In societies where women are not even meant to be seen let alone heard can clearly live a far more relaxed live with more women waiteing on the men.
But in the western world woman presume the dominant role within the house (most men beg their wives to let them buy that new PC, TV, etc..) more than one woman would mean more than one dominant control freak in the house.
Polygamy in the west is a terrible idea. I have no frigging idea how to keep up with all the demands of one woman (my wife who's reading this now agrees that I'm clueless and she feels it's for my own good) having a second woman demanding I do this, that, or the other thing would kill me in a year or two.
P.S. - My wife wishes that I relay that men are good for three things. Opening jars, killing bugs and changing tires. She says, the other things she can buy in a store with or without batteries.
I'm not entirely sure if I remember the details correctly, but I'll try to go on memory since researching it should be left to someone else.
Intel licensed the 8086 processor (and 80286 if I recall correctly) to AMD per demand of IBM. Speculatively, I believe that Intel has always considered having a second x86 CPU source as a good thing. It keeps the monopoly monkeys off their backs.
When 386 came around, AMD stayed pretty quiet. But, as you would see from NEC who produced x86 processors with 286 functions (like pusha and popa) Intel started closing up their market. x86 already had established such a strong foothold that Intel became territorial and actually did attempt to monopolize the market by not granting 80386 licenses to anyone else.
AMD responded with the 386dx-40, which for the most part was the start of the megahertz war. Cyrix (spelling), Evergreen and a few others quickly jumped on board and all started releasing 386 clones at an incredible rate.
So far as I know, at this point, NECs response was the v40 which sold primarily in the embedded and Japanese markets.
The 486 generation was a terrible era for system builders. The 486 was still using a very simplistic front side bus and didn't make use of clock multipliers. Motherboards were stills shipping with socketed crystals so that you would modify the system speed by putting in a different crystal. Since there were now so many chipsets to choose from (there must have been 15 or more brands competing) and the chipset manufacturers weren't producing stable reference implementations, motherboards came in only two quality grades... bad and worse.
Now came the absolute worst part. VESA local bus. VESA local bus was pretty much the same thing as ISA in that it connected a periperal board directly to the processor's I/O and memory busses without any logic inbetween. This was a response to the extremely overpriced and overcomplicated EISA bus and the fact that ISA was only 16Mhz x 16-bits. Since the "standard" only allowed for 33Mhz busses, most board manufacters made boards that ran at 33Mhz... maybe 33.1 but certainly not 40 or 50Mhz. So since nearly all chips coming from places other than Intel used the faster bus rates, system stability was getting to be worse than shit.
This era nearly destroyed the x86 world since the concept of name brand motherboards and video cards was only for rich people. Micronics and DFI (the only players at the time) charged $350 for their motherboards while everyone else was asking about $100-$120. I can assure you, having worked in a computer store at the time, the quality difference was worth every penny. But back then, $200 was considered a lot of money for a hobbiest to lay out. The price difference for reliable name brand memory (only Kingston existed back then) was more than double that.
Intel began sueing everyone over the x86 license. In fact, I even liked the idea since nearly half the machines I was shipping out with 40Mhz busses were coming back over and over again with serious failures. The store I worked for tried a niche market which was "reliable clones" where we tried real hard to only use parts that were quality. After losing our asses over it since noone wanted to pay $2000 for a machine when the other stores only charged $900, we dumped that business... and it didn't matter, the 40Mhz FSB motherboard were still coming back broken all the time.
Intels choice to button up the market was ideal since in a way it was non only protecting its own assets and IP, but also trying to take the crappy clone chip makers off the market. And most of them did fall off the market. Many consolidated, many disolved. But in the end it left us with Cyrix and AMD. I think I remember Cyrix being purchased by VIA or someone else.
What's important about this consolidation is that even though Intel squashed a ton of companies, the companies that remained were the companies who had managed to gain a good enough reputation duing the x86 war to make enough money to not only su
I'll conceed on this point to a certain extent. I had thought a little about how I would implement this scheme in software after the point and I'll also respond to your response.
First the response.
Routers are crap for gigabit anyway... they're 10 times worse for 10Ge. That's probably the number one reason why layer-3 switches are far more attractive in backbone environments. In fact, it took me years to find a router that cost less than $10,000 that could in fact handle 100Mbit traffic 24/7 (I mean run at full capacity 24/7) and that router implemented portions of the IP stack in FPGAs.
If you're running a software based router to handle full gigabit traffic, then I'm forced to call you a tree killer (only joking). For a software based router to efficiently handle four gigabit ports at full capacity requires some pretty substantial hardware. Of course the #1 method of making it work better is to use a server Ethernet adapter which offloads at least the checksum calculations from the host machine. Sadly, most high end ethernet adapters stop there since to do much more they'd have to have interact more closely with the TCP/IP stack which generally doesn't work out too well in most OS NIC driver models.
Now that I've attacked your line of thinking, let me make an ammendment to my line of thinking.
I made another post suggesting capping the limit at 9-bits instead of 8-bits. This would work great since it would first of all keep the numbering scheme of the current IPv4 address space, expand the number of user inputable IP addresses by 16 fold, leave a tremendous number of address space in what I'd call the "utility area" or the space where trained (I hope) network professionals would configure large banks of auto configured addresses available (which would use a more complex numbering scheme). And best of all the math would pretty much all remain the same.
The goal of the project isn't to make networking professionals lives easier. In fact, I'd rather make it harder to lay waste to all the guys out there that actually need 8 hours to perform simple router configurations. That's political though. The goal is instead to avoid making a bunch of home users and hobbiests that finally learned to recognize and IP address able to still recognize an IP address.
Imagine this, your dad called you, there's something wrong with his computer and for some reason, before you can log on to fix the system from outside, you need him to reset his router to default settings and then open a port to let you in.
So you pull up the instruction manual to the router after asking him for the model number and eventually after hearing the serial number and manufacturing date, he finally tells you the model. Now, you walk him through the reset process and the process of connecting to the proper wireless network.
Now you want to get him to find out the IP address of the router. So you do an ipconfig, ifconfig, etc... and ask him to tell you the address.
With IPv4 you can simply say "There should be something like a telephone number, but it's separated by dots instead of dashes and there are 4 groups instead of 3." From that point, it should only take two or thress tries to get him to find it.
With IPv6 numbering, you're stuck with a "Ok son, let me get my glasses, you say there are a bunch of numbers with a bunch of colons or possible some colons and some dots, wow there are a lot of numbers here. This is twice as long as the code I had to type in when you walked me through reinstalling Windows XP on my machine".
There's a value to the IPv4 numbering system in that it's short and sweet. IPv6 is too damn long. All our hopes and dreams of IPv6 being a perfect auto-configured world is crap too. So, if it means that software and hardware developers would be forced to optimize for an arithmetic operation as opposed to a binary calculation, as someone who would have to do it, I'd say "No problem, we'll find a way" since in reality it's true, we'd find a way.
