Basically it's the same idea, just on crack. It's not a quantum leap in theory - it's just more machine in the NIC than you're used to.
This Killer "NIC" is a 400Mhz computer with a NIC, that fits in a slot. They replace the entire network stack in Windows with the simplest possible stuff, and the Killer does _all_ the work, including extensively queueing, and lots of real-world software exceptions... I suspect a big part of what they do is making sure that when your CPU is bogged it doesn't context switch into dealing with the NIC as often...
If your CPU _ISN'T_ pegged you'll probably see no improvement at all, though.
Dry line DSL is when you get DSL with no phone line.
In recent years (I think in response to some FCC rule, but that's an assumption on my part to explain the sudden change) "dry line" DSL has been available - suddenly from all DSL providers here at the same time, where previously it wasn't - for about $6 more than not dry line with the same service. (It was previously available in some business packages for substantially more)
The previous time I tried to get DSL, these dry lines were very expensive. Much moreso than the monthly fees associated with a landline phone. So the cheapest option for any kind of residential DSL was to get a regular phone line - which HAD to be from SBC - whether you used it or not and no matter who your DSL provider was.
You certainly could've gotten a VoIP service as soon as they were around - and used them for, say, all your outgoing calls to be cheap - but you still had to have a phone line through SBC.
In fact, I think a phone line with SBC and SBC DSL is still the cheapest way to get broadband here by a pretty wide margin. (I think Comcast offers some good deals IF you count an expensive TV package as something you were going to get anyway...) (I haven't compared this personally, because I need a static, serverable IP, and both of those services were out of the question.)
I haven't seen this ad, but seeing it from Comcast wouldn't surprise me in the slightest - they ROUTINELY lie in their advertising. In particular they just make things up about Dish reliability (We've had both Comcast has ALWAYS been much less reliable, AND they routinely screw up scheduling and commercials so the commercial overruns part of the programming) They also always tout how hard it is to install a Dish - when both companies have constantly offered free professional installation to any new customer.
And of course they tout their "better" internet service, when in fact they have the only network LESS reliable than the crappy AT&T DSL here.
In chicagoland our phone (SBC, now AT&T) seems actually worse than cable at the greasing the pockets and buying the regulators bit, but clearly they're both very guilty. One AT&T thing I really love is that they were ordered to let anybody sell phone service and DSL service on their lines... but you can't do BOTH on a line - if you want third party DSL you MUST have AT&T phone. (Excluding dry line services which are thankfully cheaper now, but previously didn't exist as a residential service.)
Here's the point I think the other poster meant: RAID0 is striping. So every other chunk of your data is on each disk - so every file is on every disk. So you are essentially guaranteed that any time 1 disk goes bad your entire array is useless - COMPLETELY useless. Lose 1/4 of the drives, lose 100% of the data.
Unless you have an extreme need for contiguous, single file read/write speed, RAID0 is a poor choice. (For many asychronous reads a bunch of drives with your data randomly split between them is much more efficient in the average case, because seeks are much more costly than most reads. Many RAID1 implementations will choose to read from whichever drive is less utilized, so for pure asynch read-speed RAID1 is often best.)
HOWEVER, if all you have is concatenated drives - where the first part of your virtual drive is on the first physical drive, and the second part on the second one - then you skip the giant reliability disadvantage of RAID0. Essentially you then have a scenario where a single drive failure will most likely only take out its fraction of your data. (After some fun with fsck and some luck on the fs level) Lose 1/4 of your drives, lose 1/4 of your data. I can see being comfortable with this scenario.
Now, if you concat'd your drives using LVM, that's what you've got. Based on your original post, I think you did this. That is not RAID0.
So I think you're ok with having only one copy of the data, you probably have LVM concat, and that's fine.
The OP was trying to point out that for any data you even slightly care about RAID0 is a poor choice compared to meerly concating it.
I'm SURE they meant the address space - they not only pointed out the kernel option possibilities, they also explicitly said a "32bit" machine. A 32 bit address space is exactly 4Gi addresses...
If you need it, being able to use more than 4GB of RAM is the single biggest reason to get a 64bit machine - and shipping from Apple with 8GB stock for only like $10k was one of the great things about the original G5 towers, years ago. (Obviously this was loads cheaper if you bought third party RAM - the vast majority of that price was Apple's RAM markup. But, that was still cheap for a machine actually shipping with that much)
While I definitely believe in for-pay software, this analysis misses out on the really critical sharing effect. That is: Presumably every person who successfully adopts a piece of software increases the chances of more people using it. Especially now (as opposed to 94/95) that effect is huge - people like this guy talk on their blogs about software they love. This is SOMETIMES pay software... but when people get hit with the disabling features, SOME of them pay up and SOME of them simply stop using it.
If you can guarantee 100% of people who might want to use your software will hear about it and evalutate it (might approach being possible in some narrow fields, maybe) then this is unimportant to you.
I'm not saying it should always be free, but I'm saying the analysis is flawed without trying to take this into account. Personally, I think for a large class of inexpensive software I would try very hard to get users permanently addicted and using at least a basic version so they can talk about it.
(Also in my opinion competing with a bunch of other free alternatives - a la Movable Type - is the antiperfect situation for getting a lot of donations. I think the Moveable Type data was seriously flawed for comparing to potentially novel software you might write. )
Re:Flash as an application development platform
on
The Future of Flash
·
· Score: 1
How about this: artpad.art.com (drawing online) or this: www.lunchtimers.com
Or of course this: youtube.com
Or this: Aimchart.com (client side precision, configurable charting, printable, and if you save the app locally, once you have viewed the data it caches it so you can browse offline.) (This is very unfinished)
And by wrong, I mean that humans "dual-class" even if they take more than two eventually - learning one at a time as you suggest. Nonhumans "multi-class" taking on up to three simultaneously. But a nonhuman race that could class in one of those things probably could class in more than one.
Unless, of course, this all got rearranged in AD&D 3. I stopped at 2 and got just a tiny bit of a life. Or at least better RPGs.
