I saw a great comment on Slashdot once: The article was about some research being done at CERN, but it was a bit heavy on the jargon, and nobody could quite work out enough of the technical details to make really insightful comments. Lots of people made quipped along your lines: "it's necessarily complex", "jargon of the field", etc...
However, one of the comments was from a scientist working at CERN, on a very similar experiment. He didn't have a clue what the paper was about either, because the language was so cryptic, and so full of acronyms and jargon that overall it was just too hard to understand. If a guy in the same building, working on the same stuff can't understand what you're doing, how is it of value to the scientific community at large, many of whom don't even speak english as a first language?
Scientific papers in some fields have been suffering from a tendency to use language similar to patent legalese - it's obscured not to meet a specific purpose, but to try and make mundane discoveries sound complex. It's a type of showing off - the author is demonstrating that he's one of the 'elite'.
It hasn't spread to all sciences, yet. I studied Computing Science and Physics at university, in equal amounts. I can read and understand most Comp Sci research papers, but almost no modern Physics papers at all. Unfortunately, I see Comp Sci slowly getting worse. Source code is getting rare in papers, and the use of abstract maths using greek symbols is increasing. I suspect it'll get to the point pretty soon that a skilled programmer won't be able to make heads or tails of a Comp Sci paper, even if it's directly applicable to the problems encountered by a programmer day to day.
99% of all scientists in Galileo's day believed the earth was the center of the universe.
Actually... no. It was the 99% of the indoctrinated followers of the catholic church that believed that, despite mounting evidence to the contrary from an increasing number of secular scientists. Galileo was NOT the first to use a heliocentric model, he based his theories on the work of other scientists. Mind you, at the time, most astronomers were really astrologers, and hired by the church.
The parallel with the modern issue of climate change is uncanny. The minority of objective scientists are telling a truth the conservative leadership doesn't want to hear, and instead of entering into a reasoned debate, the conservatives are attempting futile smear campaigns.
Thank god we don't execute scientists any more for doubting the dogma of the leading elite, or climatology would be a dangerous profession right now!
Claims? As in no one else but the company has stated this "fact"? I wish this article waited for a review before being posted:S
It's not outside what I'd expect for a next-gen enterprise SSD. The PCI-E FusionIO cards can easily do 100K IOPS sustained. I'm just surprised the SAS bus can send that many commands per second. I guess the SCSI wire protocol scales better than I assumed.
The bandwidth numbers are actually relatively low - that's the bus speed limiting the drive. I suspect that pretty soon, most enterprise-grade SSDs will connect using the PCI bus to avoid that.
They basically admitted that they didn't have the technical strength to build something like that at their scale.
Wow.
They need to hire me, right now, for $3000 a day. Seriously, if they can't do that, I could work miracles for them, and I'd be worth every cent of that money. Scaling session IDs up is easy - in fact, it's a trivially parallelizable problem by nature.
Author misses lots of things, and makes all sorts of invalid comparisons.
For example, the invention of the electric light may seem like a big thing, but there were centrally powered lighting systems already when it was invented - such as town gas lighting. Sure, electric lights are better, but one could say it's just an 'incremental improvement'. It's just a matter of perspective.
And while the lightbulb was a big invention, it was largely unchanged for the first 50+ years. Almost every light bulb was a hot filament in a vacuum. More recently, we've been making entirely new sources of light, using entirely new chemical or physical principles.Think LEDs, OLEDs, all sorts of lasers, bioluminescence that we can now splice into rats and bunnies at will, etc... We've even made rather esoteric sources of light like beta-radiation powered lights that last for a decade.
The author also makes comments like this:
But despite daily announcements of one breakthrough or another, morbidity and mortality from cancer and stroke continue practically unabated, even in developed countries.
Well... duh. Something has got to kill us in our old age eventually, and it'll be the diseases that are hard to cure, obviously. Until we develop some sort of immortality, that's not going to change. 100% of people will die, of something, no matter how good medicine is.
Until we all become immortal, what about the major advancements, like the recent developments in growing organs? It's still in it's early stages, but even what we've got now is a massive leap forward in medicine, almost as big as the invention of modern surgical techniques.
To sum up: it's significantly more expensive, but since glass doesn't wear like asphalt does (it either works or breaks -- and it doesn't generally break from compressive stress, only torsional stress and impact), it should last longer and need less maintenance. And since you also get power out of it, displace plow crews, etc, they make the argument that it'll be a better investment if they can make the panels for $10k or less each.
Given that the one-off prototype is to cost $100k, and they have the potential for a *huge* amount of mass production, I don't think it's all that unrealistic. I'd still like to see how they handle in the real world, of course, but hey, that's why you give funding to build prototypes.;)
Oh... yes! The numbers! I love the wishful naive thinking on that page, it's just brilliant.
For example, lets examine one of the pieces of insanity on his site. He mentions embedding supercapacitors into the road surface to store energy (I assume overnight). If you don't know what those things are, they would be the filthy expensive, highly experimental, rarely used in commercial products devices with lower than battery storage capacity. I'm sure they'll improve, but I can come up with fancy plans too if I can have parts made of unobtanium.
I particularly like the plan to use the ultracaps to store sufficient power to melt ice off the roads. The inventor clearly doesn't remember his 1st year Physics, where we learnt that the the enthalpy of fusion of water is surprisingly high compared to most other chemicals.
Ok, lets get practical: I'm basing this off the technical specs (PDF) for one of the beefier ultracapacitors made by one of the top companies in the biz - Maxwell Technologies. (note: I'm sure better devices are available from somewhere else, will be soon, etc.. bear with me)
It states that a device that is about 17.6cm high and has an area of 18.9cm x 51.5cm has a total capacity of 55Wh (~200kJ). That's a big capacitor.
So if you made a road surface with it, every 973.35 cm^2 area would have 200kJ of stored power for it. That's about 200J per cm^2.
Since the enthalpy of fusion of water 333 J/g, then 200J of energy will melt 0.6g of water. A layer of water (or ice) 0.6g/cm^2 is 6mm deep.
To summarize, this guy's fancy 'invention', if 100% efficient could melt 6mm of ice (or something like 5cm of snow), assuming that the weak winter sunlight was sufficient to fully charge the capacitors during the previous day. That's assuming the entire road surface has a layer of supercapacitors in it 17.6cm thick (that's 7 inches for you yanks).
Even if you gave the benefit of doubt and assumed a 10x improvement in supercapacitor technology, you still have to factor in that he plans to use the solar power capacity for other things too, like lighting up the LED arrays built-in to the road, and to power nearby homes. Not to mention that no matter how much capacity you have, there's not enough sunlight to charge it.
Note that the cost estimates conveniently left out the cost of the ultracaps. On one of the pages, he mentions a target price of USD48 per square foot. The Maxwell ultracap is about 1 square foot, so we're looking at $48 split between a square foot of: Solar cells, the glass coating, an ultrapacitor 7 inches thick, high intensity LEDs, heating coils, power management electronics, the road substrate, and more.
Who was the moron who gave him $100K? Can I have my free money now too? I can come up with all sorts of wild plans also that make zero fiscal sense!