No, having implemented IP stacks for embedded router stacks, I'm pretty sure I do in fact understand the protocol pretty well.
What I'm pointing out is that the limitation which makes 3 digits in four groups ideal is that people find them easy to remember. By leaving 3digits by four groups standard, we can already suspect that people can continue to both remember them and not find them confusing.
But to soothe your mind, alter the standard to a 9 bit octet (yes I know what octet means, but I don't know the word for a grouping of 9) and choose to max at 0-511 instead. Therefore we can provide a direct decimal to base-2 mapping and still increase the address space by a factor of 16. Either way, it should have a pretty major positive impact.
Though now I'm getting carried away with this line of thinking. It appears that by using this method of thinking we could do something like use a double addressing system.
So we can use 0.0.0.0-511.511.511.511 to isolate the portion of the Internet which should be theoretically identifiable publicly and revert to the more complex IPv6 numbering scheme to address private and utility devices. For example, for each mobile phone device which is ALWAYS dynamically assigned IPs and therefore would not benefit from the simpler form.
I personally liked 0-999 instead of 0-511 since I feel that user inputable IP addresses should be more closely matched to human beings who think in base 10. From an IP stack development perspective, the only complication I see with this method is that IPv4 functioned so nicely since he subnet landed on a base-2 boundary. by placing it on a base-10 boundary instead, it would require an arithmetic comparison vs a binary comparison.
P.S. before you get snooty, at least take the time to read the entire sentence you quoted. I'd imagine that after making a direct reference to 2^10, it would have been obvious that I clearly do understand the importance of the value 255.
When there was no more space to build outward in Manhatten, then solution wasn't to try and produce more land. instead, they made the buildings taller (which worked well until '99)
People have no problems remembering up to four three digit groups. So why not, expand the address space to support 0-999 values instead of just 0-255. Sure, 999 isn't a byte, but it's close enough to 2^10. Sacrificing the remaining 25 values won't hurt much. But more importantly, it would increase the address pool from 4.2 billion (minus invalid values) to 1,000,000,000,000 (a trillion) which still allows something like 200 IP addresses for every person on the planet. And with technology like NAT which should be employed for security purposes should be more than we could ever use.
Not we just need some genious to figure out how best to map that mechanism to the base-2 or IPv6 world
With the upcoming release of the Nehlam platform, we would imagine that this business should become far easier to be a part of, after all, it moves the majority of the complex tasks into the CPU and the chipset is more of a peripheral and power saving feature bus. But the truth is that in upcoming Intel chips it will be necessary to implement a far more complex DMA system and so far as I can tell from the FSB documents I've read (have the real ones) it looks like most of the existing code used for developing chipsets will have to be tossed.
These days, it's pretty obvious that Via won't make a living competing with AMD on sales of AMD chipsets, so there's just not point in bothering.
I appears that to support the Nehlam processor, an entirely new memory architecture will need to be developed to compensate for PCI DMA.
At the moment, from what I understand, if Via were to pursue licensing the Nehlam FSB, pursue implementing the new memory architecture, etc.. they would use a tremendous amount of financial resources that could otherwise be allocated to protecting their currently threatened CPU business. After all, almost every single VIA device on the market has be rereleased using Atom, even if it meant putting a square peg in a round hole.
Via should take whatever money they make from supporting the Intel and AMD platforms and sink it all into making the Via platform something special.
P.S. - In case VIA is reading this, I highly recommend making a VIA optimized compiler based on one of the high performance open source projects. This way companies dependant on CODEC performance could actually tweak their code for Via as well as AMD and Intel as we do now.
Back in the late 80's I worked on System 38 and OS/400 based systems and the hottest topic of the time was graphical user interface design software for mainframe applications. It worked great too. See, mainframe applications don't generally have veyr advanced user interfaces, so to aquire and alter data through the on screen fields through a remote terminal emulation is simplistic in every way.
I believe IBM even sells some software like what I desribed, but that's not the point.
Give me a day or two in front of a Linux box with tn3270 and I'll make a system which will read all the relevant data from the mainframe ISAM to the local system so the values prior to the change can be stored for recovery later on.
Then I'll take another two days to make a program which will open up each record in the database which met the criteria concerning the minimum wage shift based on either name or social security number. Then I'll change the data to minimum wage or to the previous salary level depending on the task.
I would then make multiple test runs to ensure that it works before screwing the entire work force of public California.
Altogether, it would require about a week to do the task and maybe on that many records a few hours to a day to run.
If you want me to do it, I'll do it for $10,000 plus flight, rental car, and 4(+) star accomodations for me and my family.
Or you can probably find some college kid with no experience and no accountability and pay him/her $6.55 an hour for a month or three to do it:)
Enjoy
P.S. As to the guys all talking about how jobs are hard to get. I've worked at 3 major retail software/hardware vendors as a senior level developer. One thing I can safely say is.. apply for the job even if the skills required are something that interest you. Fact is, 90% of the time, the unrealistic requirements can be overcome if you sell yourself as someone capable of learning the technology. Most companies actually respect initiative over proficiency.
English is a more or less universal language these days... especially in the world of engineers. I worked professionally with Japanese, Chinese, Singaporian, Indian (a few different flavours), German, French, Swedish, Norwegian, Finnish, British (it's a different language in some places almost), Cuban, and more than a few from former soviet countries. Oh.. let's not forget Italians, I can write a book on that experience.
The languages I find most useful to attempt are:
- German
- Manderine
Let me justify.
French, while an attractive language is of very little use for most people in the computer business since every French person you will need to do business with will speak English. This is true for many other places, but the difference is, unless you feel you can learn to speak French so well you'd be indecipherable to a native, there's a very strong chance they'll just force you to speak English anyway. I think they take so much pride in their language that they consider Americans speaking French to sound like finger nails on chalkboards. I'm sure there are people that would differ with my findings on this topic, but I am a strong believer there's no universal answer to this issue.
Spanish, there's a niche market in computers where Spanish is a useful language, but if you can speak at least one other language and have massaged your brain to support multiple vocabularies, Spanish is a language which can be learned quite quickly on the fly and consuming financial resources to pay for university courses on the topic is probably wasteful.
Japanese, unless you intend to live and work in Japan with strong Japanese work ethic, learning the language could do you more harm than good. Collegues in Japan like to have their privacy when discussing business issues before communicating their findings to vendors or customers. If you understand their language, it may give you an edge in negotiations, but if they learn you know their language, they'll have to delay any decisions until they have a chance to speak privately. So, even if you learn it, it will work better not to use it unless you want to work in their country with their work ethics. Most American's working in Japan still work hard, but are given a little extra leg room since "He doesn't know any better". Learning the language means you've also probably learned some customs and you lose this advantage.
I can descibe the benefits or disadvantages of learning each language, but you get the idea so far. Here's the reason why to learn German and Manderine.