Re:Flash as an application development platform
on
The Future of Flash
·
· Score: 1
Let me say that I absolutely agree with you.
_IF_ old standard CSS/HTML/JS is sufficient for your site, you should do that. You shouldn't use Flash.
But that's a really big IF. How do you make an application that involves, say, arbitrary charting on the fly based on user input, in old JS? Make it entirely serverside, you say? And what if they need to be able to rapidly manipulate the views of this data - change which data is displayed, how it's displayed, etc? Put in some newer AJAX stuff and hope it works?
Here's a BIGGER if: What if you want an application they can download to their machine and make changes which will later sync to the server/database when they have internet access - AND you want it to work in a browser? Make TWO applications, you say?
Tell me how that's more professional.
Flash is definitely used in situations where it is overkill. But in situations where Flash is an appropriate application tool it has exactly 2 competitors Java applets and advanced AJAX. In my experience both fail LESS gracefully when the version isn't right, and that happens at least as commonly.
Regarding MVC Depending on the type of application you have sometimes it is appropriate to create an entire MVC application within Flash - as a fully functional OO language it's missing nothing necessary to use the MVC pattern unless you can't concieve of using it right. In the case of an app that's usable offline, it all NEEDS to be there.
This works even if the "Model" Flash is using is only a cached part of all of the data present on the server.
But you're quite right that sometimes your business logic itself should remain safely in the server. If that's true, you can still use the language of your choice as an application server, and Flash is essentially just the View.
Oh, and you can have a bunch of gateway pages, or one gateway page with a bunch of functions - which is quite a bit more common.
Re:Flash as an application development platform
on
The Future of Flash
·
· Score: 1
That's only if the site author didn't choose to include the Javascript that Flash AUTOMATICALLY MAKES FOR YOU. Then it would tell you you didn't have Flash or a new enough Flash and give you a link to download it.
As I've said a couple times - in my opinion if your app doesn't NEED the kind of functionality Flash or a Java applet provide, you should use HTML. But there's a whole class of apps where this is the best answer on the web.
You've got this pretty wrong (I've read your grandchild reply, too; some of this is a reply to that.)
1. The internet promotes communication that is very fast and very cheap. Really, that's all the openness inherently does. In fact, if you GIVE a business your physical mailing address often you will get spam there too. Just not quite as much, and relatively more targeted, because the physical spam costs more to send.
2. No amount of technical 'closedness' or email obfuscation is going to keep some friend from giving away your email address or some business you GIVE it to from sharing it. If this is your problem, you have to train your friends and give a different address to the business. There ARE two technical solutions, both available right now - and people don't use them because getting spam is basically less work than preventing it.
A. Give out unique addresses. For instance, since gmail supports "+" wildcarding you can use "youraccount+Doug@gmail.com" to send and recieve email for Doug. It'll go into your regular inbox, but if Doug gives it away you can change it to "doogie" and then set gmail's filters to block everything going to "+doug"
B. Whitelist all your incoming mail; mail only gets to you if you've already approved the recipient.
3. The fundamental problem is not that the internet is cheap, it's that many friends and many businesses don't respect your privacy. LEGAL changes certainly could change this - strong privacy laws that apply even to companies that do business with offshore companies to do the actual mailing. FINANCIAL changes could change this - if spam cost more to send the cost would be lower, but it's hardly going to be more expensive than physical spam which still exists.
4. If gmail lets the "." be anywhere that's a minor gmail flaw, it has nothing to do with the internet or the nature of mail. In fact, your link says gmail just ignores the dot completely, so essentially it was silly to use it to make a gmail acct.
5. If you get totally random mail that doesn't have your correct address (because the '.' is wrong) you probably have a simple address - perhaps one pulled from a dictionary of names and/or words. If YOU want to make YOUR internet more closed, choose a long, nondictionary, hard to accidentally type address. You will no longer be subject to this kind of attack.
6. Spammers don't get bounced emails nor care about them at all. The vast majority of spam emails are sent with a false return address - the worst spam attacks I've ever gotten were the floods of bounces when someone used one of my addresses as the "from".
7. For the most part you can't track the spam servers either. It either comes from overseas locations with lax laws or it comes from millions of compromised home user machines on the internet. If you want to stop tons of the worst spam, get the home ISPs to disconnect any machines found to be compromised - or even unpatched. (I'm NOT suggesting that they install spyware on your machine; by its very nature any remote exploit is generally detectable from the other end of the network cable. This means you can run your Windows 95 machine if you want, as long as your firewall is good enough. The mean time to a new install of Windows being compromised exposed on the internet is ~20 minutes. It takes longer than that to download the patches.)
If you want to help a little bit, make sure everyone you know patches their systems, uses a firewall, and doesn't use IE. (I prefer they don't use Windows, either, but the other things are more important.)
8. And Web 2.0 doesn't really replace anything or have much to do with email, unless you think you're going to start getting all your messages on MySpace instead of via email (which certainly some people do...) Even Web 1.0 doesn't have much to do with email - the Web and email are two different internet technologies that have some crossover (like gmail)
Web 2.0 is about the content creation moving into the hands of the users -
Linux version: 7 Newest Compiler version: 8
on
The Future of Flash
·
· Score: 1
Version 7 is a couple years old, and version 9 of the player is definitely out. However, the Flash IDE is only on version 8 - so _nothing_ in production requires player 9 yet.
verison 7 of the IDE was "Flash MX 2004" - because it came out in 2004. version 8 of the IDE only came out in 2006. So right now Linux can't play things created with the 2006 version that didn't develop with any backward compatibility.
Re:Flash as an application development platform
on
The Future of Flash
·
· Score: 1
Flash is a full-fledged application environment; only Java applets are a competitor for this and in my opinion they are not as good.
Flash has extensive and powerful accessibility controls for the visually impaired if the developer bothers to turn them on. Developing Flash content for the visually impaired is simply not a problem if you care about doing it.
Flash has possibly the highest installed userbase of any single software on the web - more than IE.
