed when discussing the particulars of that event and why there'd be no partial credit ever in his class certainly led me to believe that's why he returned to academia. He certainly didn't seem to enjoy teaching or talking with students.
And notice that AAPL is embracing their developer community.
Just got a mac pro quad as a dev box. I personally haven't touched C# as a production project yet. From several friends and my admittedly limited experience, I don't think I want to either, at least not until it's gone through another revision or two and adds some basic enterprise level features.
MS SQL is unsuitable for anything of medium load production quality. Why? Because of escalating locks and the overall poor performance per CPU. I happen to know a company that couldn't be more MS oriented that's converting to a non MS DB because their MS SQL DB, partitioned to the point of diminishing returns, can no longer be scaled (they've got 800+ DB servers).
MS just never got big business. They developed a small almost toy-like OS that they leveraged (illegally) into a monopoly, and they've been struggling ever since with making portions of it scale. You can go back to the first versions of NT for the painfully obvious shortcomings of vision. The maximum effective networked set of systems on early NT was 1 network segment (however you masked it, provided you had good network equipment). Routers caused problems with the NT network stack in communicating with clients. Each time they fixed something, they broke something else. They added WINS to get around the router issue, then ran into scalability issues because each client pings the PDC every 5 or 10 seconds or something silly like that. It was also impossible to create a multi-geographic site with WAN links in a single domain when you scaled up to just 5K users that were split among only 4 major sites. (These were big WAN pipes btw, they weren't the problem, the latency and increasing delays of PDC response is what killed it) They "fixed" this by adding ADS. They effectively broke Exchange when they did, because try sending a 250K user mailing list with ADS. (It takes upwards of 3 hours to process just 1K users on the last system I benchmarked, FYI, and it took less than 1 min before ADS).
As for Web Dev, they've just never understood it either, because it contradicts their "prime directive". Web apps almost by definition remove OS lockin, and that's something that's anathema to MS.
Wow, that chartreuse glow is blinding me.
well, I happen to have the full set of MSDN disks for 96-99. I'm sure that as I don't have an Alpha machine (I'm a little green;) that I could part with those disks if you can't find them elsewhere. Actually, I'm positive I can part with them, as I was planning to bin 13 them.
Glad you liked. And oh boy, you don't want to know what I do on ripping music. But here goes.
;)
I use EAC (Exact Audio Copy). First thing you should know - it allows for configuring threads, and 2 threads is the bare minimum - it allows for ripping a song and encoding a song simultaneously. Since I have a dual-core system, I use 3 threads. I rip and encode a CD in about 4 minutes.
FLAC is the apparent lossless compression scheme of choice. I used to use MonkeyAudio, but not lately. I just store the uncompressed WAVs on a couple of 300GB drives. I also compress to MP3 using Lame. I can bring them into iTunes by mounting the folders and AAC compress them that way. Note that I have a combination of Windows PCs and Macs, and haven't fully decided how to convert to Mac only in this case.
I don't know about album art on iTunes, other than you can set a setting to retrieve album art even post rip. I didn't really care, so never investigated it. (I know - shock of shockers, I really only care about the music
Still haven't had a chance to look through my collection yet, so no other bands come forward at this time.
For comparison, it would be hard to test OSX with a lot of CPUs because the most you can get with Apple hardware is what, 4? Why doesn't Apple have a high-end server model with a lot more? You can get a Mac Pro, and drop 2 Kentfields in it. 8 CPUs. As for servers, that's a different story.
BTW, when's the last time you saw more than 4 cores on an NT based system? Seriously? Any time I've seen a need for over 4 cores, it's Sun first, and then IBM, depending upon SLAs and the like.
Note that none of this really addresses your comment about OSX. I agree that it's not ready for server functions in that range. Hopefully they'll fix a lot of those issues soon, like with the Leopard release.
Copy the disk! :) Or, better yet, ask some friends. I'm sure I'm not the only one that still has his 3.51, 4.0 and Win2K disks in a set of sleeves. Find someone with a packrat penchant and a long MSDN subscription. :) Although unless they have the MSDN disks for around 1998/9, you'll most likely be installing SP6, I think it was. I'd have to check mine, I just fished them out of the closet last weekend.