Manderine is obvious as it is the most spoken Chinese language. In the modern world and modern economy it is likely that at some point Chinese companies will begin hiring Americans to bridge gaps. Speaking their language will make you more valuable especially since much of their designs and documents will already be in Chinese. So it's a valuable skill.
German, this is an important one. Germany is only part of the issue. Learning German will buy you respect in Germany since they consider it flattering that you've learned their language. But getting a job in Germany is easy enough with just English. Ther point is that as an American you will learn a root language which is present in English, but the German roots in English are quite weak compared to the Latin roots. By learning German you'll learn the root language to most other European languages that aren't Latin bsaed. This means that you'll find it easier to learn Latin languages since you've molded your mind to support more than one language and you'll be able to learn Germanic languages much easier. This would give you the Scandinavian languages and quite a few eastern block languages.
The benefit to this is that in times (such as current times) you would be able to leave the U.S. and work in a strong European economy, purchase real estate and convert Euro to dollars during a strong European economy. Working in Germany right now as an American is EXTREMELY profitable and should not be ov
Well Swedish chicks speak English just fine. I'm an American in Norway and speak Norwegian and understand Swedish and Danish and they humor me and pretend to understand my Norwegian.
Use English around here, it's a bit of a waste to learn the language unless you're planning on working in a day care or elementary school to hit on the single mothers.
Seriously, depending on what kind of company it is and te focus of the project types.
In my company for example, we would have to standardize on C with some cross platform GUI system like GTK+ and intially development time for every project would be inflated 50 fold over a more intuitive system since training time would be substantial and we'd have to let go a lot of junior developers that never learned how exactly object oriented languages work. Oh, we'd probably need to use Intel's compiler since then full x86 assembler would be supported.
In a company that develops retail products which should be able to run on different platforms, I highly recommend C++ with Qt as it's the only polished environment for end to end product development I've come across in 16 years of large scale retail software development. Of course, getting locked into Nokia might be an issue.
I a company that does no low level development, I would highly recommend C#, Visual Studio, Subversion, and the full kit, or Java with Eclipse and Subversion. It's a judgement call there. C# is generally a better system for Windows application development since the UI system is focussed on Windows... but it works on other platforms. For Java, even if you program with SWT, the look and feel is never quite right on each platform. So it would depend more on wht kind of developers you have and what your platform needs are.
For a company which does nothing but database work, I'd still suggest the previous two options and add either a proper.NET server environment or a proper Java server environmnent. With Java, maybe toss in some cool tech like Google widgets or another "Web 2.0" technology.
I personally would try and get TrollTech Qt or Visual C# into the system since I am a big fan of short time to market and.NET SDK and Qt both are very good for this purpose.
For years I've heard of HIV research focusing on either slowing the spread of HIV or focusing on killing the virus.
This solution instead makes it so the virus will have no effect. I would imagine that a HIV infected victim that has not been vacinated should even be able to receive immunized T-cell injections or even pill supplements.
The approach to this problem just reminds me of the kid that suggested letting the air out of the tires of the bus stuck in the tunnel. Of course it's obviously much more technical to modify T-cells in laboratories, but the beauty of the approach is equally elegant
We should always question science. We should regularly check and recheck our beliefs and gather and collect knowledge into well defined groups and refine the data sets and theories to reflect what we've learned.
If by chance evolution could be 100% wrong (I don't have an alternative, but let's assume that aliens from another universe planted seeds which grew into earth life) we should keep looking for it.
I believe we don't know everything about evolution, or most of these topics these people wish to dispute.
So here's what I hope. I hope some well educated people will in fact choose to take advantage of this to attempt to educate students on newer and more modern theories than those found in 20 year text books. Add DNA to the equation for example.
What's more likely is, Louisianna will actually manage to accomplish the most impressive and unbelievably impossible thing ever. They'll manage to force future generations to be as dumb as they are fat. (look up national health stats at the CDC)
10,000 dom objects with all the attatched properties and such can easily yield half a million pointers. It's not uncommon to have 20-30 pointers per object either, especially when their values are being altered throught scripts. It is extremely usual that the layout data alone per object is quite substantial. Remember, just because a property isn't used doesn't mean that the pointers to their data don't exist, they're just in a null state. Properties aren't efficient to store as hash tables or dictionary lists given the cost of look up. Even efficient algorithms require time to perform a million hash lookups.
Last I checked half a million times 4 does actually add up.
P.S. - I worked on one of the major 4 browsers for 6 years of my life... I also modified the source to at least another two of them and am pretty confident in my understanding of their internal workings.
Images aren't persistant either, they are unloaded or left in their compressed state most of the time. The processor cost of image decompression is negligable so there's no real point in keeping them in memory. The actual cost of doing business on a browser is state representation. The best way to store the state of a page, especially a dynamic one for the purpose of navigating forward and back through history is to leave the trees in tact as long as possible. So, instead of picturing storing 10,000 DOM objects, instead picture storing 10,000 DOM objects per page for 50 pages with 25-50 pointers per object.
Since browsers represent their data as tree structures.... each DOM element is represented by a pointer to a class, every string and token is represented by a pointer to a string (most often a class, but is some browsers, not always), ecmascript objects are very very pointer heavy especially given their bindings into the document model. CSS is a lot of strings and it's a lot of properties which are all pointer based.
So, when you double the length of the pointer from 4 bytes to 8, well guess what....
Actually, the Flash plugin is unloaded when it's not in use. Also Flash appropriately uses the NPAPI malloc and free calls to handle memory management. Browsers aren't allowed to fragment memory, when they do, they absorb all memory, all the time, so even if Flash is extremely inefficient when the plugin is unloaded, it's the nature of the browser to automatically release all memory used by the plugin.
Java is a bit of a different story since the JNI interface for Java tends to prefer to keep Java persistant. So when it uses memory, it tends to keep using it throughout the browser session. It's more efficient for the Java to keep a certail allocation pool available for the lifetime of the session as well. This can add a bunch of memory consumption.
Over the last several years, the electronics program in the high schools here in Oslo, Norway fell apart. My nephew was left stranded without an educated teacher and was forced to instead study TCP/IP networking to complete his high school diploma. The worst part is, he and his classmates were still interested in electronics and needed a tutor. So I setup a study group with zero financial resources.
This was not a problem.
There is little point in electronics kits and trainers these days as the parts they provide lack focus and the process of building the projects is more for fun and less about learning how things work. There is a much better alternative.
Simulators!
You can purchase (or accidentally duplicate) a copy of either Eletronics Workbench from National Instruments or a copy of Proteus VSM and accomplish far more in a shorter time.
A proper adult level lab requires an Oscilloscope, logic probe, function generator, a/d board, multimeter, etc... these tools virtualize both all of these tools and are extremely powerful.
