Most areas don't let you grow a "naturalized lawn" and unless you have some animals eating your lawn your just going to end up with 1-2 foot tall stands of grass everywhere. Which while not that unappealing it keeps the space from being used. You might want to look into exchange your lawn for flagstones their low matence, look nice are environmentally neutral and most homeowners associations are willing to let you get by with them.
Re:The Best Laid Plans of Mice and Men ...
on
State of the Union
·
· Score: 1
"the inevitable collapse has been pushed off another 40 years." I don't know about you but I have little faith that anyone has any clue what's going to happen in 40 years.
But, SS is has worked for a lot longer than I have been alive and I don't see the need to change much for it to keep working for the next 1000 years. There are realy only two problems with SS one is people have been living longer the other is most of the weath generated by the US economy is not seen my people who make under 88K per year and so there is little real growth in SS.
If we where to have everyone in the use pay the same % of there income into SS it would be fine. Afterall the largest problem is that the rich are getting richer while the poor get left behind and it's the poor people who pay the most money into SS.
As you seem to know a lot about this topic would you mind answering responding to a few ideas?
As I understand it the basic heating problem is your dumping a lot of kenectic energy on reentry. But why not use some form of active cooling on a wing that glides on the upper atmosphere thus dumping the kenectic energy over a longer period of time?
I mean it looks like the shuttle basicly operates by supper heating the air on it's decent and useing a skin that's thick enough to withstand that much heat untill it slows down enough to start air cooling to a reasonable temperature. But if you buildt a wing that provides lift at mach 22 and then cooled it with watter that you flash boiled and then vented at say 2000C might be able to keep the wings within a reasonable temerature for longer and thus come closer to say the SR71 than a falling rock.
You could then use jet's or even a scram jet to get to the uppper atmosphere vs a rocket that wastes a lot of energy just keeping from falling.
We don't need "something as radical as a spaceplane or nuclear rocket" We need a system to safely and cheeply move people into space. Just make a shuttle 10% of it's old size and we will save a TUN of money. The whole idea that your shuttle system should take up cargo is a huge waste. Just dock in space with whatever you want them to work with and you can use rockets which are a lot cheeper to move stuff into orbit than any space plane.
Let's say terrorism killed 10,000 people over the last 10 years. That's 280,000,000 / 10,000 * 10 or 1 in every 280,000 americans per year which is less than fucking bee stings. The year terrorism start's to kill 100k people every year is the year it's worth nothing untill then it's way to tiny an isue to bother with IMO.
WoW, a well informed and insightful political post on Slashdot I am shocked. To bad I don't have mod points but I still salute you.
The only thing I wish you would add to future rants is the disenfranchisement of anyone who has been convicted of drug offences. It's been an amazingly effective strategy for removing the right to vote from both the poor and the radical.
According to most of the people on/., Iraq was never a hotbed of terrorism either. And yet, it seems that terrorism has reared its ugly head there.
Iraq became a hotbed of terrorism because we are there. The totalitarian government was secular and would kill off religuious extreamest movements before they got to terrorism. Now days there is no true government in Iraq so terrorism is having a field day.
We have this stuff it just does not work all that well.
Automated language translation - Google.
Self-programming computers. - There are a tun of learning systems in AI. Granted they only work within a problem domain but they still work.
Natural language understanding and interfaces. - Ask Geves (SP?) Yea it sucked for a lot of things but it was still a start.
Image understanding. -There are lot's of robot's that can avoid objects based on Images.
The problem is we don't have AI that's smarter than we are but just becasue such systems are not "perfact" does not mean they are useless. A robot that just follows the red line on the floor can do usefull work as can the mouse who can only compare to images and see that you moved it to the right some. Hell image compression requires an understanding of what about the immage is important to us.
Your missing a step. Let's write a simulater for a 286 with 1MB ram and a 20meg HDD. And run it on a P4 with 2 gigs of ram and a 200GIG HDD. Now clearly you can run check to see if a state on the 286 shows up again. So you take your P4 and run it though 2^22,000,000 states if it has not haulted then it's hiting a stat it was already at so it's going to repeat forever. If it haults before then it haults. Now you can simulate your program running on the 286 with a coppy of the program and it's going to run out of memory and hault.
The basis for this is the idea that you can have 2 finite stat machens running the same program shure you can't run it on the SAME machine but it will work on something with a little over 2x the systems memory. So in the real world you can find out if it's going to hault on a users system if you run it on a system with say 3x the memory. (Might not end within a usefull time frame but we are talking about thery of finite state machens not usefull real world testing)
That's not a problem, as programs run on finite state machines. It's easy to find out if program running on a finite stat machine will Hault just run it and see if the any states ever repeat. As there are a finite number of states then the solution will show up in finite time.
PS: "The undecidability of the halting problem relies on the fact that algorithms are assumed to have potentially infinite storage" so in the real world it's not a problem.
First off Simulating a mind would still take 15-25 simulated years before the mind would be able to handle complex human level tasks.