My bad, I just googled for a random picture of solar-panel things in the desert. 8)
The fact that it hasn't been built just strengthens my point. Event the 'optimal' solar panel sites are uncompetitive with traditional forms of power, or the cheaper forms of green power (especially wind). Throwing away a bunch of efficiency on top of that is just crazy.
...Also why do they have to make roads out of them.. where did that come from? Just put them out on land somewhere, you don't have to drive all over them.
This was my first thought too. Making the solar panels into roads (or vice versa) is compounding the problem. Just put the 25,000 mi^2 of solar panels in the middle of the desert and call it even. Adding a layer of glass or some sort of protective surface is going to lessen the efficiency and raise the cost of production and maintenance. I'm all about green energy, but there are better places we could be spending our money and energy.
Back at uni, I did a mini-course on the the Solar Car challenge, because my University made some of the solar panels for the top cars, and we also had a car that entered and did fairly well (for a low budget). One of the things we learned was that solar cells lose efficiency very quickly from a variety of things. The two that most researchers ignored in the lab but mattered in the field was heat and dirt. The cars in the race are washed with cold water thoroughly at every opportunity because colder, cleaner cells are substantially more efficient. Think CPU overclocking - lower temperatures improves things a lot.
Now lets compare this situation to a typical road which is: a) Blistering hot most days. b) Really, truly, thoroughly dirty.
Sounds like the perfect place to put an expensive solar cell panel!
Another thing we learned is that a single "test" panel in a lab operates very differently to a bunch of real panels in the field. What a lot of naive researchers miss is that the amount of sunlight over the entire collecting surface in the real-world is not constant. For a one-square-foot panel, it is, but for any significant surface (the size of a car, road, whatever), it won't be. The surface will be curved or partially shadowed. This matters a lot because if you just connect a bunch of cells together, they perform roughly the same as the worst of the lot. If there's a few cells under a shadow, that's drags down the efficiency of the panels receiving sunlight. To efficiently extract energy from a bunch of panels receiving differing amounts of light takes a bunch of expensive power management electronics that can combine the different cell outputs in the right way.
In practice, cells are so expensive that the best place to put them is on huge, flat, orientable panels out in the desert where there's no clouds, no rainfall to cake dirt onto the panels, and they can be oriented to face the sun at all time, like this array in southern California.
You failed to understand that re-purchasing a vast set of licenses for some 64-bit OS just for the sake of it being 64-bit when your entire portfolio of applications and hardware (you need drivers) was built for the 32-bit platform... Well, it doesn't make any sense. It means that you intended to migrate your office from a 32-bit OS to a 64-bit OS without any other justification beyond "I felt like it", which carries a hefty price tag and absolutely no technical justification.
And moreover, the claims that 32-bit apps "just run fine on 64-bit windows" are made from the same people who claimed that your Windows XP apps "just run fine on Windows Vista" and we saw how that panned out.
Why would you have to repurchase application licenses to upgrade Windows? I'm not aware of any applications that tie licensing to the OS in such a low level way. That's just bizarre. Most apps are licensed by seat, concurrent user, or 'enterprise agreement'. I know of some apps where you can't switch between Linux and Windows like that, but the only apps I've seen prevent an OS upgrade within the same family are backup and anti-virus apps, but most allow it. Also, since Vista, Windows licenses are interchangeable between 32-bit and 64-bit editions.
You do realize that just because an application was "built for 32-bit platforms" doesn't mean it'll magically perform terribly on a 64-bit machine, right? In fact, my experience has been the opposite: A 64-bit OS is noticeably faster for 32-bit apps, because the disk cache is much larger. There's a little-known limit in 32-bit Windows: it can't create a disk cache much larger than 450MB, no matter how much physical RAM you have, because of the way the kernel address space is laid out. If you run apps that exceed that limit, your disk gets thrashed even though you have free memory. A user running a couple of Office apps can easily exceed that limit.
Most people don't realize that even 32-bit server editions can't go over that ~400MB limit. It cracks me up to see 32-bit file servers with 8GB of memory. I've seen 10-fold reductions in disk IO rate by switching to 64-bit.
There may not be any technical reason to still hold on to a 32-bit OS but there sure are economical ones. As soon as you have an entire office filled with desktops, each one running tens of thousands of dollars in software that were released only to the 32-bit platform, then you will consider that there is a very good reason to still run a 32-bit system.
Wow, the Mods are asleep or not paying attention to the dozens and dozens of post which have pointed out that 32-bit apps run just fine on 64-bit windows.
XP x64 was awful. Don't use it. Vista x64 is way better than XP x64.
Windows 7 x64 seems to run quite smoothly. I haven't run into a 32 bit application (driver, yes; application, no) that I can't run. Granted, I don't run a ton, but several old games a few apps (like Sibelius).
Mod parent down, he has no idea what he's on about.
Windows XP 64-bit is exactly the same as Windows 2003 Server 64-bit. They share the same service pack files! The only difference is some minor licensing restrictions (can't install server components, 2 CPU sockets max, etc...). Drivers that work for either will work for both.
I had XP 64-bit on both my desktop at home and my work laptop for years, basically since it was released. I never had any issues with it, even on fairly unusual hardware. It is as light-weight as XP, but has 64-bit kernel caches and buffers, so it runs much better on the same hardware, and ran circles around both 32-bit and 64-bit Vista. Only now that Windows 7 is out have I upgraded my laptop to that, because Vista was just too painful to use in comparison to XP 64-bit.
Mind you, I wouldn't install anything other than Windows 7 64-bit on desktop PCs these days, but before that, XP 64-bit was a very nice OS.
I've used an unsigned driver in vista 64 bit and I've never seen this. I used EasyBCD and selected the option for allowing unsigned drivers in 64 bit windows and everything was fine. No test mode or anything.
I just did something similar on Windows 7 x64, and no "test mode" there either!
They can just park a ship off a Manhattan and light one off if they wanted to.
Ships move at 20-30mph. Ballistic missiles move at 15,000mph +. If we make it so that our enemies have to get a ship into one our harbors, it becomes a much simpler problem. We would need to have more Coast Guard people to basically board every ship, with neutron detectors, but, its something we can do. We can track ships as they are approaching the USA, track them as they leave ports, follow them, and pretty much monitor every boat on the ocean.
Why doesn't the US do that now?
Several analysts have pointed out that the easiest way to get a nuke into a major US city is by shipping it a standard shipping container. Many cities have container docks near the central business district, where even a small nuclear explosion from a low-tech device would inflict massive damage.
Containers can't even be properly scanned for radioactive materials when they're being unloaded in the harbor. The billion dollar project to install radiation detectors at US ports is apparently a failure - it detects false positives on things like shipments of granite, but fails to detect several varieties of Uranium and Plutonium, including the forms used in nukes. Not to mention that detecting a nuke at that stage is a bit late, it's already in the right place. And even if container scanning is perfected, there's dozens of other ways of smuggling a bomb in - cruise ships, private yachts, oil tankers, scuba divers, submarines, indirectly through Mexico or Canada, etc...