Also, given the cost of aquiring all the components for a proper lab is hard. As a proof of concept, I actually developed an entire 8080 compatible processor at a transistor level with memory and graphical display in Proteus VSM. This project was great since it allowed me to teach, by example all the concepts from power supply design to software development in a simulator at ZERO cost.
I highly recommend ignoring kits and lab tools and focussing instead on the theory which can all be learned visually in these simulators.
When you've learned to design something, you can design your own boards and have them prototyped by cheap services for $20 a pop and solder the parts down. If there's a flaw in the design, figure it out in the simulator. If it requires you to access motors and sensors, then you can get a multimeter and a logic analyzer, but you should be able to fake the problems in the simulators to sort out most problems.
Just read the paper. While their research is entirely sound (no pun intended), the value of their research is pretty limited.
In circumstances like Skype (not RTP), it is possible to talk and text chat at the same time. All of it is encrypted.
The application of this type of stegonographic message is for stored data. But for that, the data would have to be stored. There's just not point in storing a voice conversation as RTP packets on the users' system. In fact, it would be almost ridiculous to store audio in a network session specific payloaded form.
Typically, it would be better to just store the audio as a.wav encapsulated G.711 or G.722.1 file. That would lose the RTP packets altogether while leaving the audio playable. But doing so would probably delete the hidden message.
So far as I can tell, there's no application for using this in the real time context and if you're trying to hide your data, storing it in RTP encapsulated audio packets on disk is pretty silly since it isn't a standard file format.
If you want to hide something, find a way to hide it in a jpeg. png, etc...
I understand your quest:) I have a 6 year old boy and a 4 year old girl and these are things which I have in fact already been working on since... before they were born. When I found out my wife was pregnant, I started building this type of educational environment. Unfortunately, here in Oslo Norway, it's hard for an American to track all those good things down, especially the Estes model rockets (which I did manage after a long time).
Throughout my house, you're likely to encounter the biggest classic erector sets (at least that's what it said on the box).
We have at last count over 25,000 lego bricks because we don't reuse, be just buy more to build new things.
We have Capsula, a large selection of small tools, bits of wood, glues, etc... for constructing small things.
We also, so far as I know have the only functional model rocket "lab" in town. Since rocket engines cost me about $20 each here in Norway, I've taken to shipping them via surface mail (REALLY SLOW) through hazmat shipping lines which actually costs me a bundle, but cuts the cost to about $5 an engine. We buy 4-5 rockets each time we're in America and spread out the task of building them.
We also have robotic components which I tend to either track down through online catalogs or design in solid works and have a local school produce (they make me buy one per student in the class, but only charge for materials).
I also keep a work bench with soldering equipment and a selection of about 2500 different electronic components, including breadboards and such.
For a chemistry set, I've been lucky, my father-in-law was a chemist (old style pharmucist) for nearly 50 years. He donated a tremendous collection of glassware, microscopes and even some controlled chemicals (his license is still valid). For a bunsen burner, I rigged one up, but it appears to function well enough.
I have just begun planning the biology and botony "lab" but since this is not my area of expertise, we're dependant on kits from educational suppliers. If it weren't for wikipedia and a subscription to britanica (it's has a fabulous children's version) I'd be lost in these areas since I don't like leaving questions unanswered for the children.
At the local technical museum, there's a huge selection of science kits, so I tend to purchase one or two each time I'm there. Too bad it's going a little too mainstream now and is carrying $5 crap items which are more likely to sell and less higher end educational components.
I hope the description of my lab helps you to plan yours since I believe that thus-far mine has been extremely sucessful in spurring scientific interest in my children.
Oh.. P.S. don't forget computers, and don't be cheap. I found that providing each child with a relatively high end computer makes them more interested in using them and asking questions about them. Building a computer with a 4-year old was a blast since I let him do most of the work.
Move to a place where earth quakes are a certainty of life and then bitch about how they can't get affordable insurance from a reliable provider on their overpriced home sitting on the fault line?
Go to resturants and look at the menus so they can catalogue the ingrediants available in the kitchen and then order food, not on the menu made of ingredients they're guaranteed not to have?
Bitch about the environment being in such poor repair and bitch about gas prices but go out and buy his and hers hummers?
Take jobs and buy houses that specifically require them to sit for 2 hours in traffic each way and then bitch about how the government is screwing them by making it hard for them to get to work?
People who bitch all day and night about illegal immigrants costing them a fortune, while they sit outside by the pool because Maria and Rosa are busy inside vaccuuming the living room?
And of course, worst of all... people who think that chicken or pineapple are in fact pizza toppings!
OEM versions of course, but still, didn't hesitate a moment to pay the Microsoft tax.
I also don't hesitate to pay for new releases of OS X if I happen to have a Mac around at the time. Haven't bothered in a few years since dealing with OS problems on two platforms (Windows, Solaris) is more than enough for one person.
I would even pay for Linux if I felt the money would get to anyone who deserved it. Instead I contribute directly to projects I feel have earned it.
The only reason I can imagine that we use such UGLY! forms of power which now we're polluting the surface of the ocean with is that we're short on alternative options and we need to try everything we can. But, this much I promise you, windmills are temporary. Maybe they'll be around 10 years, 50 years or 100 years, but they will be replaced with more technologically advanced forms of power eventually.
Personally, I'm looking forward to reactor sized fuel cells. I like the tech since unlike wind power which I believe will negatively impace global wind currents if they become too popular, fuel cells will eventually (initial models, not likely) give back to the environment by helping to sustain fresh water supplies.
So point being, these massive windmill towers that are ugly as sin and should never have been put up need to be recyclable. Since they're only a temporary solution, they shouldn't later be landfill.
So, if I can't use my smart phone to check movie times, or book other recreational services while my kids are spinning around on a ride, my kids miss out.
So how is this a good thing?
P.S. As what I'd like to call an active parent that puts his kids first, what I've learned by watching other parents is, if a parent is ignoring their kid with their phone in their hand, they'll ignore the kid without it.
This is a technique I've always been a fan of. On Linux, you can install a module as a back door and configure it to load, Windows it's exploit a handicap feature. On Mac last time I had to reset the root password, I used a similar approach.
This type of hacking is always nice, but of course this is a hack which requires physical access and can generally be avoided using full disk encryption.
Still nice to see a new, effective way to change the Administrator password, could save me hours screwing with making a new BartPE with the right drivers.
This board is known as the Intel Little Falls motherboard, it's intended for a new device market dubbed Net Top or something stupid by Intel marketing. The intended goal is to produce a low power, sub $200 PC for e-mail and web surfing.
The board should be available June 3rd worldwide from thousands of vendors. I have already placed orders here in Norway for two of them.
Also, the price listed is obscenely high, I've been seeing them typically in the $60-$70 range.
Slashdot Mirror
In societies where women are not even meant to be seen let alone heard can clearly live a far more relaxed live with more women waiteing on the men.