Also, your also talking about second by second computation learning is much more complex and involves creation / destruction of pathways in a non random fashion. There are also other pathways in the brain other than direct neuron-to-neuron communication such as chemical release that effects large numbers of neurons at the same time. AKA Endorphins and such.
This is why people will up the number of TFlops / second you would need vs. the numbers your showing.
I rember using it in an astonomy program it was usefull to have the first 12 digits of PI. Then again they want acuarcay out to +/- 1/360/60/60/1000 milli arc seconds and all that.
I find the best managers are people who like geting things done. Most problems have several solutions but people who look at problems as things to be provented go from fighting fires to increasing productivity vary quickly. People who like being productive tend to want to manage things when there faced with poor management and they tend to be happy when there part of a coheasive whole.
A clasic example is the software interfaces between billing systems. Some people just want to hack there way though each and every interface which might seem productive but they tend to create huge problems over time. Good managers look to balance the present with future needs so they will start to transform such a clugy solution into a eficent working system over time. This tends to promote productivity as people start to see there jobs becomeing easer over time as well as reducing stress as people move from 24/7 crunch time to anticipating future needs.
Bad management is often sacrificing the future to meet todays needs. They want to higher cheep labor and avoid investing in infrastructure ect. But, things can be as simple as providing donuts at break time on a shitty job. If your working a help desk the company may think of everyone as instantly replacable but a good manager will try and reduce stress cheeply thus saving on the bottom line by reducing turnover slightly. Something as small as a smile or a thank you can go a long way to increasing morale in bad situations but it is people who want to be productive that tend to think of such things as "cheep" productivty aids vs. a waste of time.
Thats:
So you go though say 100 outer layer options then 200 contires then 50 states in a file of say 30,000 data elements you found just what you where looking for after parcing only 100+200+50 data elements vs the whole file ( chanes are you got lucky and found it on closer to 50+50+25 data elements.
XSLT doesn't work with your format. Web browsers don't have a built-in parser for your format. The DOM doesn't handle your format. XPath can't reference data inside your format. CSS can't style your format.
None of which is usefull. I think XML is great as a methiod of information exchange but it sucks as a methiod of information storage / retreavel. Simple example let's say you want to know what the capital for something is.
So you search for Capital, USA, Ohio. Well in my format you do could setup your data so it's the outer layer is Capital, followed by Contry, Follwed by state's if there is any. Now as the outer layer lists the Capital you can search the outer layers skiping over any data that's not usefull because you know the size of the data that's stored in it. So you go though say 100 outer layer options then 200 contires then 50 states in a file of say 30,000 data elements you found just what you where looking for after parcing only 100*200*50 data elements vs the whole file ( chanes are you got lucky and found it on closer to 50*50*25 data elements. Now shure you could read the whole file into memory at startup and get the data into the format you want but that's can take a while and makes using XML even slower. As you have parse the while file even if you only realy need 50 or so data elements to work with.
That's an advantage to this data format vs XML which gives you human readability and the ablity to look at the data but as a what's the point in looking at that much data at in a web browser?
PS: And yes I do have well over 30,000 data elements in a setup file that my users don't want to store on a database. And yes my app reads the whole file into memory but only parces it as it needs to.
Don't forget that the intrest on the trust fund pays for SS well after 2018. It's a few trillion making 5% so that's going to take a while to go away. But, I see your point and agree but once again going extream.
If somone suddenly created a "nano-plague" that make everyone imune to any and all types of death and suddenly put everyone into a sate of perfact health than SS would colapse. But, I think we can both agree that this type of increase in life span for everyone on the same day is not going to happen.
Now let's say some people like 1% of the population get to live forever well you take them out of SS. How about a drug that makes most people live 10 years longer. Well pushing back the retirement age at a rate that's close to 0 would hirt a lot of people but if there going to be healthy then they can still work. Hell, your going to live 10 years longer you have to spend 6 of them working... That's fine with me.
The break even point for SS is when the interest on the trust fund after inflation is less than the gap between incomeing funds and outgoing funds. That where the problem shows up. It would just be gaining a head of steam and the trust fund running out is a silly idea afterall the Intrest on the trust fund is realy just tax money so when that preset tax money is running out we are going to need to raise taxes. Hell even with the trust fund in place the national debt will start to increase as soon as the trust fund goes over the intrest being pait on it's debt.
What's realy going to break is Medicare. People may be living longer but there living in worse health. Pushing back the date people start geting Medicare does little becuase it's the elderly that take most of the money. It has been said that smoker's cost the tax payers X more money but the trouth of it is most smokers die sooner and end up costing less money in the end than non smokers. Afterall, a 3-5 year stint with drugs followed by death is a hell of a lot cheeper than poping in and out of hospitals from the time your late 60's till death 20-40 years latter. Every time somone is saved from a hart attack you can see 10 of thousands of dollars sucked out of Medicare let alone SS.