Re:Flash as an application development platform
on
The Future of Flash
·
· Score: 1
As a professional developer who does a lot of work in Actionscript/Flash/Flex, I want to weigh in on a few points I've seen here:
I personally think significant Flash intros are deplorable, for the same reason I hate ever going to just about any page on MySpace. The basic reason is that I don't like people or sites that care more about proving they are "cool" to me than about providing information to me. This is particularly true of useless intros, which as a potentially-repeat visitor I have to see every time to use your site - making your site much less useful.
That said, Flash - and even moreso Flex - is an excellent web application development platform. An application that needs only a simple interface (like Google) should clearly not be Flash, but Flash can make an awesome database driven web application that is tolerant of intermittent disconnection to the server and can work with or without internet at all; it just works on a variety of platforms on and offline. AND it can do a wide array of media and server interactions along the way.
Actionscript (the language behind both) is a complete OOP language with a lot of syntactic similarity to Javascript, but with some very important and powerful Flash additions.
Flex is a really awesome server technology for generating swfs from MXML files; it's essentially exactly what you want from Flash if you're a programmer and not a designer. Flex 2.0 is supposed to have a less expensive standalone compiler version.
Flash is a Java applet killer, but it is NOT a Java killer. If you don't care about running in a browser and are making an application the runs on platforms with a JRE installed, these more traditional applications can certainly do some things Flash can't. If I wanted, say, extensive file or hardware management I would not use Flash.
Other children of your post have good points about the effects being concentrated in certain areas. But even if they weren't right, your math is irrelevant. It's not the volume of water compared to the volume of contaminated water, it's about the concentration of contaminents. No number that only talks about the volume of water itself matters a bit. I don't have appropriate numbers to recreate your math, but I'll just give an example:
By your numbers: The Gulf of Mexico, is about 7800 k cu km. The Mississippi releases about 1.1 cu km/day, or about 0.4 k cu km/year.
Let's assume that for a certain kind of agricultural contaminant 1 ppm is normally present in average Gulf water. This is an arbitrary number, but the math is all proportional for any amount the actual value is. (Which almost certainly varies by depth and location) and that a 25% change can be tolerated by the ecosystem (which is a HUGE change; imagine a 25% change in the amount of available oxygen in the air.) If that's true, Mississippi water having an average concentration of 4900 ppm would do the trick in a single year to wipe out the ecosystem.
If 5% is enough to destabilize the ecosystem, you only need 1000 ppm. If the effects are cumulative and you take a decade it would only be 490ppm (25%) or 100ppm (5%) Of course the article says that we've had about 4 decades of intensive fertilization...
And of course, that is making the wildly inaccurate assumption that these fertilizers are equally spread across all the volume in the Gulf. Let's assume that 5% of the Gulf water gets 90% of the Mississippi contaminents. I think this is probably extremely conservative, since they'll tend to settle in the delta which is relatively shallow, and the greater part of the volume comes from the depths.
For a cumulative 4 decades of buildup, now the Mississippi only needs to have 6.8 ppm for a 25% change or 1.36 (!) ppm to get a 5% change (compared to an oceans's natural value of 1ppm)
I would go one step further. Let's say that you figure out in some unlikely scientifically convincing way that 25% is an absolutely safe value for this change and that the Miss is only at perhaps 3 ppm, not 6.8. So the average of our 5% of the waters are safe. But what about the worst-off 2%? 1%? 0.1%? Since clearly it's NOT even distributed, as you move into smaller and smaller worse-off areas, you will definitely exceed any threshold you set if you look at a small enough area.
And even the 0.1% by VOLUME is 10s of thousands of square km of extremely eco-rich shallows.
[I have left out dividing by the amount of water added from the Mississippi, because as you point out the volume of water is relatively small. I've left out water interchanging outside the Gulf. I've left out whatever contaminent level was naturally present in the Miss; this is above that amount.]
I can't speak to that exact bridge, but I certainly would expect some parts to need to be replaced in less than 70 years - and the surface probably more often than that. Possibly a major replacement is needed in 70 years, but replacing that part might only cost a fraction of the entire bridge.
(For instance, all the underwater foundations can be quite involved and expensive, too.)
this is partially a reply to a gp post, but in my opinion:
1. I agree that _obscure_ degrees are great, and often great for getting jobs in or out of that obscure category. It both makes you memorable and increases the chances that you learned something hard.
2. "liberal" and particularly "artistic" degrees are generally worse, especially if they are from anything less than an extremely prestigious school. (Note that I said _degrees_ not _people_, _educations_ or _careers_)
The basic reason is essentially because the more artistic the topic is the harder it is to evaluate whether someone is really good at it - or at least there are fewer points where you can evaluate them in an absolute sense. This isn't primarily about whether the task is easy, it's primarily about whether the school can reliably tell when they should flunk you. (I further believe that many great artists would flunk out of art school in their own time - essentially because art is impossible to absolutely judge.)
I can put this another way: If your classmate couldn't flunk out of school despite being an idiot and not working, your degree itself can't be worth that much - even if YOU actually did a lot of work and learned a lot (a school's value at education and at evaluation often aren't highly correlated)
At an extremely prestigious (in whichever field) school they can overcome this to some degree by having uniformly world-class teachers evaluating you and by also setting a uniformly high standard for evaluation. In general this is not impossible, but it's certainly more rare.
If your goal is to maximize your degree, you want it to be provably difficult. In my opinion Japanese (and indeed most intensive foreign language programs) definitely qualifies. To get a certain film degree (might be only for a Masters) from the University of Chicago requires you to be at least bilingual and evaluate a number of films in their original language. That's a hardcore film degree, from a hardcore school.
Wow, I guess I have a lot to say on this subject. Disclaimer: As probably most/. posts, I am not qualified in any way to make the following statements. Do your own due diligence and check everything I say.
1. I agree that living frugally is paramount. Living frugally trumps the advantages of buying a house (unless you're going to put a lot of labor into fixing up the house) This means if you can find a house as small/cheap(#2) as your apartment or find enough roommates to make it so then it's often a great idea, but unless you can count on lots of long-term appreciation don't buy a bigger house than you need.