But in the western world woman presume the dominant role within the house (most men beg their wives to let them buy that new PC, TV, etc..) more than one woman would mean more than one dominant control freak in the house.
Polygamy in the west is a terrible idea. I have no frigging idea how to keep up with all the demands of one woman (my wife who's reading this now agrees that I'm clueless and she feels it's for my own good) having a second woman demanding I do this, that, or the other thing would kill me in a year or two.
P.S. - My wife wishes that I relay that men are good for three things. Opening jars, killing bugs and changing tires. She says, the other things she can buy in a store with or without batteries.
I'm not entirely sure if I remember the details correctly, but I'll try to go on memory since researching it should be left to someone else.
Intel licensed the 8086 processor (and 80286 if I recall correctly) to AMD per demand of IBM. Speculatively, I believe that Intel has always considered having a second x86 CPU source as a good thing. It keeps the monopoly monkeys off their backs.
When 386 came around, AMD stayed pretty quiet. But, as you would see from NEC who produced x86 processors with 286 functions (like pusha and popa) Intel started closing up their market. x86 already had established such a strong foothold that Intel became territorial and actually did attempt to monopolize the market by not granting 80386 licenses to anyone else.
AMD responded with the 386dx-40, which for the most part was the start of the megahertz war. Cyrix (spelling), Evergreen and a few others quickly jumped on board and all started releasing 386 clones at an incredible rate.
So far as I know, at this point, NECs response was the v40 which sold primarily in the embedded and Japanese markets.
The 486 generation was a terrible era for system builders. The 486 was still using a very simplistic front side bus and didn't make use of clock multipliers. Motherboards were stills shipping with socketed crystals so that you would modify the system speed by putting in a different crystal. Since there were now so many chipsets to choose from (there must have been 15 or more brands competing) and the chipset manufacturers weren't producing stable reference implementations, motherboards came in only two quality grades... bad and worse.
Now came the absolute worst part. VESA local bus. VESA local bus was pretty much the same thing as ISA in that it connected a periperal board directly to the processor's I/O and memory busses without any logic inbetween. This was a response to the extremely overpriced and overcomplicated EISA bus and the fact that ISA was only 16Mhz x 16-bits. Since the "standard" only allowed for 33Mhz busses, most board manufacters made boards that ran at 33Mhz... maybe 33.1 but certainly not 40 or 50Mhz. So since nearly all chips coming from places other than Intel used the faster bus rates, system stability was getting to be worse than shit.
This era nearly destroyed the x86 world since the concept of name brand motherboards and video cards was only for rich people. Micronics and DFI (the only players at the time) charged $350 for their motherboards while everyone else was asking about $100-$120. I can assure you, having worked in a computer store at the time, the quality difference was worth every penny. But back then, $200 was considered a lot of money for a hobbiest to lay out. The price difference for reliable name brand memory (only Kingston existed back then) was more than double that.
Intel began sueing everyone over the x86 license. In fact, I even liked the idea since nearly half the machines I was shipping out with 40Mhz busses were coming back over and over again with serious failures. The store I worked for tried a niche market which was "reliable clones" where we tried real hard to only use parts that were quality. After losing our asses over it since noone wanted to pay $2000 for a machine when the other stores only charged $900, we dumped that business... and it didn't matter, the 40Mhz FSB motherboard were still coming back broken all the time.
Intels choice to button up the market was ideal since in a way it was non only protecting its own assets and IP, but also trying to take the crappy clone chip makers off the market. And most of them did fall off the market. Many consolidated, many disolved. But in the end it left us with Cyrix and AMD. I think I remember Cyrix being purchased by VIA or someone else.
What's important about this consolidation is that even though Intel squashed a ton of companies, the companies that remained were the companies who had managed to gain a good enough reputation duing the x86 war to make enough money to not only su
I'll conceed on this point to a certain extent. I had thought a little about how I would implement this scheme in software after the point and I'll also respond to your response.
First the response.
Routers are crap for gigabit anyway... they're 10 times worse for 10Ge. That's probably the number one reason why layer-3 switches are far more attractive in backbone environments. In fact, it took me years to find a router that cost less than $10,000 that could in fact handle 100Mbit traffic 24/7 (I mean run at full capacity 24/7) and that router implemented portions of the IP stack in FPGAs.
If you're running a software based router to handle full gigabit traffic, then I'm forced to call you a tree killer (only joking). For a software based router to efficiently handle four gigabit ports at full capacity requires some pretty substantial hardware. Of course the #1 method of making it work better is to use a server Ethernet adapter which offloads at least the checksum calculations from the host machine. Sadly, most high end ethernet adapters stop there since to do much more they'd have to have interact more closely with the TCP/IP stack which generally doesn't work out too well in most OS NIC driver models.
Now that I've attacked your line of thinking, let me make an ammendment to my line of thinking.
I made another post suggesting capping the limit at 9-bits instead of 8-bits. This would work great since it would first of all keep the numbering scheme of the current IPv4 address space, expand the number of user inputable IP addresses by 16 fold, leave a tremendous number of address space in what I'd call the "utility area" or the space where trained (I hope) network professionals would configure large banks of auto configured addresses available (which would use a more complex numbering scheme). And best of all the math would pretty much all remain the same.
The goal of the project isn't to make networking professionals lives easier. In fact, I'd rather make it harder to lay waste to all the guys out there that actually need 8 hours to perform simple router configurations. That's political though. The goal is instead to avoid making a bunch of home users and hobbiests that finally learned to recognize and IP address able to still recognize an IP address.
Imagine this, your dad called you, there's something wrong with his computer and for some reason, before you can log on to fix the system from outside, you need him to reset his router to default settings and then open a port to let you in.
So you pull up the instruction manual to the router after asking him for the model number and eventually after hearing the serial number and manufacturing date, he finally tells you the model. Now, you walk him through the reset process and the process of connecting to the proper wireless network.
Now you want to get him to find out the IP address of the router. So you do an ipconfig, ifconfig, etc... and ask him to tell you the address.
With IPv4 you can simply say "There should be something like a telephone number, but it's separated by dots instead of dashes and there are 4 groups instead of 3." From that point, it should only take two or thress tries to get him to find it.
With IPv6 numbering, you're stuck with a "Ok son, let me get my glasses, you say there are a bunch of numbers with a bunch of colons or possible some colons and some dots, wow there are a lot of numbers here. This is twice as long as the code I had to type in when you walked me through reinstalling Windows XP on my machine".
There's a value to the IPv4 numbering system in that it's short and sweet. IPv6 is too damn long. All our hopes and dreams of IPv6 being a perfect auto-configured world is crap too. So, if it means that software and hardware developers would be forced to optimize for an arithmetic operation as opposed to a binary calculation, as someone who would have to do it, I'd say "No problem, we'll find a way" since in reality it's true, we'd find a way.