The way I see it Medicare is going to need to cap both yearly and life time spending if it's going to work over the long haul. You can give a hart transplant to an 95 year old man but that's not going to give him 20 years. Medicare needs to stop treating extended hospital stays as a reasonable expence and start sending people home yes there going to die but they will die with or with out help. They might live 2 more months in a hospital but at some point we need to start letting people go with out extream measures.
However, even if you assume that software is 50 times harder than engineering (I'll contest that!!!), you're back to looking at something the complexity of the Linux kernel.
OK, now I am back to going bug eyed. 5Lines of code per min is nuts. Now Cassini-Huygens is extreamly complex system but you have to look at what makes it complex. When you stat saying that people create software at say 5 lines of code per unit time you saying "code" has a given complexity which is clearly not the case.
At one point in time NASA was paying 1000$ per line of code. Now the reason for this is that they test there code. As a basic rule of thumb any random coder of any level will be somewhere around 1/10 to 10 times as efecent as a coder of the same background. So take two people with a BS from MIT and give them the same problem and one guy may show up with code in 1/2 the time as the other person and that code will be so much better than the other persons that they would have to spend 5x as long to create code that is as bug free / feature rich.
However, if your juding the value / effert put into coding software then a well tested piece of code will be worth 6-10 times as much as a new pice of code that does the same thing. Now all these ideas are why a most software projects tend to create software closer to 1-5 Lines of Code per hour over there life time than 5 Lines per Min.
2ndly as I said if your taking the full number of man hours for a given project like Cassini-Huygens you have to seperate the construction times from the develepment time and evaluate them sepratly. Afterall building a bridge is not a simple step but there is a lot more posiblity of failure at the design stage than when somone is adding the rivits. Shure they could fuck up but visual inspection makes applying rivits a reasonably safe endever aka you can see if you fucked up by looking at it. Where as a desginer who miscaluclates the harmonic's that are setup by trafic can easly cause the system to fail. Not only that but a lot of there mistakes can be glossed over by building it with 20% overkill over what they think would ever be needed. AKA Max stress is trafic at a sandstill. Max harmonic load is Max stress X 5. Let's build it to take Max stress X 6 and things should be fine. Where there is no whay to do the same thing with software.
Now you can say Cassini-Huygens is a lot more complex than a bridge and I would agree with you but they can still build it with the idea that there going to need X power at max so let's set it up to give 20% more than that to be safe. Now as to power sources there may be a lot of man hours to add the power source but as several space based power sources have been tested in space there is little risk that there idea will not work if they used a slight change to the same basic design. Thus most of the man hours put into the powersource are vary low risk were there is no area in most software projects that are vary low risk.
I would say Cassini-Huygens would be much more complex than Linux if you where starting from scratch say in 1900 and wanted to build it but at this point most of the work is minor changes to extreamly well tested systems. Hell Linux started out as a vertion of UNIX that was not totaly unix compatible. And as such was much less complex than the work NASA(and it's contractors ect.) did just for Cassini-Huygens but right now I would say it's a much more complex system than the new work done just for Cassini-Huygens.
In the real world people do care how you dress and your grooming habits. So HS does teach you some useful skills as far as human interaction goes. But knowing your never going to be around most of these people ever again you can work on other things...
1. Pick a group of people and Lie all the time to them for a week and see if you can get away with it. Which lies are easy to pick out and which are hard. What type of actions give away lies and which type of lie is so benign that people forget you said it.
2. Pick some group and start bending them to your will. See if you can change speech patterns, modes of dress, thought's feelings, or what ever. But, keep these people from knowing what your doing / keep them from noticing that they changed the way they wave at each other or if they notice keep them from noticing that your the trend setter.
3. Get someone or some group to warship you. Either as the person who decides things for the group or the person who they look to for advice. Ect.
4. Keep someone from dropping out of something either a group or activity. Or if you want to really work for it go for someone that was going to drop out of high school.
5. Get someone to drop out of some group or activity.
Some of these skills become invaluable in life. Others give you some moral issues to deal with. EX: Do you feel glad that you destroyed someone faith in God? But, you can practice most of these skills with little fallout latter in life. And it becomes a lot more obvious when someone is trying to manipulate you, which makes it a lot easer to discover their motivations, and thus how to control them.
At first I went a little bug eyed at the "5 lines per minute on average" but as 2 mil / 5 = 400k I think your either talking about hours or suck at math...
Anyway, the point your missing with your argument is making software is a vary decision intense (yea I made that one up) endeavor where say putting together a solar cell quickly comes down to a few 100 choices made in repetition. And it's not the first probe sent into space, which means that a lot of choices have already been made. I mean the chose to use a rocket type that was already in use so that was one choice that was both easy and created a lot of man hours worth of work people already did.
Think of it this way the first man on the moon took a LOT of work the 7th was not such a big deal. Yea there was a lot of work but most of the decisions had already been made and a lot of those left where more a long the lines of "does this look like that (Y/N)" vs. "What should I do if I get data that is clearly wrong at this point?"