2. However, if you CAN find a cheap enough place, note that: the principle part of your payment you're investing in the long-term, so don't compare that part to your mortgage and you get a tax deduction for the entire amount of mortgage interest, so reduce that part by your taxes (and you won't pay gains on the first $250,000 of profit, either).
So my rule of thumb is that if your 30-year mortgage, before tax advantages is only 25% more than your rent it's probably a good deal. This does not assume any significant appreciation, but it doesn't apply in a town that's dying or if you're in a coastal city right now and there really is a huge housing crash.
Also if you HAVE any equity in your house when you buy it and have good credit you can lock-in a home equity line of credit, not borrow from it, and have a guaranteed credit reserve at minimal or no cost.
3. I'm going to arbitrarily define a long-term investment as one where you can reasonably guarantee you won't need the money for 5 years. Generally I wouldn't get into the stock market without that kind of horizon, which I don't think you have now. Because a volatile investment can go down with little reason, and often if it goes down without a great reason the right answer is to hang onto it. So don't do this unless you're sure you have enough of a liquid reserve and then you have maybe twice that:)
4. However, if you want the short version of doing pretty well with relatively low risk in stocks go to fool.com If you want the shorter version: buy S&P500 and Russell index funds. Generally avoid most mutual funds.
5. There are three basic strategies I'd address to having available, relatively short-term money. How much of it does depend on whether you have the bank of Mom and Dad for a really big emergency... And really it's a tiering process... you should have SOME very liquid cash, some that's a little less liquid, etc.
One option is certainly a money market. Another is a short-term CD. You might look into some treasury securities.
6. A technique for investing in vehicles that have a slightly longer term than you like is to invest, say 1/4 of your cash in a 3 month CD, then in a month invest another quarter and in two months another. (Leaving 1/4 as cash) Then everytime a CD matures you can roll it into another (many banks will do this automatically) but at any given moment you have to wait no longer than 1 month to get your money.
7. Roth IRAs. If your cash now is from actual earned (taxable) income, you are probably eligible for a Roth. And if you are, you should be putting money in it up to your income or the Roth cap. Things to know:
A Roth is a tax designation, NOT a vehicle. So you can have a "Roth savings account" or a "Roth CD" (but you can't mix it with your other money, of course)
Because you contributed after-tax money to the Roth, you can generally withdraw anything up to the entire amount you contributed without penalty*. (You can't ever withdraw the INTEREST until your retire, though) This makes a Roth unique among retirement tax options in that it can double as a liquid reserve (if it's in a liquid vehicle)
*(You still don't especially WANT to withdraw it, because what you can't do is ever put that money back into the Roth. If you withdraw it you are essentially in the same position as having never put it in, but because there are caps on per-year Roth contributions there's a big incentive to start early and keep contributing.)
To my knowledge - which is admittedly a year or so old - basically there are three relevant points.
1. Most people have 3 color receptors that they actually use, while some are colorblind to varying degrees including a relatively high number are red-green colorblind having effectively one RG and one B receptor. HOWEVER, where (what wavelength) the "R" "G" and "B" receptors is is NOT exactly the same for each person. So it is very possible that a perfect match for one person is not a perfect match for another especially for colors that are a complex mixture of wavelengths (eg most real-life pigments in sunlight) Note that generally matching the amount of the same pigment should generally be very, very close - to demonstrate this effect you mostly need to be combining very different wavelengths that "should" be the same added together.
The take-home geek message is that you can use an RGB monitor to match every color you can see - IF the monitor's RGB match yours. Otherwise it's not perfect. (Also see point 3)
Have two receptors very close together eventually becomes indistinguishable from just having one as they approach being in the same spot.
2. Some people are known as "tetrachromats" All examples I've heard about have been the mothers of red-green colorblind men. Essentially they have an extra receptor between R & G. This means that they can determine that two colors don't match in situations where everyone with three receptors would think they matched.
3. Apparently we may also have a 4th (or 5th, depending on pt 2) receptor in the ultraviolet range. However, most of the light in this range is blocked by the alchohol in our eye fluids, so this receptor is mostly pretty useless. However, this doesn't mean we don't see SOME color with this receptor right at the edge where it's not blocked by the alchohol - it's just not a very large part of our sight.
These colors definitely don't exist in monitors, which I personally and nonscientifically think is why I love staring at the LED on a PS2.
And don't redistribute. As a software author, beyond it being illegal I consider it to be quite disrespectful to the authors if you were to redistribute these files in violation of their license. On the other hand, it certainly seems like a good idea to make a personal copy at this moment, and that certainly IS in accordance with their license.
GP makes sense, but is not right. Note: I'm not an authority on the Shuttle program by any means, but this is just basic science. What I'm saying below is even more true for the hydrogen, but I used the oxygen as an example. There are two basic reasons it would be quite cold.
1. To make Liquid Oxygen at room temperature does require extremely low temperatures. But they aren't keeping it THAT cold, because it would be prohibitively difficult.
It doesn't need to be that cold, because under pressure the temperature where it stays liquid goes up. You can keep a strong tank of O at room temperature and it will still be liquid. So they likely keep it cold to reduce the pressures they have to maintain somewhat, but it's not required.
2. Whenever you reduce the pressure on a gas (or especially when a liquid becomes a gas) this is a very endothermic (heat absorbing) process. So a home oxygen tank IN USE is cold, but one being stored isn't. (Random metal at room temp will FEEL cold, but the tank isn't _especially_ cold)
When you pour the liquid oxygen into a large temperature tank - even if you do it really fast and in a sealed way - any room left in the tank instantly becomes filled with much higher-pressure gaseous Oxygen. This expansion makes it very cold. And there is a lot of room in the tank when you START filling, even if by the end you filled it all.
Slashdot Mirror
Basically it's the same idea, just on crack. It's not a quantum leap in theory - it's just more machine in the NIC than you're used to.