I read a great American Scientist article on why the current border protections against nuclear materials are a joke, but I can't find it now. I did find another research paper that covers most of the key points though.
Hmm... one thing that might be going on is that when you upgrade to an SSD, suddenly everything becomes CPU limited.
I noticed that before my SSD, the CPU would only rarely go above 30% for any length of time, but now it hits 100% regularly.
The Intel Core i7 in your desktop is at least 3x the speed of your laptop CPU: it has double the cores, and each core is more efficient.
I saw a similar thing recently: I have a Vertex drive in a 2.66 GHz C2D laptop (4GB, running Windows 2008 x64), and a co-worker put an identical drive into a quad-core desktop. The desktop system was quite noticeably faster.
We've all become used to CPU upgrades having little effect because of the bottleneck of the disk system, so when the disk IO issues go away, it can be surprising how much difference a decent CPU makes.
My strategy has always been the same: I liked to find the ultimate 'combo' and push my character far beyond what the original designers intended. For example, in WoW I played an Alchemist Paladin - the theory being that by using Paladin shields and spells, I could survive just long enough in high-level zones to pick some uber herb, make uber potions, and then use said potions to turn myself into the Ubermensch. I could pull it off too. I was able to go into zones 10 levels above me, and walk out with bags full of herbs. I did things like jump off cliffs on low-level zones into corners of high level zones while shielded, so I would be able to pick the one herb that was there out of range of any monsters. I carefully waited for patrols of elite mobs to wander past, and I'd carefully sneak past them into high-level zones.
My strategy would have worked, except that Blizzard put in totally artificial limits into the game, so that if you did manage to cleverly set your character up like that, it wouldn't actually do you any good at all. The best strategy (in terms of efficiency) by far was always to grind roughly equal level mobs while doing boring quests (kill X of Y mobs). That pissed me off. The game actively forced you to play in a boring, linear, repetitive manner. Creativity was not only discouraged, but in many cases virtually impossible.
I'm sure they did it to prevent 'twinking', or 'exploiting', or whatever, but it made the game deadly boring for me. I much prefer games like Nethack, where you can find a wand of wishing after playing the game for 3 turns, and still die four turns later. (Yes, it happened to me. I now know that I shouldn't wish for artifacts that blast the wielder with their power when picked up if I'm still at level 1.)
I put an OCZ 256GB Vertex into my laptop, and it got so fast it's just unbelievable. I did have to upgrade the firmware, the stock firmware was both buggy and slow. Since then, several major revisions of the firmware have been released. I'll try this new GC firmware when I upgrade to Windows 7.
Still, I have more confidence in the long term reliability of the Intel drives. A friend of mine purchased a Vertex drive as well, and it died a week after he got it, and mine had data corruption issues at first until I patched the firmware.
This is all easy to do. Why aren't you doing it? For a small office, it wouldn't even be expensive.
Especially in a small business, your users will rebel if they can't install (or use) their software... which is quite reasonable given most people are still running Windows XP, and most XP software is not capable of being installed or sometimes even used without admin access... this is especially troublesome if that user happens to be the CEO/Owner.
You hardly ever have time/resources to "do it properly" in a small business, unless what you're "doing right" is a core competency of the business. The trick is to convince the guy who signs the checks that it is business/mission critical (often non-trivial).
If you have full-time IT staff at the company, it's possible to convince users that they're better off, you just have to approach things from the right direction. For example, making users "non-admin" and "user-initiated software installation" are not mutually exclusive. Under Windows, you can publish a list of pre-prepared applications, and users can be allowed to install anything from that list at will simply by clicking an icon. It's much harder if you're not always on site, ready to install applications for users in an ad-hoc manner.
Also, before you remove user's admin rights, make sure that their new desktop image contains every app they have previously used. If they don't feel the need to install any additional software, they won't even notice that they can't. For example, you can go a long way by making sure that every workstation has all the basic 'frameworks' like Java,.NET, Flash, etc... that way, every website will work.
There are several use cases where a timely, user-intervention-free shutdown is critical to the correct functioning of an operating system. It's an operation that must always succeed, or the OS is broken.
Then you should be using the command outlined above: "shutdown/s/t 1/f"
This will shutdown the computer and close any force-close any open applications.
Hell, go through the PID list and KILL any applications you want, then issue a shutdown request.
That's a complex, factual, obscure, smug, and incorrect answer. Typical developer attitude.
I do know about that command, and have been forced to use it, but I avoid it if at all possible, because, as you said yourself, it force-closes applications. On a server, that can lead to data loss.
That's like saying that if I have trouble going to sleep, then I should keep a revolver next to my bed, because that's the only way to make sure that I'll go down, and stay down.
Is it so hard to respond to a "machine shutdown" event by serializing the application state to a temporary file, and then restoring it when the user runs the app again next boot?
It is very easy, it's called "Hibernate". All modern computer/OS combos can do it. Just don't expect any connection-dependent applications to work correctly after.
Shutdown is not the same as hibernate. You can't apply an OS patch, and then hibernate to apply it, for example. My original example of Firefox is a great example of how most apps should work. After the "apply patch and restart now?" prompt, Firefox returns to where it was before.
One of the reasons I hate Windows so much is that I can't even rely on the piece of crap OS to shut down if I tell it to shut down and then walk away. It literally expects me to sit there for up to five minutes while it 'saves my settings' and stops all the processes to ensure the bloody thing turns itself off.
Sounds like you hate computers in general and have likely chosen the wrong profession.
Why is his requirement unrealistic? I'm a computer professional, I love computers, and I fully agree with the original poster. Why should I wait and watch what should be an automatic, guaranteed to succeed operation? No user-mode application should ever be able to interrupt a critical kernel-mode operation like 'shutdown'. Not ever. There are several use cases where a timely, user-intervention-free shutdown is critical to the correct functioning of an operating system. It's an operation that must always succeed, or the OS is broken.
Have you ever done a remote reboot and had the machine not come back up, because the OS hung during shutdown, for whatever reason? I have, many times, and it's not fun. If you're not working on a server with an integrated management board, but a PC or a beige-box server in a remote lights-out environment, you're basically out of options if that happens.
What if it's a shutdown triggered by a UPS? The server now has just a couple of minutes to shut down cleanly. If it just sits there waiting for the user, it won't be a clean shutdown when it finally loses power, not to mention that it's wasting precious battery power when it doesn't have to.
Laptops and batteries come to mind also. I've once put a laptop into its bag, only to realize 10 minutes later from the hideous burning smell that the OS hadn't really shut off, it had just turned the screen black.
Your reasoning sounds like the excuse of a lazy developer. Why can't applications be written in such a way that they can be shut down quickly and reliably without user intervention? This has been standard for database systems for decades, but GUI application developers are only now catching up. Firefox can now recover almost all of its 'state' after even a crash, which is a good start, but why do trivial applications like text editors ask stupid questions like "Would you like to save this file?" and HANG the machine during shutdown? Is it so hard to respond to a "machine shutdown" event by serializing the application state to a temporary file, and then restoring it when the user runs the app again next boot?