Now I am not saying it's easy but it's a lot simpler than making the modern Linux kernel.
Ok let's say 10% of the pople in the use are over 70 and on SS today.
Ok now let's say in 10 years. 9% of the population is over 80 because some people die and some people are born well now you let anyone over 79 into SS.
Ok now let's say in 10 years. 9% of the population is over 89 because some people die and some people are born well now you let anyone over 88 into SS.
The idea is that the closer to SS the sooner you retire but it's not at a 1:1 ratio so people that are 1 year away should retire in 1.5 years say but peple 5 years away retire in 7 years. And people 10 years away retire in 25 years ect.
Hell if nobody died from old age starting today then SS would still take a few years to colapse. The idea that imortality would break the system in a sigle year is just as silly as the idea that people can keep retiring at the same age regardless of how long they live.
My first year out of collage I was living in my own apartment. Making 38k/year and I spend 1grand on a snow bord and gave my parents a 200$ gift (mom got 1500$ laptop dad got a 19" monitor). And I still saved 7.5% of my income. I use the snow bord enough that it's worth it and I just wanted to say thanks to my parents.
Now I am making 46k/year and am about to get the next anual raise.
I live 5 miles from work so there is little comute and drive and while I drive an old car I don't see the point in getting a new one. And other than a grand on my PC this year I am still not buying junk. Now I am looking at making more money every year for a while so what do I do with that money?
I am thinking of going back to school but WTF am I going to do when I start making 100K+ a year? I have been giving away way to much money but it's not realy making me happy and some of my friends seem to be creaped out by it a little. Then again I think there in the same situation where they can do things that cost money but they don't realy want to. I mean hell you show up to a party with two 40$ bottles of wine nobody seems to want to drink them after a while. Hell, I was thinking of picking up one of those huge HDTV flat screans for a while but I don't realy watch much tv so it's seems more silly than anything else to pick one of those up.
At this rate I think I will either retire mid to late 40's or start a family. As for right now I feel way to young at 24 to have kids. Then agian I will probably change my mind in a few years so I figure I might as well just dump money into the stock market.
Anyway, it's the #1 reason to realize private act's don't make since is there is no advantage to a private investment system. Where the government to take up to 1/4 of the trust fund and invest it in a diversified world wide portfolio I would be happy but there is no reason to have private investors mucking around with private act's there net performance is not going to beat random stock picks and your going to end up with an industry trying to advise them.
By the way I pick 1/4 as a cap because it's fairly low risk I mean the odds of a balanced portfolio losing 50% of it's value over 20 years is really low which means the risk's are reasonable. And it also reduces the IOU nature of Social Security savings.
PS: By diversified I mean no industry would have more than 10% of the capital. 40% of savings outside the US no country receiving more than twice there % of savings over there GDP as compared to the worlds GDP. AKA it's not going to be used promote economic growth so much as make money. No company can have more than 0.1% of the net savings. That and you export the profits out of the stock market so it's never more than 1/4 of the trust fund. Also investments should never fall under the domain of a single entity so you would have 3 separate institutions providing investment advice and an arbitration entity that can only approve investments made by these organizations. Thus helping to limit fraud.
I say limit the % of the population that can be in SS at any one time say 5-10% of the workforce. That would work with any life span. And there could never be a colapse.
1. But, I already have the code to do this. I have code that reads / writes to this format in C, C++, Pascal, and Java as well as some great libraries to edit and display this format. It's faster than XML. I can write/read to this format at speeds that are limited more by the HDD than the CPU where the last time I checked XML was realy slow. I can even jump though the file with out parcing the whole thing. And there is some built it redundancy to show if the file format is valid. On the other hand I if you can tell me some usefull tools for use with XML I will look into it as I can already export this format to XML but then again it's been a while sence I looked at it so there may be some great XML tools out there.
2. It's a data storage format that's designed to handle tree's and list's natively. It aslo has a little redundancy so I can notice if I fucked up when writting the data. What new features can XML give me when dealing with setting information for custom apps? I mean it's already vary extensible in that code ignores tag names and format id's that it does not understand while still being able to pass them along.
As to the GUI a lot of this code does things like setup window locations and format's with drag and drop editing which you need a custom GUI for so I don't see how using XML will help with this. Now where I going to look at say a list of names and locations that would be one thing but things are a lot more complex than that. Hmm, I am starting to think I should just show you some of these tools so it makes more sence to you but I think we are not realy talking abou the same things. When you compare XML to some other format your saying "Look at all the tools for XML out there." Where when I looked into XML I was like "that's interesting but there are no usefull tools for working with it." Hell, my first thought was wow they reinvented STML and gave it a new name.
I mean if you where to compare C to Pascal I would say Pascal wins hands down. Shure C has some great liburary's but you can use them from Pascal so there is little point. Which is the same thing with this format I can make XML code out of it where I ever to need to use some XML tool but I get a lot of use out of the native format so what's the advantage to the format?