P.S. Let me
No, having implemented IP stacks for embedded router stacks, I'm pretty sure I do in fact understand the protocol pretty well.
What I'm pointing out is that the limitation which makes 3 digits in four groups ideal is that people find them easy to remember. By leaving 3digits by four groups standard, we can already suspect that people can continue to both remember them and not find them confusing.
But to soothe your mind, alter the standard to a 9 bit octet (yes I know what octet means, but I don't know the word for a grouping of 9) and choose to max at 0-511 instead. Therefore we can provide a direct decimal to base-2 mapping and still increase the address space by a factor of 16. Either way, it should have a pretty major positive impact.
Though now I'm getting carried away with this line of thinking. It appears that by using this method of thinking we could do something like use a double addressing system.
So we can use 0.0.0.0-511.511.511.511 to isolate the portion of the Internet which should be theoretically identifiable publicly and revert to the more complex IPv6 numbering scheme to address private and utility devices. For example, for each mobile phone device which is ALWAYS dynamically assigned IPs and therefore would not benefit from the simpler form.
I personally liked 0-999 instead of 0-511 since I feel that user inputable IP addresses should be more closely matched to human beings who think in base 10. From an IP stack development perspective, the only complication I see with this method is that IPv4 functioned so nicely since he subnet landed on a base-2 boundary. by placing it on a base-10 boundary instead, it would require an arithmetic comparison vs a binary comparison.
P.S. before you get snooty, at least take the time to read the entire sentence you quoted. I'd imagine that after making a direct reference to 2^10, it would have been obvious that I clearly do understand the importance of the value 255.
When there was no more space to build outward in Manhatten, then solution wasn't to try and produce more land. instead, they made the buildings taller (which worked well until '99)
People have no problems remembering up to four three digit groups. So why not, expand the address space to support 0-999 values instead of just 0-255. Sure, 999 isn't a byte, but it's close enough to 2^10. Sacrificing the remaining 25 values won't hurt much. But more importantly, it would increase the address pool from 4.2 billion (minus invalid values) to 1,000,000,000,000 (a trillion) which still allows something like 200 IP addresses for every person on the planet. And with technology like NAT which should be employed for security purposes should be more than we could ever use.
Not we just need some genious to figure out how best to map that mechanism to the base-2 or IPv6 world
With the upcoming release of the Nehlam platform, we would imagine that this business should become far easier to be a part of, after all, it moves the majority of the complex tasks into the CPU and the chipset is more of a peripheral and power saving feature bus. But the truth is that in upcoming Intel chips it will be necessary to implement a far more complex DMA system and so far as I can tell from the FSB documents I've read (have the real ones) it looks like most of the existing code used for developing chipsets will have to be tossed.
These days, it's pretty obvious that Via won't make a living competing with AMD on sales of AMD chipsets, so there's just not point in bothering.
I appears that to support the Nehlam processor, an entirely new memory architecture will need to be developed to compensate for PCI DMA.
At the moment, from what I understand, if Via were to pursue licensing the Nehlam FSB, pursue implementing the new memory architecture, etc.. they would use a tremendous amount of financial resources that could otherwise be allocated to protecting their currently threatened CPU business. After all, almost every single VIA device on the market has be rereleased using Atom, even if it meant putting a square peg in a round hole.
Via should take whatever money they make from supporting the Intel and AMD platforms and sink it all into making the Via platform something special.
P.S. - In case VIA is reading this, I highly recommend making a VIA optimized compiler based on one of the high performance open source projects. This way companies dependant on CODEC performance could actually tweak their code for Via as well as AMD and Intel as we do now.
Back in the late 80's I worked on System 38 and OS/400 based systems and the hottest topic of the time was graphical user interface design software for mainframe applications. It worked great too. See, mainframe applications don't generally have veyr advanced user interfaces, so to aquire and alter data through the on screen fields through a remote terminal emulation is simplistic in every way.
:)
I believe IBM even sells some software like what I desribed, but that's not the point.
Give me a day or two in front of a Linux box with tn3270 and I'll make a system which will read all the relevant data from the mainframe ISAM to the local system so the values prior to the change can be stored for recovery later on.
Then I'll take another two days to make a program which will open up each record in the database which met the criteria concerning the minimum wage shift based on either name or social security number. Then I'll change the data to minimum wage or to the previous salary level depending on the task.
I would then make multiple test runs to ensure that it works before screwing the entire work force of public California.
Altogether, it would require about a week to do the task and maybe on that many records a few hours to a day to run.
If you want me to do it, I'll do it for $10,000 plus flight, rental car, and 4(+) star accomodations for me and my family.
Or you can probably find some college kid with no experience and no accountability and pay him/her $6.55 an hour for a month or three to do it
Enjoy
P.S. As to the guys all talking about how jobs are hard to get. I've worked at 3 major retail software/hardware vendors as a senior level developer. One thing I can safely say is.. apply for the job even if the skills required are something that interest you. Fact is, 90% of the time, the unrealistic requirements can be overcome if you sell yourself as someone capable of learning the technology. Most companies actually respect initiative over proficiency.
Isn't that the same place that brought us the serial killer birth place capitol of the world, greater Kenosha Wisconsin?
English is a more or less universal language these days... especially in the world of engineers. I worked professionally with Japanese, Chinese, Singaporian, Indian (a few different flavours), German, French, Swedish, Norwegian, Finnish, British (it's a different language in some places almost), Cuban, and more than a few from former soviet countries. Oh.. let's not forget Italians, I can write a book on that experience.
:
The languages I find most useful to attempt are
- German
- Manderine
Let me justify.
French, while an attractive language is of very little use for most people in the computer business since every French person you will need to do business with will speak English. This is true for many other places, but the difference is, unless you feel you can learn to speak French so well you'd be indecipherable to a native, there's a very strong chance they'll just force you to speak English anyway. I think they take so much pride in their language that they consider Americans speaking French to sound like finger nails on chalkboards. I'm sure there are people that would differ with my findings on this topic, but I am a strong believer there's no universal answer to this issue.
Spanish, there's a niche market in computers where Spanish is a useful language, but if you can speak at least one other language and have massaged your brain to support multiple vocabularies, Spanish is a language which can be learned quite quickly on the fly and consuming financial resources to pay for university courses on the topic is probably wasteful.
Japanese, unless you intend to live and work in Japan with strong Japanese work ethic, learning the language could do you more harm than good. Collegues in Japan like to have their privacy when discussing business issues before communicating their findings to vendors or customers. If you understand their language, it may give you an edge in negotiations, but if they learn you know their language, they'll have to delay any decisions until they have a chance to speak privately. So, even if you learn it, it will work better not to use it unless you want to work in their country with their work ethics. Most American's working in Japan still work hard, but are given a little extra leg room since "He doesn't know any better". Learning the language means you've also probably learned some customs and you lose this advantage.