Now to Visual Studio. If there ever were a bigger POS out there, I don't know what it is. Notepad, maybe? It's the most rudimentary of IDEs in existence, falls woefully short of real productivity enhancing features, and most horridly produces lots of useless boilerplate code that you wind up having to rewrite anyways. The GUI portion works, but that's about it. Since I don't write thick client, that aspect is useless to me. And god help you if you need to do heavy refactoring of code with VS. (BTW, I own 2 copies of VS, although not the latest. A friend I trust owns that one and stated its usability had not improved). Maybe should invest in a laptop that isn't a piece of crap. This argument has been outdated since Win2000 came out; I thought you guys had finally given up trying to get people to keep believing it. Is around a $5K top end laptop with Windows XP Pro good enough for you? I'll get the exact model from another friend of mine who's about to buy a MBP. Another draw is multi-media editing software. The software on OSX just works better and easier than anything I've seen on an MS system.
And another argument from 1998. You really need to pull your head out. Even software like Photoshop just plain works better on a Mac.
(Especially now that CS3 is out) You should also look at the other packages available. Heck, the out of the box Mac comes with better multimedia editing software than most people buy for their PCs. I'm not even sure how expensive I'd have to go to match a PC package with the Mac packages I've got. I know the $100+ packages won't do. I love how Apple fanboys always seem to cover up that Apple participates in more vendor lock-in that Microsoft, and instead explain that you can just "hack it". Oh, and breaking the licensing agreement and giving Apple the option of revoking your software license is always a great idea too. Do check Vista's licensing agreement and get back to me on that. Especially the OEM ones. Let's see, Intel processor, Intel chipset... Probably about the same as a 16 or 32 core PC! I think you missed the sly aside there. I was indirectly referencing AMD. But that's no surprise. You appear to be living in a constant state of self-righteous denial, so don't let me disuade you from joining your PAL Ballme on stage.
Yep, and I changed it when I got my system. It's a simple change, and better yet, all your applications respect it.
I can only say I'm happy I'm not paying your salary. C++ sucks for web development. Way too slow to develop and costly to maintain. Pearl[sic] sucks in the maintenance arena. It's a great little scripting language and for single use cases, but for rapidly changing web sites in production - been there, done that, converted to Java and had a good night's sleep. .NET just can't support the load. Try doing 35K concurrent pricing requests with yield management sometime. By the time you get that running, you'll be sucking a block's worth of power. I suppose next you'll tell me that MS SQL is the best DB server ever? (This is a huge joke in case you're too dense to catch it. MS SQL has so many issues no self-respecting architect would allow it in their systems) BTW, those 35K users are served off of 6 physical machines in the app server layer (where your C# code would run) and a single DB machine. There's more to the system, but that's all I'll say.
You sound like a junior programmer. You haven't yet learned enough about large systems to talk realistically about them, and it shows. Just ask anyone who's been supporting a large dynamic web site w/ Perl. It's been done, I know one group, and they don't sleep much. They're also about 8 revisions behind the Java version that went live 3 years ago. For some reason they haven't rolled it over yet.
As for Macs, they seem to be selling just fine, shipping almost 5% of the worldwide laptop market last year. That's 5% of the units, think about what that means in $'s.
Egads - you must have used NT 4.0 post SP1. The initial release was guaranteed to BSOD if you kept it up and running, and didn't run on laptops worth a crap. Why was it guaranteed to BSOD you ask? Because MS, in its godlike power, didn't think a thing about synchronizing a 20 bit kernel memory page counter and a 32 bit user space page counter... (when the kernel's counter rolled over, it was merely a matter of time.)
The rest of windows systems don't hibernate/sleep for shit. Win2K flip a coin. Half the time it BSOD'd on a roughly $5k at the time laptop. WinXP? Unless I strip it down it's unstable as hell. Then again, the frequent windows update forced reboots probably kick the system enough that you'll almost never run into a system issue caused by long living processes. We won't even talk about non-NT based systems.