Slashdot Mirror
Most areas don't let you grow a "naturalized lawn" and unless you have some animals eating your lawn your just going to end up with 1-2 foot tall stands of grass everywhere. Which while not that unappealing it keeps the space from being used. You might want to look into exchange your lawn for flagstones their low matence, look nice are environmentally neutral and most homeowners associations are willing to let you get by with them.
NOT FOR A ALL*
*ok they can kill them selves...
"the inevitable collapse has been pushed off another 40 years." I don't know about you but I have little faith that anyone has any clue what's going to happen in 40 years.
But, SS is has worked for a lot longer than I have been alive and I don't see the need to change much for it to keep working for the next 1000 years. There are realy only two problems with SS one is people have been living longer the other is most of the weath generated by the US economy is not seen my people who make under 88K per year and so there is little real growth in SS.
If we where to have everyone in the use pay the same % of there income into SS it would be fine. Afterall the largest problem is that the rich are getting richer while the poor get left behind and it's the poor people who pay the most money into SS.
As you seem to know a lot about this topic would you mind answering responding to a few ideas?
As I understand it the basic heating problem is your dumping a lot of kenectic energy on reentry. But why not use some form of active cooling on a wing that glides on the upper atmosphere thus dumping the kenectic energy over a longer period of time?
I mean it looks like the shuttle basicly operates by supper heating the air on it's decent and useing a skin that's thick enough to withstand that much heat untill it slows down enough to start air cooling to a reasonable temperature. But if you buildt a wing that provides lift at mach 22 and then cooled it with watter that you flash boiled and then vented at say 2000C might be able to keep the wings within a reasonable temerature for longer and thus come closer to say the SR71 than a falling rock.
You could then use jet's or even a scram jet to get to the uppper atmosphere vs a rocket that wastes a lot of energy just keeping from falling.
We don't need "something as radical as a spaceplane or nuclear rocket" We need a system to safely and cheeply move people into space. Just make a shuttle 10% of it's old size and we will save a TUN of money. The whole idea that your shuttle system should take up cargo is a huge waste. Just dock in space with whatever you want them to work with and you can use rockets which are a lot cheeper to move stuff into orbit than any space plane.
Let's say terrorism killed 10,000 people over the last 10 years. That's 280,000,000 / 10,000 * 10 or 1 in every 280,000 americans per year which is less than fucking bee stings. The year terrorism start's to kill 100k people every year is the year it's worth nothing untill then it's way to tiny an isue to bother with IMO.
WoW, a well informed and insightful political post on Slashdot I am shocked. To bad I don't have mod points but I still salute you.
The only thing I wish you would add to future rants is the disenfranchisement of anyone who has been convicted of drug offences. It's been an amazingly effective strategy for removing the right to vote from both the poor and the radical.
According to most of the people on /., Iraq was never a hotbed of terrorism either. And yet, it seems that terrorism has reared its ugly head there.
Iraq became a hotbed of terrorism because we are there. The totalitarian government was secular and would kill off religuious extreamest movements before they got to terrorism. Now days there is no true government in Iraq so terrorism is having a field day.
We have this stuff it just does not work all that well.
Automated language translation - Google.
Self-programming computers. - There are a tun of learning systems in AI. Granted they only work within a problem domain but they still work.
Natural language understanding and interfaces. - Ask Geves (SP?) Yea it sucked for a lot of things but it was still a start.
Image understanding. -There are lot's of robot's that can avoid objects based on Images.
The problem is we don't have AI that's smarter than we are but just becasue such systems are not "perfact" does not mean they are useless. A robot that just follows the red line on the floor can do usefull work as can the mouse who can only compare to images and see that you moved it to the right some. Hell image compression requires an understanding of what about the immage is important to us.
Your missing a step. Let's write a simulater for a 286 with 1MB ram and a 20meg HDD. And run it on a P4 with 2 gigs of ram and a 200GIG HDD. Now clearly you can run check to see if a state on the 286 shows up again. So you take your P4 and run it though 2^22,000,000 states if it has not haulted then it's hiting a stat it was already at so it's going to repeat forever. If it haults before then it haults. Now you can simulate your program running on the 286 with a coppy of the program and it's going to run out of memory and hault.
The basis for this is the idea that you can have 2 finite stat machens running the same program shure you can't run it on the SAME machine but it will work on something with a little over 2x the systems memory. So in the real world you can find out if it's going to hault on a users system if you run it on a system with say 3x the memory. (Might not end within a usefull time frame but we are talking about thery of finite state machens not usefull real world testing)
That's not a problem, as programs run on finite state machines. It's easy to find out if program running on a finite stat machine will Hault just run it and see if the any states ever repeat. As there are a finite number of states then the solution will show up in finite time.
PS: "The undecidability of the halting problem relies on the fact that algorithms are assumed to have potentially infinite storage" so in the real world it's not a problem.
First off Simulating a mind would still take 15-25 simulated years before the mind would be able to handle complex human level tasks.