This Killer "NIC" is a 400Mhz computer with a NIC, that fits in a slot. They replace the entire network stack in Windows with the simplest possible stuff, and the Killer does _all_ the work, including extensively queueing, and lots of real-world software exceptions... I suspect a big part of what they do is making sure that when your CPU is bogged it doesn't context switch into dealing with the NIC as often...
If your CPU _ISN'T_ pegged you'll probably see no improvement at all, though.
I still have at least one old (486 era) box - I remember it had an amazing amount of RAM, but at this point I have no idea how much that was :)
Sad.
Dry line DSL is when you get DSL with no phone line.
In recent years (I think in response to some FCC rule, but that's an assumption on my part to explain the sudden change) "dry line" DSL has been available - suddenly from all DSL providers here at the same time, where previously it wasn't - for about $6 more than not dry line with the same service. (It was previously available in some business packages for substantially more)
The previous time I tried to get DSL, these dry lines were very expensive. Much moreso than the monthly fees associated with a landline phone. So the cheapest option for any kind of residential DSL was to get a regular phone line - which HAD to be from SBC - whether you used it or not and no matter who your DSL provider was.
You certainly could've gotten a VoIP service as soon as they were around - and used them for, say, all your outgoing calls to be cheap - but you still had to have a phone line through SBC.
In fact, I think a phone line with SBC and SBC DSL is still the cheapest way to get broadband here by a pretty wide margin. (I think Comcast offers some good deals IF you count an expensive TV package as something you were going to get anyway...) (I haven't compared this personally, because I need a static, serverable IP, and both of those services were out of the question.)
I haven't seen this ad, but seeing it from Comcast wouldn't surprise me in the slightest - they ROUTINELY lie in their advertising. In particular they just make things up about Dish reliability (We've had both Comcast has ALWAYS been much less reliable, AND they routinely screw up scheduling and commercials so the commercial overruns part of the programming) They also always tout how hard it is to install a Dish - when both companies have constantly offered free professional installation to any new customer.
And of course they tout their "better" internet service, when in fact they have the only network LESS reliable than the crappy AT&T DSL here.
In chicagoland our phone (SBC, now AT&T) seems actually worse than cable at the greasing the pockets and buying the regulators bit, but clearly they're both very guilty. One AT&T thing I really love is that they were ordered to let anybody sell phone service and DSL service on their lines... but you can't do BOTH on a line - if you want third party DSL you MUST have AT&T phone. (Excluding dry line services which are thankfully cheaper now, but previously didn't exist as a residential service.)
Here's the point I think the other poster meant:
RAID0 is striping. So every other chunk of your data is on each disk - so every file is on every disk. So you are essentially guaranteed that any time 1 disk goes bad your entire array is useless - COMPLETELY useless. Lose 1/4 of the drives, lose 100% of the data.
Unless you have an extreme need for contiguous, single file read/write speed, RAID0 is a poor choice. (For many asychronous reads a bunch of drives with your data randomly split between them is much more efficient in the average case, because seeks are much more costly than most reads. Many RAID1 implementations will choose to read from whichever drive is less utilized, so for pure asynch read-speed RAID1 is often best.)
HOWEVER, if all you have is concatenated drives - where the first part of your virtual drive is on the first physical drive, and the second part on the second one - then you skip the giant reliability disadvantage of RAID0. Essentially you then have a scenario where a single drive failure will most likely only take out its fraction of your data. (After some fun with fsck and some luck on the fs level) Lose 1/4 of your drives, lose 1/4 of your data. I can see being comfortable with this scenario.
Now, if you concat'd your drives using LVM, that's what you've got. Based on your original post, I think you did this. That is not RAID0.
So I think you're ok with having only one copy of the data, you probably have LVM concat, and that's fine.
The OP was trying to point out that for any data you even slightly care about RAID0 is a poor choice compared to meerly concating it.
I'm SURE they meant the address space - they not only pointed out the kernel option possibilities, they also explicitly said a "32bit" machine. A 32 bit address space is exactly 4Gi addresses...
If you need it, being able to use more than 4GB of RAM is the single biggest reason to get a 64bit machine - and shipping from Apple with 8GB stock for only like $10k was one of the great things about the original G5 towers, years ago. (Obviously this was loads cheaper if you bought third party RAM - the vast majority of that price was Apple's RAM markup. But, that was still cheap for a machine actually shipping with that much)
http://www.latimes.com/news/local/oceans/la-oceans -series,0,7842752.special
While I definitely believe in for-pay software, this analysis misses out on the really critical sharing effect. That is: Presumably every person who successfully adopts a piece of software increases the chances of more people using it. Especially now (as opposed to 94/95) that effect is huge - people like this guy talk on their blogs about software they love. This is SOMETIMES pay software... but when people get hit with the disabling features, SOME of them pay up and SOME of them simply stop using it.
If you can guarantee 100% of people who might want to use your software will hear about it and evalutate it (might approach being possible in some narrow fields, maybe) then this is unimportant to you.
I'm not saying it should always be free, but I'm saying the analysis is flawed without trying to take this into account. Personally, I think for a large class of inexpensive software I would try very hard to get users permanently addicted and using at least a basic version so they can talk about it.
(Also in my opinion competing with a bunch of other free alternatives - a la Movable Type - is the antiperfect situation for getting a lot of donations. I think the Moveable Type data was seriously flawed for comparing to potentially novel software you might write. )
How about this:
artpad.art.com (drawing online) or this:
www.lunchtimers.com
Or of course this:
youtube.com
Or this:
Aimchart.com (client side precision, configurable charting, printable, and if you save the app locally, once you have viewed the data it caches it so you can browse offline.) (This is very unfinished)
And by wrong, I mean that humans "dual-class" even if they take more than two eventually - learning one at a time as you suggest. Nonhumans "multi-class" taking on up to three simultaneously. But a nonhuman race that could class in one of those things probably could class in more than one.
Unless, of course, this all got rearranged in AD&D 3. I stopped at 2 and got just a tiny bit of a life. Or at least better RPGs.
Let me say that I absolutely agree with you.