This is actually probably quite doable, but would be filthy expensive.
Most people don't realize, but digital electronics is way, WAY ahead of what you get in your home PC, if you're willing to pony up the cash.
For example, non-Silicon based semiconductors often outperform the good old standard stuff significantly. Silicon is by no means the fastest, it's just the cheapest. Gallium Arsenide and Indium-based materials can both clock many gigahertz higher than Silicion for the same process size and power dissipation. They're toxic, fragile, and the largest wafer sizes are tiny, so not exactly mainstream, but available now.
The real performance king though is the Rapid Single Flux Quantum process, which can go over 100 GHz easily. It's used in things like radio telescope amplifiers and high-performance DSPs for military radar. Sure, it requires liquid helium cooling, but it also only requires milliwatts per gigaflop, so it's just about the only technology that'll let you squeeze a petaflop into a box and not have it melt into slag. That still means you'd need something like a kilowatt of cryogenic cooling, which is nontrivial, but still, I'd say it's doable with a bit of engineering wizardry.
Microsoft goes out of their way to make their.NET libraries skip 10% of what a typical developer actually requires.
To my eyes, there are deliberate looking 'holes' in the API. If you Google for a solution, lots of helpful web pages explain how to do native Win32 interop. Some of these come from Microsoft themselves. Seems helpful, right? Except that if they know that there's a problem, and they know that enough developers are experiencing it that they need to put up a technet article about it, then why isn't it already solved in the standard library? It's up to version 3.5 now, surely, they've had the time!
The answer is that Microsoft is forcing developers to calls small bits of the Win32 API so that their.NET applications will never be truly portable.
Similarly, there's a 'convenient' WebBrowser control that you can use to display HTML. Of course, it's always the IE control, and behave just like IE does. Good luck porting that piece-of-shit to any other platform! Sure, you can embed Webkit or Gecko or something on another platform, but then no application that relies on specific behavior of the WebBrowser control will work 'as-is' on Mono.
How many patents does Microsoft have related to IE? Dozens? More? How are you, some hobbyist developer, going to avoid THAT landmine?
Firefox deployment is seriously hampered by the lack of official MSI packages and administrative templates for Group Policy.
Lots of people have pointed out that some random third parties have MSI packages on their website. That's nice, but my boss won't deploy a core application downloaded from some guy's homepage. So what's left... I can roll my own package, except that Firefox has hundreds of fiddly little files, which change rapidly across versions, so I have to do this over.. and over.. and over. By hand. For every release.
If Mozilla devs had half a brain, and actually read any one of the dozens of feature requests that have been sitting in bugzilla for years now, they would have added an MSI build step to the build script, and with almost zero effort they could be spitting out MSI packages automatically. They wouldn't have to change their setup program, they'd just have to add a link to a "administrative download" page for network admins with the MSI packages there. Lots of vendors do this. For example, you can get MSI packages for the Flash plugin.
I just can't understand why they wouldn't have done this years ago. Maybe they feel like anything Microsoft is 'teh enemy', but then again something like 85% of Firefox installations are on Windows... so that would be silly.
I suspect that the real problem is that the Mozilla devs completely and totally ignore bugzilla issue reports. I don't know why they even bother running the website. They should just turn that server off to save power.
I can understand if they ignore random feature requests by 'AweSomeHax0r57', but the devs also seem to ignore common showstopper error reports that cause severe data corruption, so I'm fairly certain they just ignore the whole site.
For example, back around the 2.0 days there was a bug that would wipe all your settings, including bookmarks, history, cookies, everything. If you were a Thunderbird user, it would also blow away your mail too, for your convenience. Even without the source, I figured out the problem in 30 minutes: It's the same issue that plagued ext4 - Firefox was writing settings out by simply overwriting your files in-place, one line at a time. If it crashed during shutdown, you ended up with almost nothing left, and your bookmarks, history, everything would be wiped. I cheerfully created a Bugzilla account so I could report my findings, but found to my dismay that the error had been reported already... four fucking times. Each duplicate had hundreds of messages from panicked users begging for help in restoring their data. Some of the reports were old. They had accumulated serious history, across major product releases. At least three of the reports had users reporting the precise cause of the issue, the area of the source where it was happening, and at least one guy had proposed the correct fix (the atomic rename method for replacing files).
I still sometimes lose all of my settings in Firefox. It's rare, but it happens. I suspect it's something else these days, but as far as I know, that file-overwrite bug wasn't fixed for something like two or three years.
If the Mozilla dev team can't be bother to fix a showstopper data corruption bug, I'm not holding my breath for 'enterprise features'. Until they get their act together, Firefox will always be that 'other' browser in the corporate world.
True, but in my experience the difference between Java and the.NET stuff is performance.
Sure, some synthetic benchmarks might show that Java beats.NET by some margin, but in practice, that's not what matters.
What users notice is that Silverlight loads almost instantly (as fast as Flash), and that desktop.NET apps run just as fast as native, and look as good, or better than native apps.
I can always tell when a Java app starts because the JVM startup brings my machine to its knees, and the end result is inevitably some "non platform specific" GUI that looks like a Solaris desktop app from the late 1990s.
The % utilization metric is a red herring. Most servers are underutilized by that metric, which is why VMware is making so much money consolidating them!
Users don't actually notice, or care, about CPU utilization. What users notice, is latency. If my computer is 99% idle, that's fine, but I want it to respond to mouse clicks in a timely fashion. I don't want to wait, even if it's just a few hundred milliseconds. This is where parallel computation can bring big wins.
One thing I noticed is that MS SQL Server still has its default "threshold for query parallelism" set to "5", which AFAIK means that if the query planner estimates that a query will take more than 5 seconds, it'll attempt a parallel query plan instead. That's insane! I don't know what kind of users Microsoft is thinking of, but in my world, if a form takes 5 seconds to display, it's way too slow to be considered acceptable. Many servers now have 8 or more cores, and 24 (4x hexacore) is going to be common for database servers very soon. In that picture, even if you only consider a 15x speedup due to overhead, 5 seconds becomes something like 300 milliseconds!
Ordinary Windows applications can benefit from the same kind of speedup. For example, a huge number of applications use compression internally (all Java JAR files, of the docx-style Office 2007 files, etc...), yet the only parallel compressor I know of is WinRAR, which really does get 4x the speed on my quad-core. Did you know that the average compression rate for a normal algorithm like zip is something like 10MB/sec/core? That's pathetic. A Core i7 with 8 threads could probably do the same thing at 60 MB/sec or more, which is more in line with, say, gigabit ethernet speeds, or a typical hard-drive.
In other words, for a large class of apps, your hard-drive is not the bottleneck, your CPU is. How pathetic is that? A modern CPU has 4 or more cores, and it's busy hammering just one of those while your hard-drive, a mechanical component, is waiting to send it more data.
You wait until you get an SSD. Suddenly, a whole range of apps become "cpu limited".
Slashdot Mirror
It's true and false.
I saw a great comment on Slashdot once: The article was about some research being done at CERN, but it was a bit heavy on the jargon, and nobody could quite work out enough of the technical details to make really insightful comments. Lots of people made quipped along your lines: "it's necessarily complex", "jargon of the field", etc...