Also, your also talking about second by second computation learning is much more complex and involves creation / destruction of pathways in a non random fashion. There are also other pathways in the brain other than direct neuron-to-neuron communication such as chemical release that effects large numbers of neurons at the same time. AKA Endorphins and such.
This is why people will up the number of TFlops / second you would need vs. the numbers your showing.
I rember using it in an astonomy program it was usefull to have the first 12 digits of PI. Then again they want acuarcay out to +/- 1/360/60/60/1000 milli arc seconds and all that.
I find the best managers are people who like geting things done. Most problems have several solutions but people who look at problems as things to be provented go from fighting fires to increasing productivity vary quickly. People who like being productive tend to want to manage things when there faced with poor management and they tend to be happy when there part of a coheasive whole.
A clasic example is the software interfaces between billing systems. Some people just want to hack there way though each and every interface which might seem productive but they tend to create huge problems over time. Good managers look to balance the present with future needs so they will start to transform such a clugy solution into a eficent working system over time. This tends to promote productivity as people start to see there jobs becomeing easer over time as well as reducing stress as people move from 24/7 crunch time to anticipating future needs.
Bad management is often sacrificing the future to meet todays needs. They want to higher cheep labor and avoid investing in infrastructure ect. But, things can be as simple as providing donuts at break time on a shitty job. If your working a help desk the company may think of everyone as instantly replacable but a good manager will try and reduce stress cheeply thus saving on the bottom line by reducing turnover slightly. Something as small as a smile or a thank you can go a long way to increasing morale in bad situations but it is people who want to be productive that tend to think of such things as "cheep" productivty aids vs. a waste of time.
Thats: So you go though say 100 outer layer options then 200 contires then 50 states in a file of say 30,000 data elements you found just what you where looking for after parcing only 100+200+50 data elements vs the whole file ( chanes are you got lucky and found it on closer to 50+50+25 data elements.
XSLT doesn't work with your format. Web browsers don't have a built-in parser for your format. The DOM doesn't handle your format. XPath can't reference data inside your format. CSS can't style your format.
None of which is usefull. I think XML is great as a methiod of information exchange but it sucks as a methiod of information storage / retreavel. Simple example let's say you want to know what the capital for something is. So you search for Capital, USA, Ohio. Well in my format you do could setup your data so it's the outer layer is Capital, followed by Contry, Follwed by state's if there is any. Now as the outer layer lists the Capital you can search the outer layers skiping over any data that's not usefull because you know the size of the data that's stored in it. So you go though say 100 outer layer options then 200 contires then 50 states in a file of say 30,000 data elements you found just what you where looking for after parcing only 100*200*50 data elements vs the whole file ( chanes are you got lucky and found it on closer to 50*50*25 data elements. Now shure you could read the whole file into memory at startup and get the data into the format you want but that's can take a while and makes using XML even slower. As you have parse the while file even if you only realy need 50 or so data elements to work with.
That's an advantage to this data format vs XML which gives you human readability and the ablity to look at the data but as a what's the point in looking at that much data at in a web browser?
PS: And yes I do have well over 30,000 data elements in a setup file that my users don't want to store on a database. And yes my app reads the whole file into memory but only parces it as it needs to.
Don't forget that the intrest on the trust fund pays for SS well after 2018. It's a few trillion making 5% so that's going to take a while to go away. But, I see your point and agree but once again going extream.
If somone suddenly created a "nano-plague" that make everyone imune to any and all types of death and suddenly put everyone into a sate of perfact health than SS would colapse. But, I think we can both agree that this type of increase in life span for everyone on the same day is not going to happen.
Now let's say some people like 1% of the population get to live forever well you take them out of SS. How about a drug that makes most people live 10 years longer. Well pushing back the retirement age at a rate that's close to 0 would hirt a lot of people but if there going to be healthy then they can still work. Hell, your going to live 10 years longer you have to spend 6 of them working... That's fine with me.
The break even point for SS is when the interest on the trust fund after inflation is less than the gap between incomeing funds and outgoing funds. That where the problem shows up. It would just be gaining a head of steam and the trust fund running out is a silly idea afterall the Intrest on the trust fund is realy just tax money so when that preset tax money is running out we are going to need to raise taxes. Hell even with the trust fund in place the national debt will start to increase as soon as the trust fund goes over the intrest being pait on it's debt.
What's realy going to break is Medicare. People may be living longer but there living in worse health. Pushing back the date people start geting Medicare does little becuase it's the elderly that take most of the money. It has been said that smoker's cost the tax payers X more money but the trouth of it is most smokers die sooner and end up costing less money in the end than non smokers. Afterall, a 3-5 year stint with drugs followed by death is a hell of a lot cheeper than poping in and out of hospitals from the time your late 60's till death 20-40 years latter. Every time somone is saved from a hart attack you can see 10 of thousands of dollars sucked out of Medicare let alone SS.