_IF_ old standard CSS/HTML/JS is sufficient for your site, you should do that. You shouldn't use Flash.
But that's a really big IF. How do you make an application that involves, say, arbitrary charting on the fly based on user input, in old JS? Make it entirely serverside, you say? And what if they need to be able to rapidly manipulate the views of this data - change which data is displayed, how it's displayed, etc? Put in some newer AJAX stuff and hope it works?
Here's a BIGGER if: What if you want an application they can download to their machine and make changes which will later sync to the server/database when they have internet access - AND you want it to work in a browser? Make TWO applications, you say?
Tell me how that's more professional.
Flash is definitely used in situations where it is overkill. But in situations where Flash is an appropriate application tool it has exactly 2 competitors Java applets and advanced AJAX. In my experience both fail LESS gracefully when the version isn't right, and that happens at least as commonly.
Regarding MVC
Depending on the type of application you have sometimes it is appropriate to create an entire MVC application within Flash - as a fully functional OO language it's missing nothing necessary to use the MVC pattern unless you can't concieve of using it right. In the case of an app that's usable offline, it all NEEDS to be there.
This works even if the "Model" Flash is using is only a cached part of all of the data present on the server.
But you're quite right that sometimes your business logic itself should remain safely in the server. If that's true, you can still use the language of your choice as an application server, and Flash is essentially just the View.
Oh, and you can have a bunch of gateway pages, or one gateway page with a bunch of functions - which is quite a bit more common.
That's only if the site author didn't choose to include the Javascript that Flash AUTOMATICALLY MAKES FOR YOU. Then it would tell you you didn't have Flash or a new enough Flash and give you a link to download it.
As I've said a couple times - in my opinion if your app doesn't NEED the kind of functionality Flash or a Java applet provide, you should use HTML. But there's a whole class of apps where this is the best answer on the web.
You've got it backwards - the cars are dangerous to pedestrians, themselves, and other cars. You need to stop everyone from DRIVING to make it safer.
You've got this pretty wrong (I've read your grandchild reply, too; some of this is a reply to that.)
1. The internet promotes communication that is very fast and very cheap. Really, that's all the openness inherently does. In fact, if you GIVE a business your physical mailing address often you will get spam there too. Just not quite as much, and relatively more targeted, because the physical spam costs more to send.
2. No amount of technical 'closedness' or email obfuscation is going to keep some friend from giving away your email address or some business you GIVE it to from sharing it. If this is your problem, you have to train your friends and give a different address to the business. There ARE two technical solutions, both available right now - and people don't use them because getting spam is basically less work than preventing it.
A. Give out unique addresses. For instance, since gmail supports "+" wildcarding you can use "youraccount+Doug@gmail.com" to send and recieve email for Doug. It'll go into your regular inbox, but if Doug gives it away you can change it to "doogie" and then set gmail's filters to block everything going to "+doug"
B. Whitelist all your incoming mail; mail only gets to you if you've already approved the recipient.
3. The fundamental problem is not that the internet is cheap, it's that many friends and many businesses don't respect your privacy. LEGAL changes certainly could change this - strong privacy laws that apply even to companies that do business with offshore companies to do the actual mailing. FINANCIAL changes could change this - if spam cost more to send the cost would be lower, but it's hardly going to be more expensive than physical spam which still exists.
4. If gmail lets the "." be anywhere that's a minor gmail flaw, it has nothing to do with the internet or the nature of mail. In fact, your link says gmail just ignores the dot completely, so essentially it was silly to use it to make a gmail acct.
5. If you get totally random mail that doesn't have your correct address (because the '.' is wrong) you probably have a simple address - perhaps one pulled from a dictionary of names and/or words. If YOU want to make YOUR internet more closed, choose a long, nondictionary, hard to accidentally type address. You will no longer be subject to this kind of attack.
6. Spammers don't get bounced emails nor care about them at all. The vast majority of spam emails are sent with a false return address - the worst spam attacks I've ever gotten were the floods of bounces when someone used one of my addresses as the "from".
7. For the most part you can't track the spam servers either. It either comes from overseas locations with lax laws or it comes from millions of compromised home user machines on the internet. If you want to stop tons of the worst spam, get the home ISPs to disconnect any machines found to be compromised - or even unpatched. (I'm NOT suggesting that they install spyware on your machine; by its very nature any remote exploit is generally detectable from the other end of the network cable. This means you can run your Windows 95 machine if you want, as long as your firewall is good enough. The mean time to a new install of Windows being compromised exposed on the internet is ~20 minutes. It takes longer than that to download the patches.)
If you want to help a little bit, make sure everyone you know patches their systems, uses a firewall, and doesn't use IE. (I prefer they don't use Windows, either, but the other things are more important.)
8. And Web 2.0 doesn't really replace anything or have much to do with email, unless you think you're going to start getting all your messages on MySpace instead of via email (which certainly some people do...) Even Web 1.0 doesn't have much to do with email - the Web and email are two different internet technologies that have some crossover (like gmail)
Web 2.0 is about the content creation moving into the hands of the users -
Version 7 is a couple years old, and version 9 of the player is definitely out. However, the Flash IDE is only on version 8 - so _nothing_ in production requires player 9 yet.
verison 7 of the IDE was "Flash MX 2004" - because it came out in 2004. version 8 of the IDE only came out in 2006. So right now Linux can't play things created with the 2006 version that didn't develop with any backward compatibility.
Flash is a full-fledged application environment; only Java applets are a competitor for this and in my opinion they are not as good.
Flash has extensive and powerful accessibility controls for the visually impaired if the developer bothers to turn them on. Developing Flash content for the visually impaired is simply not a problem if you care about doing it.
Flash has possibly the highest installed userbase of any single software on the web - more than IE.
As a professional developer who does a lot of work in Actionscript/Flash/Flex, I want to weigh in on a few points I've seen here:
I personally think significant Flash intros are deplorable, for the same reason I hate ever going to just about any page on MySpace. The basic reason is that I don't like people or sites that care more about proving they are "cool" to me than about providing information to me. This is particularly true of useless intros, which as a potentially-repeat visitor I have to see every time to use your site - making your site much less useful.