However, one of the comments was from a scientist working at CERN, on a very similar experiment. He didn't have a clue what the paper was about either, because the language was so cryptic, and so full of acronyms and jargon that overall it was just too hard to understand. If a guy in the same building, working on the same stuff can't understand what you're doing, how is it of value to the scientific community at large, many of whom don't even speak english as a first language?
Scientific papers in some fields have been suffering from a tendency to use language similar to patent legalese - it's obscured not to meet a specific purpose, but to try and make mundane discoveries sound complex. It's a type of showing off - the author is demonstrating that he's one of the 'elite'.
It hasn't spread to all sciences, yet. I studied Computing Science and Physics at university, in equal amounts. I can read and understand most Comp Sci research papers, but almost no modern Physics papers at all. Unfortunately, I see Comp Sci slowly getting worse. Source code is getting rare in papers, and the use of abstract maths using greek symbols is increasing. I suspect it'll get to the point pretty soon that a skilled programmer won't be able to make heads or tails of a Comp Sci paper, even if it's directly applicable to the problems encountered by a programmer day to day.
99% of all scientists in Galileo's day believed the earth was the center of the universe.
Actually... no. It was the 99% of the indoctrinated followers of the catholic church that believed that, despite mounting evidence to the contrary from an increasing number of secular scientists. Galileo was NOT the first to use a heliocentric model, he based his theories on the work of other scientists. Mind you, at the time, most astronomers were really astrologers, and hired by the church.
The parallel with the modern issue of climate change is uncanny. The minority of objective scientists are telling a truth the conservative leadership doesn't want to hear, and instead of entering into a reasoned debate, the conservatives are attempting futile smear campaigns.
Thank god we don't execute scientists any more for doubting the dogma of the leading elite, or climatology would be a dangerous profession right now!
Congrats! Oh wait...
Start-up Claims SSD Achieves 180,000 IOPS
Claims? As in no one else but the company has stated this "fact"? I wish this article waited for a review before being posted :S
It's not outside what I'd expect for a next-gen enterprise SSD. The PCI-E FusionIO cards can easily do 100K IOPS sustained. I'm just surprised the SAS bus can send that many commands per second. I guess the SCSI wire protocol scales better than I assumed.
The bandwidth numbers are actually relatively low - that's the bus speed limiting the drive. I suspect that pretty soon, most enterprise-grade SSDs will connect using the PCI bus to avoid that.
They basically admitted that they didn't have the technical strength to build something like that at their scale.
Wow.
They need to hire me, right now, for $3000 a day. Seriously, if they can't do that, I could work miracles for them, and I'd be worth every cent of that money. Scaling session IDs up is easy - in fact, it's a trivially parallelizable problem by nature.
Author misses lots of things, and makes all sorts of invalid comparisons.
For example, the invention of the electric light may seem like a big thing, but there were centrally powered lighting systems already when it was invented - such as town gas lighting. Sure, electric lights are better, but one could say it's just an 'incremental improvement'. It's just a matter of perspective.
And while the lightbulb was a big invention, it was largely unchanged for the first 50+ years. Almost every light bulb was a hot filament in a vacuum. More recently, we've been making entirely new sources of light, using entirely new chemical or physical principles.Think LEDs, OLEDs, all sorts of lasers, bioluminescence that we can now splice into rats and bunnies at will, etc... We've even made rather esoteric sources of light like beta-radiation powered lights that last for a decade.
The author also makes comments like this:
But despite daily announcements of one breakthrough or another, morbidity and mortality from cancer and stroke continue practically unabated, even in developed countries.
Well... duh. Something has got to kill us in our old age eventually, and it'll be the diseases that are hard to cure, obviously. Until we develop some sort of immortality, that's not going to change. 100% of people will die, of something, no matter how good medicine is.
Until we all become immortal, what about the major advancements, like the recent developments in growing organs? It's still in it's early stages, but even what we've got now is a massive leap forward in medicine, almost as big as the invention of modern surgical techniques.
People should really read the FAQ and the numbers.
To sum up: it's significantly more expensive, but since glass doesn't wear like asphalt does (it either works or breaks -- and it doesn't generally break from compressive stress, only torsional stress and impact), it should last longer and need less maintenance. And since you also get power out of it, displace plow crews, etc, they make the argument that it'll be a better investment if they can make the panels for $10k or less each.
Given that the one-off prototype is to cost $100k, and they have the potential for a *huge* amount of mass production, I don't think it's all that unrealistic. I'd still like to see how they handle in the real world, of course, but hey, that's why you give funding to build prototypes. ;)
Oh... yes! The numbers! I love the wishful naive thinking on that page, it's just brilliant.
For example, lets examine one of the pieces of insanity on his site. He mentions embedding supercapacitors into the road surface to store energy (I assume overnight). If you don't know what those things are, they would be the filthy expensive, highly experimental, rarely used in commercial products devices with lower than battery storage capacity. I'm sure they'll improve, but I can come up with fancy plans too if I can have parts made of unobtanium.
I particularly like the plan to use the ultracaps to store sufficient power to melt ice off the roads. The inventor clearly doesn't remember his 1st year Physics, where we learnt that the the enthalpy of fusion of water is surprisingly high compared to most other chemicals.
Ok, lets get practical: I'm basing this off the technical specs (PDF) for one of the beefier ultracapacitors made by one of the top companies in the biz - Maxwell Technologies. (note: I'm sure better devices are available from somewhere else, will be soon, etc.. bear with me)
It states that a device that is about 17.6cm high and has an area of 18.9cm x 51.5cm has a total capacity of 55Wh (~200kJ). That's a big capacitor.
So if you made a road surface with it, every 973.35 cm^2 area would have 200kJ of stored power for it. That's about 200J per cm^2.
Since the enthalpy of fusion of water 333 J/g, then 200J of energy will melt 0.6g of water. A layer of water (or ice) 0.6g/cm^2 is 6mm deep.
To summarize, this guy's fancy 'invention', if 100% efficient could melt 6mm of ice (or something like 5cm of snow), assuming that the weak winter sunlight was sufficient to fully charge the capacitors during the previous day. That's assuming the entire road surface has a layer of supercapacitors in it 17.6cm thick (that's 7 inches for you yanks).
Even if you gave the benefit of doubt and assumed a 10x improvement in supercapacitor technology, you still have to factor in that he plans to use the solar power capacity for other things too, like lighting up the LED arrays built-in to the road, and to power nearby homes. Not to mention that no matter how much capacity you have, there's not enough sunlight to charge it.
Note that the cost estimates conveniently left out the cost of the ultracaps. On one of the pages, he mentions a target price of USD48 per square foot. The Maxwell ultracap is about 1 square foot, so we're looking at $48 split between a square foot of: Solar cells, the glass coating, an ultrapacitor 7 inches thick, high intensity LEDs, heating coils, power management electronics, the road substrate, and more.
Who was the moron who gave him $100K? Can I have my free money now too? I can come up with all sorts of wild plans also that make zero fiscal sense!