You should get the native Windows compiled Unix tools. It really helps when you're popping back and forth. ls works. Oh, and cd \ works just fine too, taking you to the root of the drive you're on. If you configure your disks so that they mount in directories, you'll only have a single drive letter and no issues utilizing \ for root. :)
I also vaguely recall something somewhere about how to get the command line to recognize / for \. Why a visionary like Bill couldn't make this configurable is beyond me, since '/' is reserved as far as filenames go. But wait, Bill's like the l337 h4x0rz! It'll only take him 10000010000010001 days to figure it out.
OK, so that was perhaps a bit harsh. But who runs this? I've never seen it in the wild. I've seen Cygwin everywhere, which is a much richer *nix shell environment, but I don't like it. (BTW, this is how I found out how shitty Windows file system case support is. I had to rename the files in Cygwin to see them. I believe it defaulted to all uppercase first.) It's too bad that OSX has such kernel scaling problems, what with very coarse locking (somewhat improved in Tiger) and the necessity to use slow BSD user threads (as opposed to Mach kernel threads). Those things are going to need to be fixed before 16 or 32 cores are worthwhile, and I hope they don't have to break too much compatibility to do it. It'd be interesting to see what those numbers are on Intel cores. That paper is almost 2 years old. I'm thinking it may have improved some. It is true that the coarse locking was ameliorated in Tiger but not removed. Either way, that article is virtually irrelevant to today's systems other than to point out the pThread issue, and that MySQL does/did have issues on OSX. It's fine for my dev usage, but I've never heavily loaded it, so can't say I've ever run into this particular set of bottlenecks.
As great as Parallels is, I don't think I'd want to run OSX in a VM on Linux or Windows. I don't even have a real reason to run Parallels yet, it's just cool to be able to run XP in a window. :) Most of my current software needs are met by OSX software, and those that aren't do not currently run in a VM.
I had the same stance. But those features didn't matter as much for a laptop. I bought a powerbook 2 years ago, and upgraded to a MBP this year and haven't regretted it for a moment. I'm even considering a mini.
Both you and the GP have it wrong. Neither of those matter. How many of the machines out there play games vs how many are used just to browse the web, word process, and play with those nifty digital pictures all those cameras pump out?
:)
But let's put aside the consumer market for the moment. There's something much more important afoot. Remember Ballmer's little prancing mantra? "Developers. Developers! DEVELOPERS!!!"? Guess what? There's a ton of Java developers out there. Their code does not run on MS OSes in general in production. Their tools are generally OS agnostic as well. In fact, in general, their tools run better on non-MS OSes. (Something about case-sensitive file systems)
Enter a nice, lightweight, reliable laptop (MBP) which about half the Java programmers I know have moved to over the last 6 months. Everyone that has loves them. There's the additional advantage that it's a *nix subsystem, which happens to mesh nicely with our targeted deploy environments. Add to that the hugely user friendly user features, and a bunch of us have come to realize that with Macs we get work done, we're not working on our systems. We don't have to do maintenance, configure them just so, or wait for them them to boot forever when they've failed to recover from a sleep/hibernate situation.
Now let's tie this back to Ballmer's rant, even though I'm discussing Java developers here. The developers I'm talking about are your top end developers, the ones who have to design and document architectures and give presentations. They start using user friendly non-MS applications that do what they want. They don't suffer embarassing BSODs on waking from sleep, or during their presentations. Lastly, they're also generally multiple language programmers, including C/C++ (and now add Objective C to their arsenal) and all of a sudden, there's more developers for Apple than MS.
Another draw is multi-media editing software. The software on OSX just works better and easier than anything I've seen on an MS system. Even software that covers both systems, Photoshop and Capture One, the Mac version either runs better or there is an equivalent Mac version that's just plain better.
As for making your own system with your own hardware, Apple sells their software already and if you're willing to hack it, you can run it on other hardware. It's just unsupported, which reduces Apple's liability. They're very successful at what they do, and it should be interesting to see where they go. I personally hope to see them use the best hardware and improve their threading issues in the OS in the future. I'd love to see a 16 or 32 core Mac Pro in the near future - imagine the processing ability of such a system.