That said, Flash - and even moreso Flex - is an excellent web application development platform. An application that needs only a simple interface (like Google) should clearly not be Flash, but Flash can make an awesome database driven web application that is tolerant of intermittent disconnection to the server and can work with or without internet at all; it just works on a variety of platforms on and offline. AND it can do a wide array of media and server interactions along the way.
Actionscript (the language behind both) is a complete OOP language with a lot of syntactic similarity to Javascript, but with some very important and powerful Flash additions.
Flex is a really awesome server technology for generating swfs from MXML files; it's essentially exactly what you want from Flash if you're a programmer and not a designer. Flex 2.0 is supposed to have a less expensive standalone compiler version.
Flash is a Java applet killer, but it is NOT a Java killer. If you don't care about running in a browser and are making an application the runs on platforms with a JRE installed, these more traditional applications can certainly do some things Flash can't. If I wanted, say, extensive file or hardware management I would not use Flash.
Other children of your post have good points about the effects being concentrated in certain areas. But even if they weren't right, your math is irrelevant. It's not the volume of water compared to the volume of contaminated water, it's about the concentration of contaminents. No number that only talks about the volume of water itself matters a bit. I don't have appropriate numbers to recreate your math, but I'll just give an example:
By your numbers: The Gulf of Mexico, is about 7800 k cu km. The Mississippi releases about 1.1 cu km/day, or about 0.4 k cu km/year.
Let's assume that for a certain kind of agricultural contaminant 1 ppm is normally present in average Gulf water. This is an arbitrary number, but the math is all proportional for any amount the actual value is. (Which almost certainly varies by depth and location) and that a 25% change can be tolerated by the ecosystem (which is a HUGE change; imagine a 25% change in the amount of available oxygen in the air.) If that's true, Mississippi water having an average concentration of 4900 ppm would do the trick in a single year to wipe out the ecosystem.
If 5% is enough to destabilize the ecosystem, you only need 1000 ppm. If the effects are cumulative and you take a decade it would only be 490ppm (25%) or 100ppm (5%) Of course the article says that we've had about 4 decades of intensive fertilization...
And of course, that is making the wildly inaccurate assumption that these fertilizers are equally spread across all the volume in the Gulf. Let's assume that 5% of the Gulf water gets 90% of the Mississippi contaminents. I think this is probably extremely conservative, since they'll tend to settle in the delta which is relatively shallow, and the greater part of the volume comes from the depths.
For a cumulative 4 decades of buildup, now the Mississippi only needs to have 6.8 ppm for a 25% change or 1.36 (!) ppm to get a 5% change (compared to an oceans's natural value of 1ppm)
I would go one step further. Let's say that you figure out in some unlikely scientifically convincing way that 25% is an absolutely safe value for this change and that the Miss is only at perhaps 3 ppm, not 6.8. So the average of our 5% of the waters are safe. But what about the worst-off 2%? 1%? 0.1%? Since clearly it's NOT even distributed, as you move into smaller and smaller worse-off areas, you will definitely exceed any threshold you set if you look at a small enough area.
And even the 0.1% by VOLUME is 10s of thousands of square km of extremely eco-rich shallows.
[I have left out dividing by the amount of water added from the Mississippi, because as you point out the volume of water is relatively small. I've left out water interchanging outside the Gulf. I've left out whatever contaminent level was naturally present in the Miss; this is above that amount.]
I can't speak to that exact bridge, but I certainly would expect some parts to need to be replaced in less than 70 years - and the surface probably more often than that. Possibly a major replacement is needed in 70 years, but replacing that part might only cost a fraction of the entire bridge.
(For instance, all the underwater foundations can be quite involved and expensive, too.)
Generally I'm pretty fond of a roll-your-own multi-site online-over-ssh backup solution.
But if you weren't going to do that, Amazon offers:
15c / GB/month + 20c / GB xfer
Some middlewares already exist, like Jungledisk.
this is partially a reply to a gp post, but in my opinion:
1. I agree that _obscure_ degrees are great, and often great for getting jobs in or out of that obscure category. It both makes you memorable and increases the chances that you learned something hard.
2. "liberal" and particularly "artistic" degrees are generally worse, especially if they are from anything less than an extremely prestigious school. (Note that I said _degrees_ not _people_, _educations_ or _careers_)
The basic reason is essentially because the more artistic the topic is the harder it is to evaluate whether someone is really good at it - or at least there are fewer points where you can evaluate them in an absolute sense. This isn't primarily about whether the task is easy, it's primarily about whether the school can reliably tell when they should flunk you. (I further believe that many great artists would flunk out of art school in their own time - essentially because art is impossible to absolutely judge.)
I can put this another way: If your classmate couldn't flunk out of school despite being an idiot and not working, your degree itself can't be worth that much - even if YOU actually did a lot of work and learned a lot (a school's value at education and at evaluation often aren't highly correlated)
At an extremely prestigious (in whichever field) school they can overcome this to some degree by having uniformly world-class teachers evaluating you and by also setting a uniformly high standard for evaluation. In general this is not impossible, but it's certainly more rare.
If your goal is to maximize your degree, you want it to be provably difficult. In my opinion Japanese (and indeed most intensive foreign language programs) definitely qualifies. To get a certain film degree (might be only for a Masters) from the University of Chicago requires you to be at least bilingual and evaluate a number of films in their original language. That's a hardcore film degree, from a hardcore school.
Wow, I guess I have a lot to say on this subject. Disclaimer: As probably most /. posts, I am not qualified in any way to make the following statements. Do your own due diligence and check everything I say.
:)
1. I agree that living frugally is paramount. Living frugally trumps the advantages of buying a house (unless you're going to put a lot of labor into fixing up the house) This means if you can find a house as small/cheap(#2) as your apartment or find enough roommates to make it so then it's often a great idea, but unless you can count on lots of long-term appreciation don't buy a bigger house than you need.