My bad, I just googled for a random picture of solar-panel things in the desert. 8)
The fact that it hasn't been built just strengthens my point. Event the 'optimal' solar panel sites are uncompetitive with traditional forms of power, or the cheaper forms of green power (especially wind). Throwing away a bunch of efficiency on top of that is just crazy.
...Also why do they have to make roads out of them.. where did that come from? Just put them out on land somewhere, you don't have to drive all over them.
This was my first thought too. Making the solar panels into roads (or vice versa) is compounding the problem. Just put the 25,000 mi^2 of solar panels in the middle of the desert and call it even. Adding a layer of glass or some sort of protective surface is going to lessen the efficiency and raise the cost of production and maintenance. I'm all about green energy, but there are better places we could be spending our money and energy.
Back at uni, I did a mini-course on the the Solar Car challenge, because my University made some of the solar panels for the top cars, and we also had a car that entered and did fairly well (for a low budget). One of the things we learned was that solar cells lose efficiency very quickly from a variety of things. The two that most researchers ignored in the lab but mattered in the field was heat and dirt. The cars in the race are washed with cold water thoroughly at every opportunity because colder, cleaner cells are substantially more efficient. Think CPU overclocking - lower temperatures improves things a lot.
Now lets compare this situation to a typical road which is:
a) Blistering hot most days.
b) Really, truly, thoroughly dirty.
Sounds like the perfect place to put an expensive solar cell panel!
Another thing we learned is that a single "test" panel in a lab operates very differently to a bunch of real panels in the field. What a lot of naive researchers miss is that the amount of sunlight over the entire collecting surface in the real-world is not constant. For a one-square-foot panel, it is, but for any significant surface (the size of a car, road, whatever), it won't be. The surface will be curved or partially shadowed. This matters a lot because if you just connect a bunch of cells together, they perform roughly the same as the worst of the lot. If there's a few cells under a shadow, that's drags down the efficiency of the panels receiving sunlight. To efficiently extract energy from a bunch of panels receiving differing amounts of light takes a bunch of expensive power management electronics that can combine the different cell outputs in the right way.
In practice, cells are so expensive that the best place to put them is on huge, flat, orientable panels out in the desert where there's no clouds, no rainfall to cake dirt onto the panels, and they can be oriented to face the sun at all time, like this array in southern California.
You failed to understand that re-purchasing a vast set of licenses for some 64-bit OS just for the sake of it being 64-bit when your entire portfolio of applications and hardware (you need drivers) was built for the 32-bit platform... Well, it doesn't make any sense. It means that you intended to migrate your office from a 32-bit OS to a 64-bit OS without any other justification beyond "I felt like it", which carries a hefty price tag and absolutely no technical justification.
And moreover, the claims that 32-bit apps "just run fine on 64-bit windows" are made from the same people who claimed that your Windows XP apps "just run fine on Windows Vista" and we saw how that panned out.
Why would you have to repurchase application licenses to upgrade Windows? I'm not aware of any applications that tie licensing to the OS in such a low level way. That's just bizarre. Most apps are licensed by seat, concurrent user, or 'enterprise agreement'. I know of some apps where you can't switch between Linux and Windows like that, but the only apps I've seen prevent an OS upgrade within the same family are backup and anti-virus apps, but most allow it. Also, since Vista, Windows licenses are interchangeable between 32-bit and 64-bit editions.
You do realize that just because an application was "built for 32-bit platforms" doesn't mean it'll magically perform terribly on a 64-bit machine, right? In fact, my experience has been the opposite: A 64-bit OS is noticeably faster for 32-bit apps, because the disk cache is much larger. There's a little-known limit in 32-bit Windows: it can't create a disk cache much larger than 450MB, no matter how much physical RAM you have, because of the way the kernel address space is laid out. If you run apps that exceed that limit, your disk gets thrashed even though you have free memory. A user running a couple of Office apps can easily exceed that limit.
Most people don't realize that even 32-bit server editions can't go over that ~400MB limit. It cracks me up to see 32-bit file servers with 8GB of memory. I've seen 10-fold reductions in disk IO rate by switching to 64-bit.
There may not be any technical reason to still hold on to a 32-bit OS but there sure are economical ones. As soon as you have an entire office filled with desktops, each one running tens of thousands of dollars in software that were released only to the 32-bit platform, then you will consider that there is a very good reason to still run a 32-bit system.
Wow, the Mods are asleep or not paying attention to the dozens and dozens of post which have pointed out that 32-bit apps run just fine on 64-bit windows.
XP x64 was awful. Don't use it. Vista x64 is way better than XP x64.
Windows 7 x64 seems to run quite smoothly. I haven't run into a 32 bit application (driver, yes; application, no) that I can't run. Granted, I don't run a ton, but several old games a few apps (like Sibelius).
Mod parent down, he has no idea what he's on about.
Windows XP 64-bit is exactly the same as Windows 2003 Server 64-bit. They share the same service pack files! The only difference is some minor licensing restrictions (can't install server components, 2 CPU sockets max, etc...). Drivers that work for either will work for both.
I had XP 64-bit on both my desktop at home and my work laptop for years, basically since it was released. I never had any issues with it, even on fairly unusual hardware. It is as light-weight as XP, but has 64-bit kernel caches and buffers, so it runs much better on the same hardware, and ran circles around both 32-bit and 64-bit Vista. Only now that Windows 7 is out have I upgraded my laptop to that, because Vista was just too painful to use in comparison to XP 64-bit.
Mind you, I wouldn't install anything other than Windows 7 64-bit on desktop PCs these days, but before that, XP 64-bit was a very nice OS.
I've used an unsigned driver in vista 64 bit and I've never seen this. I used EasyBCD and selected the option for allowing unsigned drivers in 64 bit windows and everything was fine. No test mode or anything.
I just did something similar on Windows 7 x64, and no "test mode" there either!
They can just park a ship off a Manhattan and light one off if they wanted to.
Ships move at 20-30mph. Ballistic missiles move at 15,000mph +. If we make it so that our enemies have to get a ship into one our harbors, it becomes a much simpler problem. We would need to have more Coast Guard people to basically board every ship, with neutron detectors, but, its something we can do. We can track ships as they are approaching the USA, track them as they leave ports, follow them, and pretty much monitor every boat on the ocean.
Why doesn't the US do that now?
Several analysts have pointed out that the easiest way to get a nuke into a major US city is by shipping it a standard shipping container. Many cities have container docks near the central business district, where even a small nuclear explosion from a low-tech device would inflict massive damage.
Containers can't even be properly scanned for radioactive materials when they're being unloaded in the harbor. The billion dollar project to install radiation detectors at US ports is apparently a failure - it detects false positives on things like shipments of granite, but fails to detect several varieties of Uranium and Plutonium, including the forms used in nukes. Not to mention that detecting a nuke at that stage is a bit late, it's already in the right place. And even if container scanning is perfected, there's dozens of other ways of smuggling a bomb in - cruise ships, private yachts, oil tankers, scuba divers, submarines, indirectly through Mexico or Canada, etc...