The way I see it Medicare is going to need to cap both yearly and life time spending if it's going to work over the long haul. You can give a hart transplant to an 95 year old man but that's not going to give him 20 years. Medicare needs to stop treating extended hospital stays as a reasonable expence and start sending people home yes there going to die but they will die with or with out help. They might live 2 more months in a hospital but at some point we need to start letting people go with out extream measures.
However, even if you assume that software is 50 times harder than engineering (I'll contest that!!!), you're back to looking at something the complexity of the Linux kernel.
OK, now I am back to going bug eyed. 5Lines of code per min is nuts. Now Cassini-Huygens is extreamly complex system but you have to look at what makes it complex. When you stat saying that people create software at say 5 lines of code per unit time you saying "code" has a given complexity which is clearly not the case.
At one point in time NASA was paying 1000$ per line of code. Now the reason for this is that they test there code. As a basic rule of thumb any random coder of any level will be somewhere around 1/10 to 10 times as efecent as a coder of the same background. So take two people with a BS from MIT and give them the same problem and one guy may show up with code in 1/2 the time as the other person and that code will be so much better than the other persons that they would have to spend 5x as long to create code that is as bug free / feature rich.
However, if your juding the value / effert put into coding software then a well tested piece of code will be worth 6-10 times as much as a new pice of code that does the same thing. Now all these ideas are why a most software projects tend to create software closer to 1-5 Lines of Code per hour over there life time than 5 Lines per Min.
2ndly as I said if your taking the full number of man hours for a given project like Cassini-Huygens you have to seperate the construction times from the develepment time and evaluate them sepratly. Afterall building a bridge is not a simple step but there is a lot more posiblity of failure at the design stage than when somone is adding the rivits. Shure they could fuck up but visual inspection makes applying rivits a reasonably safe endever aka you can see if you fucked up by looking at it. Where as a desginer who miscaluclates the harmonic's that are setup by trafic can easly cause the system to fail. Not only that but a lot of there mistakes can be glossed over by building it with 20% overkill over what they think would ever be needed. AKA Max stress is trafic at a sandstill. Max harmonic load is Max stress X 5. Let's build it to take Max stress X 6 and things should be fine. Where there is no whay to do the same thing with software.
Now you can say Cassini-Huygens is a lot more complex than a bridge and I would agree with you but they can still build it with the idea that there going to need X power at max so let's set it up to give 20% more than that to be safe. Now as to power sources there may be a lot of man hours to add the power source but as several space based power sources have been tested in space there is little risk that there idea will not work if they used a slight change to the same basic design. Thus most of the man hours put into the powersource are vary low risk were there is no area in most software projects that are vary low risk.
I would say Cassini-Huygens would be much more complex than Linux if you where starting from scratch say in 1900 and wanted to build it but at this point most of the work is minor changes to extreamly well tested systems. Hell Linux started out as a vertion of UNIX that was not totaly unix compatible. And as such was much less complex than the work NASA(and it's contractors ect.) did just for Cassini-Huygens but right now I would say it's a much more complex system than the new work done just for Cassini-Huygens.
In the real world people do care how you dress and your grooming habits. So HS does teach you some useful skills as far as human interaction goes. But knowing your never going to be around most of these people ever again you can work on other things...
1. Pick a group of people and Lie all the time to them for a week and see if you can get away with it. Which lies are easy to pick out and which are hard. What type of actions give away lies and which type of lie is so benign that people forget you said it.
2. Pick some group and start bending them to your will. See if you can change speech patterns, modes of dress, thought's feelings, or what ever. But, keep these people from knowing what your doing / keep them from noticing that they changed the way they wave at each other or if they notice keep them from noticing that your the trend setter.
3. Get someone or some group to warship you. Either as the person who decides things for the group or the person who they look to for advice. Ect.
4. Keep someone from dropping out of something either a group or activity. Or if you want to really work for it go for someone that was going to drop out of high school.
5. Get someone to drop out of some group or activity.
Some of these skills become invaluable in life. Others give you some moral issues to deal with. EX: Do you feel glad that you destroyed someone faith in God? But, you can practice most of these skills with little fallout latter in life. And it becomes a lot more obvious when someone is trying to manipulate you, which makes it a lot easer to discover their motivations, and thus how to control them.
At first I went a little bug eyed at the "5 lines per minute on average" but as 2 mil / 5 = 400k I think your either talking about hours or suck at math...
Anyway, the point your missing with your argument is making software is a vary decision intense (yea I made that one up) endeavor where say putting together a solar cell quickly comes down to a few 100 choices made in repetition. And it's not the first probe sent into space, which means that a lot of choices have already been made. I mean the chose to use a rocket type that was already in use so that was one choice that was both easy and created a lot of man hours worth of work people already did.
Think of it this way the first man on the moon took a LOT of work the 7th was not such a big deal. Yea there was a lot of work but most of the decisions had already been made and a lot of those left where more a long the lines of "does this look like that (Y/N)" vs. "What should I do if I get data that is clearly wrong at this point?"
Now I am not saying it's easy but it's a lot simpler than making the modern Linux kernel.