I can descibe the benefits or disadvantages of learning each language, but you get the idea so far. Here's the reason why to learn German and Manderine.
Manderine is obvious as it is the most spoken Chinese language. In the modern world and modern economy it is likely that at some point Chinese companies will begin hiring Americans to bridge gaps. Speaking their language will make you more valuable especially since much of their designs and documents will already be in Chinese. So it's a valuable skill.
German, this is an important one. Germany is only part of the issue. Learning German will buy you respect in Germany since they consider it flattering that you've learned their language. But getting a job in Germany is easy enough with just English. Ther point is that as an American you will learn a root language which is present in English, but the German roots in English are quite weak compared to the Latin roots. By learning German you'll learn the root language to most other European languages that aren't Latin bsaed. This means that you'll find it easier to learn Latin languages since you've molded your mind to support more than one language and you'll be able to learn Germanic languages much easier. This would give you the Scandinavian languages and quite a few eastern block languages.
The benefit to this is that in times (such as current times) you would be able to leave the U.S. and work in a strong European economy, purchase real estate and convert Euro to dollars during a strong European economy. Working in Germany right now as an American is EXTREMELY profitable and should not be ov
Well Swedish chicks speak English just fine. I'm an American in Norway and speak Norwegian and understand Swedish and Danish and they humor me and pretend to understand my Norwegian.
Use English around here, it's a bit of a waste to learn the language unless you're planning on working in a day care or elementary school to hit on the single mothers.
Seriously, depending on what kind of company it is and te focus of the project types.
.NET server environment or a proper Java server environmnent. With Java, maybe toss in some cool tech like Google widgets or another "Web 2.0" technology.
.NET SDK and Qt both are very good for this purpose.
In my company for example, we would have to standardize on C with some cross platform GUI system like GTK+ and intially development time for every project would be inflated 50 fold over a more intuitive system since training time would be substantial and we'd have to let go a lot of junior developers that never learned how exactly object oriented languages work. Oh, we'd probably need to use Intel's compiler since then full x86 assembler would be supported.
In a company that develops retail products which should be able to run on different platforms, I highly recommend C++ with Qt as it's the only polished environment for end to end product development I've come across in 16 years of large scale retail software development. Of course, getting locked into Nokia might be an issue.
I a company that does no low level development, I would highly recommend C#, Visual Studio, Subversion, and the full kit, or Java with Eclipse and Subversion. It's a judgement call there. C# is generally a better system for Windows application development since the UI system is focussed on Windows... but it works on other platforms. For Java, even if you program with SWT, the look and feel is never quite right on each platform. So it would depend more on wht kind of developers you have and what your platform needs are.
For a company which does nothing but database work, I'd still suggest the previous two options and add either a proper
I personally would try and get TrollTech Qt or Visual C# into the system since I am a big fan of short time to market and
I wish you luck whatever you choose.
For years I've heard of HIV research focusing on either slowing the spread of HIV or focusing on killing the virus.
This solution instead makes it so the virus will have no effect. I would imagine that a HIV infected victim that has not been vacinated should even be able to receive immunized T-cell injections or even pill supplements.
The approach to this problem just reminds me of the kid that suggested letting the air out of the tires of the bus stuck in the tunnel. Of course it's obviously much more technical to modify T-cells in laboratories, but the beauty of the approach is equally elegant
I highly suspect it's not.
We should always question science. We should regularly check and recheck our beliefs and gather and collect knowledge into well defined groups and refine the data sets and theories to reflect what we've learned.
If by chance evolution could be 100% wrong (I don't have an alternative, but let's assume that aliens from another universe planted seeds which grew into earth life) we should keep looking for it.
I believe we don't know everything about evolution, or most of these topics these people wish to dispute.
So here's what I hope. I hope some well educated people will in fact choose to take advantage of this to attempt to educate students on newer and more modern theories than those found in 20 year text books. Add DNA to the equation for example.
What's more likely is, Louisianna will actually manage to accomplish the most impressive and unbelievably impossible thing ever. They'll manage to force future generations to be as dumb as they are fat. (look up national health stats at the CDC)
10,000 dom objects with all the attatched properties and such can easily yield half a million pointers. It's not uncommon to have 20-30 pointers per object either, especially when their values are being altered throught scripts. It is extremely usual that the layout data alone per object is quite substantial. Remember, just because a property isn't used doesn't mean that the pointers to their data don't exist, they're just in a null state. Properties aren't efficient to store as hash tables or dictionary lists given the cost of look up. Even efficient algorithms require time to perform a million hash lookups.
... I also modified the source to at least another two of them and am pretty confident in my understanding of their internal workings.
Last I checked half a million times 4 does actually add up.
P.S. - I worked on one of the major 4 browsers for 6 years of my life
Images aren't persistant either, they are unloaded or left in their compressed state most of the time. The processor cost of image decompression is negligable so there's no real point in keeping them in memory. The actual cost of doing business on a browser is state representation. The best way to store the state of a page, especially a dynamic one for the purpose of navigating forward and back through history is to leave the trees in tact as long as possible. So, instead of picturing storing 10,000 DOM objects, instead picture storing 10,000 DOM objects per page for 50 pages with 25-50 pointers per object.
Still think those 4 bytes don't count?
Since browsers represent their data as tree structures.... each DOM element is represented by a pointer to a class, every string and token is represented by a pointer to a string (most often a class, but is some browsers, not always), ecmascript objects are very very pointer heavy especially given their bindings into the document model. CSS is a lot of strings and it's a lot of properties which are all pointer based.
So, when you double the length of the pointer from 4 bytes to 8, well guess what....
Actually, the Flash plugin is unloaded when it's not in use. Also Flash appropriately uses the NPAPI malloc and free calls to handle memory management. Browsers aren't allowed to fragment memory, when they do, they absorb all memory, all the time, so even if Flash is extremely inefficient when the plugin is unloaded, it's the nature of the browser to automatically release all memory used by the plugin.
Java is a bit of a different story since the JNI interface for Java tends to prefer to keep Java persistant. So when it uses memory, it tends to keep using it throughout the browser session. It's more efficient for the Java to keep a certail allocation pool available for the lifetime of the session as well. This can add a bunch of memory consumption.
Over the last several years, the electronics program in the high schools here in Oslo, Norway fell apart. My nephew was left stranded without an educated teacher and was forced to instead study TCP/IP networking to complete his high school diploma. The worst part is, he and his classmates were still interested in electronics and needed a tutor. So I setup a study group with zero financial resources.
This was not a problem.
There is little point in electronics kits and trainers these days as the parts they provide lack focus and the process of building the projects is more for fun and less about learning how things work. There is a much better alternative.
Simulators!
You can purchase (or accidentally duplicate) a copy of either Eletronics Workbench from National Instruments or a copy of Proteus VSM and accomplish far more in a shorter time.
A proper adult level lab requires an Oscilloscope, logic probe, function generator, a/d board, multimeter, etc... these tools virtualize both all of these tools and are extremely powerful.