I read a great American Scientist article on why the current border protections against nuclear materials are a joke, but I can't find it now. I did find another research paper that covers most of the key points though.
Hmm... one thing that might be going on is that when you upgrade to an SSD, suddenly everything becomes CPU limited.
I noticed that before my SSD, the CPU would only rarely go above 30% for any length of time, but now it hits 100% regularly.
The Intel Core i7 in your desktop is at least 3x the speed of your laptop CPU: it has double the cores, and each core is more efficient.
I saw a similar thing recently: I have a Vertex drive in a 2.66 GHz C2D laptop (4GB, running Windows 2008 x64), and a co-worker put an identical drive into a quad-core desktop. The desktop system was quite noticeably faster.
We've all become used to CPU upgrades having little effect because of the bottleneck of the disk system, so when the disk IO issues go away, it can be surprising how much difference a decent CPU makes.
That's why I stopped playing WoW.
My strategy has always been the same: I liked to find the ultimate 'combo' and push my character far beyond what the original designers intended. For example, in WoW I played an Alchemist Paladin - the theory being that by using Paladin shields and spells, I could survive just long enough in high-level zones to pick some uber herb, make uber potions, and then use said potions to turn myself into the Ubermensch. I could pull it off too. I was able to go into zones 10 levels above me, and walk out with bags full of herbs. I did things like jump off cliffs on low-level zones into corners of high level zones while shielded, so I would be able to pick the one herb that was there out of range of any monsters. I carefully waited for patrols of elite mobs to wander past, and I'd carefully sneak past them into high-level zones.
My strategy would have worked, except that Blizzard put in totally artificial limits into the game, so that if you did manage to cleverly set your character up like that, it wouldn't actually do you any good at all. The best strategy (in terms of efficiency) by far was always to grind roughly equal level mobs while doing boring quests (kill X of Y mobs). That pissed me off. The game actively forced you to play in a boring, linear, repetitive manner. Creativity was not only discouraged, but in many cases virtually impossible.
I'm sure they did it to prevent 'twinking', or 'exploiting', or whatever, but it made the game deadly boring for me. I much prefer games like Nethack, where you can find a wand of wishing after playing the game for 3 turns, and still die four turns later. (Yes, it happened to me. I now know that I shouldn't wish for artifacts that blast the wielder with their power when picked up if I'm still at level 1.)
You must have had a faulty setup in some way.
I put an OCZ 256GB Vertex into my laptop, and it got so fast it's just unbelievable. I did have to upgrade the firmware, the stock firmware was both buggy and slow. Since then, several major revisions of the firmware have been released. I'll try this new GC firmware when I upgrade to Windows 7.
Still, I have more confidence in the long term reliability of the Intel drives. A friend of mine purchased a Vertex drive as well, and it died a week after he got it, and mine had data corruption issues at first until I patched the firmware.
Especially in a small business, your users will rebel if they can't install (or use) their software... which is quite reasonable given most people are still running Windows XP, and most XP software is not capable of being installed or sometimes even used without admin access... this is especially troublesome if that user happens to be the CEO/Owner.
You hardly ever have time/resources to "do it properly" in a small business, unless what you're "doing right" is a core competency of the business. The trick is to convince the guy who signs the checks that it is business/mission critical (often non-trivial).
If you have full-time IT staff at the company, it's possible to convince users that they're better off, you just have to approach things from the right direction. For example, making users "non-admin" and "user-initiated software installation" are not mutually exclusive. Under Windows, you can publish a list of pre-prepared applications, and users can be allowed to install anything from that list at will simply by clicking an icon. It's much harder if you're not always on site, ready to install applications for users in an ad-hoc manner.
Also, before you remove user's admin rights, make sure that their new desktop image contains every app they have previously used. If they don't feel the need to install any additional software, they won't even notice that they can't. For example, you can go a long way by making sure that every workstation has all the basic 'frameworks' like Java, .NET, Flash, etc... that way, every website will work.
There are several use cases where a timely, user-intervention-free shutdown is critical to the correct functioning of an operating system. It's an operation that must always succeed, or the OS is broken.
Then you should be using the command outlined above: "shutdown /s /t 1 /f"
This will shutdown the computer and close any force-close any open applications.
Hell, go through the PID list and KILL any applications you want, then issue a shutdown request.
That's a complex, factual, obscure, smug, and incorrect answer. Typical developer attitude.
I do know about that command, and have been forced to use it, but I avoid it if at all possible, because, as you said yourself, it force-closes applications. On a server, that can lead to data loss.
That's like saying that if I have trouble going to sleep, then I should keep a revolver next to my bed, because that's the only way to make sure that I'll go down, and stay down.
Is it so hard to respond to a "machine shutdown" event by serializing the application state to a temporary file, and then restoring it when the user runs the app again next boot?
It is very easy, it's called "Hibernate". All modern computer/OS combos can do it.
Just don't expect any connection-dependent applications to work correctly after.
Shutdown is not the same as hibernate. You can't apply an OS patch, and then hibernate to apply it, for example. My original example of Firefox is a great example of how most apps should work. After the "apply patch and restart now?" prompt, Firefox returns to where it was before.
One of the reasons I hate Windows so much is that I can't even rely on the piece of crap OS to shut down if I tell it to shut down and then walk away. It literally expects me to sit there for up to five minutes while it 'saves my settings' and stops all the processes to ensure the bloody thing turns itself off.
Sounds like you hate computers in general and have likely chosen the wrong profession.
Why is his requirement unrealistic? I'm a computer professional, I love computers, and I fully agree with the original poster. Why should I wait and watch what should be an automatic, guaranteed to succeed operation? No user-mode application should ever be able to interrupt a critical kernel-mode operation like 'shutdown'. Not ever. There are several use cases where a timely, user-intervention-free shutdown is critical to the correct functioning of an operating system. It's an operation that must always succeed, or the OS is broken.
Have you ever done a remote reboot and had the machine not come back up, because the OS hung during shutdown, for whatever reason? I have, many times, and it's not fun. If you're not working on a server with an integrated management board, but a PC or a beige-box server in a remote lights-out environment, you're basically out of options if that happens.
What if it's a shutdown triggered by a UPS? The server now has just a couple of minutes to shut down cleanly. If it just sits there waiting for the user, it won't be a clean shutdown when it finally loses power, not to mention that it's wasting precious battery power when it doesn't have to.
Laptops and batteries come to mind also. I've once put a laptop into its bag, only to realize 10 minutes later from the hideous burning smell that the OS hadn't really shut off, it had just turned the screen black.
Your reasoning sounds like the excuse of a lazy developer. Why can't applications be written in such a way that they can be shut down quickly and reliably without user intervention? This has been standard for database systems for decades, but GUI application developers are only now catching up. Firefox can now recover almost all of its 'state' after even a crash, which is a good start, but why do trivial applications like text editors ask stupid questions like "Would you like to save this file?" and HANG the machine during shutdown? Is it so hard to respond to a "machine shutdown" event by serializing the application state to a temporary file, and then restoring it when the user runs the app again next boot?