As for multitasking, only the power consumption is in question.
But the point I was making wasn't whether it was the most power conservative solution, but that 3+ year old tech can equal brand new tech from Intel.
So go ahead and pick nits all you want. Considering that AMD should get at least 10% improvement from the move to 65nm alone (increased clock speeds @ same or lower price) plus lower TDPs, and not counting on any of the myriad of other improvements they're adding to K8L, the fact that Intel can't trounce AMD's current old tech across the board would worry me if I were Intel.
Considering the laptop is brand new tech, and AMD is 3+ year old tech, I don't really consider it much of a win. My laptop, btw, is hooked up with RAID0 FW 800 drives. That certainly helps with performance tests that test more than the CPU/RAM, especially considering that the desktop has no RAID drives.
Intel's FSB has been the major bottleneck for years now. They're running a Bugotti on a potholed dirt road with a small smooth section in the middle. Once they dump the FSB which sounds like it may happen with Nehalma, they may be back in the running with multi-processor machines. Why may? Because the FSB is only one major bottleneck in multiprocessor machines. NUMA is another approach to better utilize resources, especially in the world of virtualized machines. AMD has both today, which is why they dominate the server space.
And speaking of innovations, AMDs 64bit extensions, built in memory controller, NUMA support, and hardware virtualization weren't worthy? Yes, Intel "innovated" the superscalar deep pipeline (P4) and the disaster that is the Itanium (could they have picked a more appropriate name?) but bad designs generally don't make for good longevity in the business world and I'm not sure they'll be listed as innovations, but more like "mistakes".
Lastly, AMD already licenses the HTT to others, so it is conceivable that someone could build a co-processor and plug it into a multi processor board and work with an AMD CPU. Pluggable cores implies multi-processor boards, and currently Intel isn't even in the parking lot. I think it funny that Intel announces a GPU core on chip after everyone already stated that was a driving reason for AMD to buy ATI. That goes right back to innovations. I think Intel has lost their way. They were a progressive company at one time, only because they had to be because of competition. They thought they'd "won", and sat on their laurels. AMD acquired a bunch of really smart folk and IP, and now their kicking their butt technologically, Core 2 not withstanding. The impression I get is that they're truly struggling to stay in the ball-game. That's based on previous AMD iterations, so we'll have to wait and see how far the K8L leaps, and if it's enough to surpass the C2D in single threaded performance. I have no doubt it's going to eat their lunch in 3+ threaded applications, and in 2+P systems they'll be untouchable for years yet.
Consumer Reports and others have not ranked Dyson anything more than average. I used to have other links, but that was 2 years ago when I was in the market for a new vacuum cleaner. I needed one that would clean dog hair out of the carpet without jamming the power head (3 chows produce lots of hair). The top rated vacuum for pet hair was the $500 Kenmore model. I bought it and I've been happy with everything but robustness of the powerhead. Fortunately, the 5 year extended warranty has already proven worth the $60. :)
:-/
In the reviews I read at the time, Dyson was at or near the bottom of every review set. Rainbows above $1200 were also considered, so it wasn't price that was a driver here. That link above lists someone saying something along the lines of "OMG - look how much crap came out of my carpet!" and somehow psychologically that states that the clear cylinder Dyson "cleans" the carpet better than, say, a bagged vacuum. Rainbow owners I know said the same thing, since you get to see the dirt every time you vacuum as you have to dump the water out. I wonder if these folks have ever looked into the vacuum bag after a single vacuuming? During shedding season, that's about how long a bag lasts.
In any case, the reviews that convinced me Dyson stunk were the ones that measured exhaust particulates. Dyson came in dead last in the class of vacuums claiming to have hepa filters etc. I was also turned off by having to dump the dirt in a messy operation into the trash. Same issue with the Rainbows, although they performed near the top for particulates. Something about sucking air through water really filters out a lot of airborne crap.