Also, given the cost of aquiring all the components for a proper lab is hard. As a proof of concept, I actually developed an entire 8080 compatible processor at a transistor level with memory and graphical display in Proteus VSM. This project was great since it allowed me to teach, by example all the concepts from power supply design to software development in a simulator at ZERO cost.
I highly recommend ignoring kits and lab tools and focussing instead on the theory which can all be learned visually in these simulators.
When you've learned to design something, you can design your own boards and have them prototyped by cheap services for $20 a pop and solder the parts down. If there's a flaw in the design, figure it out in the simulator. If it requires you to access motors and sensors, then you can get a multimeter and a logic analyzer, but you should be able to fake the problems in the simulators to sort out most problems.
Good luck with your education.
Just read the paper. While their research is entirely sound (no pun intended), the value of their research is pretty limited.
.wav encapsulated G.711 or G.722.1 file. That would lose the RTP packets altogether while leaving the audio playable. But doing so would probably delete the hidden message.
In circumstances like Skype (not RTP), it is possible to talk and text chat at the same time. All of it is encrypted.
The application of this type of stegonographic message is for stored data. But for that, the data would have to be stored. There's just not point in storing a voice conversation as RTP packets on the users' system. In fact, it would be almost ridiculous to store audio in a network session specific payloaded form.
Typically, it would be better to just store the audio as a
So far as I can tell, there's no application for using this in the real time context and if you're trying to hide your data, storing it in RTP encapsulated audio packets on disk is pretty silly since it isn't a standard file format.
If you want to hide something, find a way to hide it in a jpeg. png, etc...
I understand your quest :) I have a 6 year old boy and a 4 year old girl and these are things which I have in fact already been working on since... before they were born. When I found out my wife was pregnant, I started building this type of educational environment. Unfortunately, here in Oslo Norway, it's hard for an American to track all those good things down, especially the Estes model rockets (which I did manage after a long time).
Throughout my house, you're likely to encounter the biggest classic erector sets (at least that's what it said on the box).
We have at last count over 25,000 lego bricks because we don't reuse, be just buy more to build new things.
We have Capsula, a large selection of small tools, bits of wood, glues, etc... for constructing small things.
We also, so far as I know have the only functional model rocket "lab" in town. Since rocket engines cost me about $20 each here in Norway, I've taken to shipping them via surface mail (REALLY SLOW) through hazmat shipping lines which actually costs me a bundle, but cuts the cost to about $5 an engine. We buy 4-5 rockets each time we're in America and spread out the task of building them.
We also have robotic components which I tend to either track down through online catalogs or design in solid works and have a local school produce (they make me buy one per student in the class, but only charge for materials).
I also keep a work bench with soldering equipment and a selection of about 2500 different electronic components, including breadboards and such.
For a chemistry set, I've been lucky, my father-in-law was a chemist (old style pharmucist) for nearly 50 years. He donated a tremendous collection of glassware, microscopes and even some controlled chemicals (his license is still valid). For a bunsen burner, I rigged one up, but it appears to function well enough.
I have just begun planning the biology and botony "lab" but since this is not my area of expertise, we're dependant on kits from educational suppliers. If it weren't for wikipedia and a subscription to britanica (it's has a fabulous children's version) I'd be lost in these areas since I don't like leaving questions unanswered for the children.
At the local technical museum, there's a huge selection of science kits, so I tend to purchase one or two each time I'm there. Too bad it's going a little too mainstream now and is carrying $5 crap items which are more likely to sell and less higher end educational components.
I hope the description of my lab helps you to plan yours since I believe that thus-far mine has been extremely sucessful in spurring scientific interest in my children.
Oh.. P.S. don't forget computers, and don't be cheap. I found that providing each child with a relatively high end computer makes them more interested in using them and asking questions about them. Building a computer with a 4-year old was a blast since I let him do most of the work.
Move to a place where earth quakes are a certainty of life and then bitch about how they can't get affordable insurance from a reliable provider on their overpriced home sitting on the fault line?
Go to resturants and look at the menus so they can catalogue the ingrediants available in the kitchen and then order food, not on the menu made of ingredients they're guaranteed not to have?
Bitch about the environment being in such poor repair and bitch about gas prices but go out and buy his and hers hummers?
Take jobs and buy houses that specifically require them to sit for 2 hours in traffic each way and then bitch about how the government is screwing them by making it hard for them to get to work?
People who bitch all day and night about illegal immigrants costing them a fortune, while they sit outside by the pool because Maria and Rosa are busy inside vaccuuming the living room?
And of course, worst of all... people who think that chicken or pineapple are in fact pizza toppings!
OEM versions of course, but still, didn't hesitate a moment to pay the Microsoft tax.
I also don't hesitate to pay for new releases of OS X if I happen to have a Mac around at the time. Haven't bothered in a few years since dealing with OS problems on two platforms (Windows, Solaris) is more than enough for one person.
I would even pay for Linux if I felt the money would get to anyone who deserved it. Instead I contribute directly to projects I feel have earned it.
The only reason I can imagine that we use such UGLY! forms of power which now we're polluting the surface of the ocean with is that we're short on alternative options and we need to try everything we can. But, this much I promise you, windmills are temporary. Maybe they'll be around 10 years, 50 years or 100 years, but they will be replaced with more technologically advanced forms of power eventually.
Personally, I'm looking forward to reactor sized fuel cells. I like the tech since unlike wind power which I believe will negatively impace global wind currents if they become too popular, fuel cells will eventually (initial models, not likely) give back to the environment by helping to sustain fresh water supplies.
So point being, these massive windmill towers that are ugly as sin and should never have been put up need to be recyclable. Since they're only a temporary solution, they shouldn't later be landfill.
So, if I can't use my smart phone to check movie times, or book other recreational services while my kids are spinning around on a ride, my kids miss out.
So how is this a good thing?
P.S. As what I'd like to call an active parent that puts his kids first, what I've learned by watching other parents is, if a parent is ignoring their kid with their phone in their hand, they'll ignore the kid without it.
This is a technique I've always been a fan of. On Linux, you can install a module as a back door and configure it to load, Windows it's exploit a handicap feature. On Mac last time I had to reset the root password, I used a similar approach.
This type of hacking is always nice, but of course this is a hack which requires physical access and can generally be avoided using full disk encryption.
Still nice to see a new, effective way to change the Administrator password, could save me hours screwing with making a new BartPE with the right drivers.
This board is known as the Intel Little Falls motherboard, it's intended for a new device market dubbed Net Top or something stupid by Intel marketing. The intended goal is to produce a low power, sub $200 PC for e-mail and web surfing.
The board should be available June 3rd worldwide from thousands of vendors. I have already placed orders here in Norway for two of them.
Also, the price listed is obscenely high, I've been seeing them typically in the $60-$70 range.