This is actually probably quite doable, but would be filthy expensive.
Most people don't realize, but digital electronics is way, WAY ahead of what you get in your home PC, if you're willing to pony up the cash.
For example, non-Silicon based semiconductors often outperform the good old standard stuff significantly. Silicon is by no means the fastest, it's just the cheapest. Gallium Arsenide and Indium-based materials can both clock many gigahertz higher than Silicion for the same process size and power dissipation. They're toxic, fragile, and the largest wafer sizes are tiny, so not exactly mainstream, but available now.
The real performance king though is the Rapid Single Flux Quantum process, which can go over 100 GHz easily. It's used in things like radio telescope amplifiers and high-performance DSPs for military radar. Sure, it requires liquid helium cooling, but it also only requires milliwatts per gigaflop, so it's just about the only technology that'll let you squeeze a petaflop into a box and not have it melt into slag. That still means you'd need something like a kilowatt of cryogenic cooling, which is nontrivial, but still, I'd say it's doable with a bit of engineering wizardry.
It's even worse than that.
Microsoft goes out of their way to make their .NET libraries skip 10% of what a typical developer actually requires.
To my eyes, there are deliberate looking 'holes' in the API. If you Google for a solution, lots of helpful web pages explain how to do native Win32 interop. Some of these come from Microsoft themselves. Seems helpful, right? Except that if they know that there's a problem, and they know that enough developers are experiencing it that they need to put up a technet article about it, then why isn't it already solved in the standard library? It's up to version 3.5 now, surely, they've had the time!
The answer is that Microsoft is forcing developers to calls small bits of the Win32 API so that their .NET applications will never be truly portable.
Similarly, there's a 'convenient' WebBrowser control that you can use to display HTML. Of course, it's always the IE control, and behave just like IE does. Good luck porting that piece-of-shit to any other platform! Sure, you can embed Webkit or Gecko or something on another platform, but then no application that relies on specific behavior of the WebBrowser control will work 'as-is' on Mono.
How many patents does Microsoft have related to IE? Dozens? More? How are you, some hobbyist developer, going to avoid THAT landmine?
Mod parent up.
Firefox deployment is seriously hampered by the lack of official MSI packages and administrative templates for Group Policy.
Lots of people have pointed out that some random third parties have MSI packages on their website. That's nice, but my boss won't deploy a core application downloaded from some guy's homepage. So what's left... I can roll my own package, except that Firefox has hundreds of fiddly little files, which change rapidly across versions, so I have to do this over.. and over.. and over. By hand. For every release.
If Mozilla devs had half a brain, and actually read any one of the dozens of feature requests that have been sitting in bugzilla for years now, they would have added an MSI build step to the build script, and with almost zero effort they could be spitting out MSI packages automatically. They wouldn't have to change their setup program, they'd just have to add a link to a "administrative download" page for network admins with the MSI packages there. Lots of vendors do this. For example, you can get MSI packages for the Flash plugin.
I just can't understand why they wouldn't have done this years ago. Maybe they feel like anything Microsoft is 'teh enemy', but then again something like 85% of Firefox installations are on Windows... so that would be silly.
I suspect that the real problem is that the Mozilla devs completely and totally ignore bugzilla issue reports. I don't know why they even bother running the website. They should just turn that server off to save power.
I can understand if they ignore random feature requests by 'AweSomeHax0r57', but the devs also seem to ignore common showstopper error reports that cause severe data corruption, so I'm fairly certain they just ignore the whole site.
For example, back around the 2.0 days there was a bug that would wipe all your settings, including bookmarks, history, cookies, everything. If you were a Thunderbird user, it would also blow away your mail too, for your convenience. Even without the source, I figured out the problem in 30 minutes: It's the same issue that plagued ext4 - Firefox was writing settings out by simply overwriting your files in-place, one line at a time. If it crashed during shutdown, you ended up with almost nothing left, and your bookmarks, history, everything would be wiped. I cheerfully created a Bugzilla account so I could report my findings, but found to my dismay that the error had been reported already... four fucking times. Each duplicate had hundreds of messages from panicked users begging for help in restoring their data. Some of the reports were old. They had accumulated serious history, across major product releases. At least three of the reports had users reporting the precise cause of the issue, the area of the source where it was happening, and at least one guy had proposed the correct fix (the atomic rename method for replacing files).
I still sometimes lose all of my settings in Firefox. It's rare, but it happens. I suspect it's something else these days, but as far as I know, that file-overwrite bug wasn't fixed for something like two or three years.
If the Mozilla dev team can't be bother to fix a showstopper data corruption bug, I'm not holding my breath for 'enterprise features'. Until they get their act together, Firefox will always be that 'other' browser in the corporate world.
True, but in my experience the difference between Java and the .NET stuff is performance.
Sure, some synthetic benchmarks might show that Java beats .NET by some margin, but in practice, that's not what matters.
What users notice is that Silverlight loads almost instantly (as fast as Flash), and that desktop .NET apps run just as fast as native, and look as good, or better than native apps.
I can always tell when a Java app starts because the JVM startup brings my machine to its knees, and the end result is inevitably some "non platform specific" GUI that looks like a Solaris desktop app from the late 1990s.
The % utilization metric is a red herring. Most servers are underutilized by that metric, which is why VMware is making so much money consolidating them!
Users don't actually notice, or care, about CPU utilization. What users notice, is latency. If my computer is 99% idle, that's fine, but I want it to respond to mouse clicks in a timely fashion. I don't want to wait, even if it's just a few hundred milliseconds. This is where parallel computation can bring big wins.
One thing I noticed is that MS SQL Server still has its default "threshold for query parallelism" set to "5", which AFAIK means that if the query planner estimates that a query will take more than 5 seconds, it'll attempt a parallel query plan instead. That's insane! I don't know what kind of users Microsoft is thinking of, but in my world, if a form takes 5 seconds to display, it's way too slow to be considered acceptable. Many servers now have 8 or more cores, and 24 (4x hexacore) is going to be common for database servers very soon. In that picture, even if you only consider a 15x speedup due to overhead, 5 seconds becomes something like 300 milliseconds!
Ordinary Windows applications can benefit from the same kind of speedup. For example, a huge number of applications use compression internally (all Java JAR files, of the docx-style Office 2007 files, etc...), yet the only parallel compressor I know of is WinRAR, which really does get 4x the speed on my quad-core. Did you know that the average compression rate for a normal algorithm like zip is something like 10MB/sec/core? That's pathetic. A Core i7 with 8 threads could probably do the same thing at 60 MB/sec or more, which is more in line with, say, gigabit ethernet speeds, or a typical hard-drive.
In other words, for a large class of apps, your hard-drive is not the bottleneck, your CPU is. How pathetic is that? A modern CPU has 4 or more cores, and it's busy hammering just one of those while your hard-drive, a mechanical component, is waiting to send it more data.
You wait until you get an SSD. Suddenly, a whole range of apps become "cpu limited".
That's an interesting new metric: Average number of arms remaining per researcher.
"Ooo... he really didn't like that at all! Bob's going to need help feeding himself tomorrow!"