OK extream and simple example:
Ok let's say 10% of the pople in the use are over 70 and on SS today.
Ok now let's say in 10 years. 9% of the population is over 80 because some people die and some people are born well now you let anyone over 79 into SS.
Ok now let's say in 10 years. 9% of the population is over 89 because some people die and some people are born well now you let anyone over 88 into SS.
The idea is that the closer to SS the sooner you retire but it's not at a 1:1 ratio so people that are 1 year away should retire in 1.5 years say but peple 5 years away retire in 7 years. And people 10 years away retire in 25 years ect.
Hell if nobody died from old age starting today then SS would still take a few years to colapse. The idea that imortality would break the system in a sigle year is just as silly as the idea that people can keep retiring at the same age regardless of how long they live.
I mostly don't buy useless crap.
My first year out of collage I was living in my own apartment. Making 38k/year and I spend 1grand on a snow bord and gave my parents a 200$ gift (mom got 1500$ laptop dad got a 19" monitor). And I still saved 7.5% of my income. I use the snow bord enough that it's worth it and I just wanted to say thanks to my parents.
Now I am making 46k/year and am about to get the next anual raise.
I live 5 miles from work so there is little comute and drive and while I drive an old car I don't see the point in getting a new one. And other than a grand on my PC this year I am still not buying junk. Now I am looking at making more money every year for a while so what do I do with that money?
I am thinking of going back to school but WTF am I going to do when I start making 100K+ a year? I have been giving away way to much money but it's not realy making me happy and some of my friends seem to be creaped out by it a little. Then again I think there in the same situation where they can do things that cost money but they don't realy want to. I mean hell you show up to a party with two 40$ bottles of wine nobody seems to want to drink them after a while. Hell, I was thinking of picking up one of those huge HDTV flat screans for a while but I don't realy watch much tv so it's seems more silly than anything else to pick one of those up.
At this rate I think I will either retire mid to late 40's or start a family. As for right now I feel way to young at 24 to have kids. Then agian I will probably change my mind in a few years so I figure I might as well just dump money into the stock market.
Personally I think you where being trolled.
Anyway, it's the #1 reason to realize private act's don't make since is there is no advantage to a private investment system. Where the government to take up to 1/4 of the trust fund and invest it in a diversified world wide portfolio I would be happy but there is no reason to have private investors mucking around with private act's there net performance is not going to beat random stock picks and your going to end up with an industry trying to advise them.
By the way I pick 1/4 as a cap because it's fairly low risk I mean the odds of a balanced portfolio losing 50% of it's value over 20 years is really low which means the risk's are reasonable. And it also reduces the IOU nature of Social Security savings.
PS: By diversified I mean no industry would have more than 10% of the capital. 40% of savings outside the US no country receiving more than twice there % of savings over there GDP as compared to the worlds GDP. AKA it's not going to be used promote economic growth so much as make money. No company can have more than 0.1% of the net savings. That and you export the profits out of the stock market so it's never more than 1/4 of the trust fund. Also investments should never fall under the domain of a single entity so you would have 3 separate institutions providing investment advice and an arbitration entity that can only approve investments made by these organizations. Thus helping to limit fraud.
Your thoughts?
I say limit the % of the population that can be in SS at any one time say 5-10% of the workforce. That would work with any life span. And there could never be a colapse.
1. But, I already have the code to do this. I have code that reads / writes to this format in C, C++, Pascal, and Java as well as some great libraries to edit and display this format. It's faster than XML. I can write/read to this format at speeds that are limited more by the HDD than the CPU where the last time I checked XML was realy slow. I can even jump though the file with out parcing the whole thing. And there is some built it redundancy to show if the file format is valid. On the other hand I if you can tell me some usefull tools for use with XML I will look into it as I can already export this format to XML but then again it's been a while sence I looked at it so there may be some great XML tools out there.
2. It's a data storage format that's designed to handle tree's and list's natively. It aslo has a little redundancy so I can notice if I fucked up when writting the data. What new features can XML give me when dealing with setting information for custom apps? I mean it's already vary extensible in that code ignores tag names and format id's that it does not understand while still being able to pass them along.
As to the GUI a lot of this code does things like setup window locations and format's with drag and drop editing which you need a custom GUI for so I don't see how using XML will help with this. Now where I going to look at say a list of names and locations that would be one thing but things are a lot more complex than that. Hmm, I am starting to think I should just show you some of these tools so it makes more sence to you but I think we are not realy talking abou the same things. When you compare XML to some other format your saying "Look at all the tools for XML out there." Where when I looked into XML I was like "that's interesting but there are no usefull tools for working with it." Hell, my first thought was wow they reinvented STML and gave it a new name.
I mean if you where to compare C to Pascal I would say Pascal wins hands down. Shure C has some great liburary's but you can use them from Pascal so there is little point. Which is the same thing with this format I can make XML code out of it where I ever to need to use some XML tool but I get a lot of use out of the native format so what's the advantage to the format?