2. However, if you CAN find a cheap enough place, note that: the principle part of your payment you're investing in the long-term, so don't compare that part to your mortgage and you get a tax deduction for the entire amount of mortgage interest, so reduce that part by your taxes (and you won't pay gains on the first $250,000 of profit, either).
So my rule of thumb is that if your 30-year mortgage, before tax advantages is only 25% more than your rent it's probably a good deal. This does not assume any significant appreciation, but it doesn't apply in a town that's dying or if you're in a coastal city right now and there really is a huge housing crash.
Also if you HAVE any equity in your house when you buy it and have good credit you can lock-in a home equity line of credit, not borrow from it, and have a guaranteed credit reserve at minimal or no cost.
3. I'm going to arbitrarily define a long-term investment as one where you can reasonably guarantee you won't need the money for 5 years. Generally I wouldn't get into the stock market without that kind of horizon, which I don't think you have now. Because a volatile investment can go down with little reason, and often if it goes down without a great reason the right answer is to hang onto it. So don't do this unless you're sure you have enough of a liquid reserve and then you have maybe twice that
4. However, if you want the short version of doing pretty well with relatively low risk in stocks go to fool.com If you want the shorter version: buy S&P500 and Russell index funds. Generally avoid most mutual funds.
5. There are three basic strategies I'd address to having available, relatively short-term money. How much of it does depend on whether you have the bank of Mom and Dad for a really big emergency...
And really it's a tiering process... you should have SOME very liquid cash, some that's a little less liquid, etc.
One option is certainly a money market. Another is a short-term CD. You might look into some treasury securities.
6. A technique for investing in vehicles that have a slightly longer term than you like is to invest, say 1/4 of your cash in a 3 month CD, then in a month invest another quarter and in two months another. (Leaving 1/4 as cash) Then everytime a CD matures you can roll it into another (many banks will do this automatically) but at any given moment you have to wait no longer than 1 month to get your money.
7. Roth IRAs. If your cash now is from actual earned (taxable) income, you are probably eligible for a Roth. And if you are, you should be putting money in it up to your income or the Roth cap. Things to know:
A Roth is a tax designation, NOT a vehicle. So you can have a "Roth savings account" or a "Roth CD" (but you can't mix it with your other money, of course)
Because you contributed after-tax money to the Roth, you can generally withdraw anything up to the entire amount you contributed without penalty*. (You can't ever withdraw the INTEREST until your retire, though) This makes a Roth unique among retirement tax options in that it can double as a liquid reserve (if it's in a liquid vehicle)
*(You still don't especially WANT to withdraw it, because what you can't do is ever put that money back into the Roth. If you withdraw it you are essentially in the same position as having never put it in, but because there are caps on per-year Roth contributions there's a big incentive to start early and keep contributing.)
To my knowledge - which is admittedly a year or so old - basically there are three relevant points.
1. Most people have 3 color receptors that they actually use, while some are colorblind to varying degrees including a relatively high number are red-green colorblind having effectively one RG and one B receptor. HOWEVER, where (what wavelength) the "R" "G" and "B" receptors is is NOT exactly the same for each person. So it is very possible that a perfect match for one person is not a perfect match for another especially for colors that are a complex mixture of wavelengths (eg most real-life pigments in sunlight) Note that generally matching the amount of the same pigment should generally be very, very close - to demonstrate this effect you mostly need to be combining very different wavelengths that "should" be the same added together.
The take-home geek message is that you can use an RGB monitor to match every color you can see - IF the monitor's RGB match yours. Otherwise it's not perfect. (Also see point 3)
Have two receptors very close together eventually becomes indistinguishable from just having one as they approach being in the same spot.
2. Some people are known as "tetrachromats" All examples I've heard about have been the mothers of red-green colorblind men. Essentially they have an extra receptor between R & G. This means that they can determine that two colors don't match in situations where everyone with three receptors would think they matched.
3. Apparently we may also have a 4th (or 5th, depending on pt 2) receptor in the ultraviolet range. However, most of the light in this range is blocked by the alchohol in our eye fluids, so this receptor is mostly pretty useless. However, this doesn't mean we don't see SOME color with this receptor right at the edge where it's not blocked by the alchohol - it's just not a very large part of our sight.
These colors definitely don't exist in monitors, which I personally and nonscientifically think is why I love staring at the LED on a PS2.
If you're going to wget, may I suggest playing nice with rates, delays, and non-essential files:
t ml -X /Video,/Chat,/Forum,/Blog
wget -w 2 --limit-rate=5k -m http://www.sysinternals.com/SysinternalsSiteMap.h
A relatively reasonable 81MB and 553 files.
And don't redistribute. As a software author, beyond it being illegal I consider it to be quite disrespectful to the authors if you were to redistribute these files in violation of their license. On the other hand, it certainly seems like a good idea to make a personal copy at this moment, and that certainly IS in accordance with their license.
GP makes sense, but is not right. Note: I'm not an authority on the Shuttle program by any means, but this is just basic science. What I'm saying below is even more true for the hydrogen, but I used the oxygen as an example. There are two basic reasons it would be quite cold.
1. To make Liquid Oxygen at room temperature does require extremely low temperatures. But they aren't keeping it THAT cold, because it would be prohibitively difficult.
It doesn't need to be that cold, because under pressure the temperature where it stays liquid goes up. You can keep a strong tank of O at room temperature and it will still be liquid. So they likely keep it cold to reduce the pressures they have to maintain somewhat, but it's not required.
2. Whenever you reduce the pressure on a gas (or especially when a liquid becomes a gas) this is a very endothermic (heat absorbing) process. So a home oxygen tank IN USE is cold, but one being stored isn't. (Random metal at room temp will FEEL cold, but the tank isn't _especially_ cold)
When you pour the liquid oxygen into a large temperature tank - even if you do it really fast and in a sealed way - any room left in the tank instantly becomes filled with much higher-pressure gaseous Oxygen. This expansion makes it very cold. And there is a lot of room in the tank when you START filling, even if by the end you filled it all.