Some tasks can't be done in parallel and this is the Achilles heel of massive parallel architectures. See for instance http://en.wikipedia.org/wiki/Amdahl's_law. No parallel hardware and no parallel algoritms - no matter how clever - will help you, if you have a task of sequential nature (and even only help you somewhat no matter how massively parallel it is, if your task is partly sequential).
If one truckdriver can drive a truck 50 miles in one hour, how far can two drivers then drive the truck in the same amount of time?
Unlike Microsoft products, it's not like many of us are locked into using Google Have you ever tried to migrate pictures sorted and catalogued by Picasa? Can't be done - you'll lose for instance all your album definitions. The information is stored in some kind of proprietory database. Ever tried moving your pictures between harddrives? Same story - you'll need to "export" and re-"import" if the path to the pictures ever changes - or else live with the need to redo most of your cataloging again.
And yes, I'm bitter, because I had to learn it the hard way...
Excellent timeline. What it misses though is that the original article in Jyllandsposten was posted in response to an incident where the danish writer Kåre Bluitgen couldn't find an illustrator for his childrens book about Muhammed. The illustrators Mr. Bluitgen had contacted politely declined citing fear of personal safety.
At the time I saw the article as a provocation - but in my opinion the provocation was not directed against muslims or the arabic world - it was directed against the danes giving in to the atmosfere of threats and violence apparently originating from orthodox muslims. Fear and self-censorship is very slippery slope indeed.
It aches my heart to see how it all is blown totally out of proportions - how extremes on both sides is provoking each other and how every voice of moderation and reason was ignored. The seamingly calm waters had a strong undercurrent.
Why is a password that a user has committed to memory that never changes worse than a password that changes every three months that a user has to write down?
Because compromising written passwords do require physical access. If you have physical access to where my written password would have been stored (if I did write them down) then password security would be the least of my problems.
It is called an appeal. That means that the case is found to be of sufficient general interest to be retried at a higher level. As there are a finite number of levels with which to appeal the process will terminate at some point.
It is a very handy construct found in many legal system around the world.
OKOKRIM sound to me like a abbreviation of "Okonomisk Kriminalitet" (the first letter should be an "Oslash") which lead me to suspect that it is the prosecutor for economic criminality.
Databases containing personal information is stricly regulated by the way. Cross reference is VERY illegal, even for government institutions.
Make that "even more so for governmental institutions". Some years ago I worked at a Danish business school where we had employment data for approx. 10% of the danish workforce (where you had been employed and for how long - it was based on social security data (primarily "ATP-indberetninger")). Data where anonymized both regarding workers and workplaces (and some other dimensions I don't remenber today) but even so available to only a very limited group of scientists.
Even so we weren't allowed to keep the data for very long. The authorities was afraid that we might be able to track information about individuals and therefore afterwards we weren't allowed to have access to anything but aggregated data.
Having access to data that were immidiately crossreferencable was a goldmine. Take for instance the publication "Danish economy May 1988, The Labour Market" from the Danish Economic Counsil: For the first time economist where able to get behind the truth of unemployment rates. They could see that 75% of the workforce (the "A-team") where fulltime employed and 25% of the workforce (the "B-team") walked in and out for employment. If the unemploymentrate was 10% it only ment that the B-team where unemployed for shorter or longer periods on average 40% of their time.
What is my point? Well, there are actually two: 1) Danish laws regarding privacy are being rigorously upheld by the relevant authorities and 2) universially crossreferencable data is a goldmine for researchers (here are a list of publications from the Danish National Center for Register based Research) and used wisely benificial for everyone.
I remember talking to a sysadmin 10 years+ ago, that had a hard time not rolling over from laugther. Someone somewhere on the DECnet in the middle of nowhere in American had by mistake announced their transatlantic routing-cost as zero. Then two things happened:
1) as most machines on both side of the pond automaticly updated their routing tables this unfortunate machine got almost all transatlantic traffic routed to it. 2) as transatlantic traffic virtually stopped then all of this machines immitiate neighbors rather quickly manually updated their routing cost to this machine to close to infinity.
...was to see that users like my parent couldn't care less about security. Anything that hinders them in doing what they want to do is considered a nuisance.
I am also having a hard time explaining to my wife why I have revoked most priviliges for the "Internet zone" in IE (yes, I prefer Phoenix too, but the sad fact is, that there are many sites that won't work unless you use IE) - somehow it is still to much of a bother to add sites that we trust to the "trusted site" list the first time we visit them.
Maybe you just need to get burned once to respect fire (and of course understand what happened). Within the first 48 hours of my job as a student programmer I managed to wipe all files of several projects - my current directory wasn't what I thought is was and I had become more priviliged than I should be. That day I learnt not to invoke all priviliges in the login-script but only as needed. I also learnt something about proper backup routines that day - the nightly backup really saved me (thank you dear sysadmin for saving my from the wrath of my colleges...!).
Oh - that reminds me of another story. As student programmers we were given group-privileges (this was VAX/VMS). It was very practical to be able to start and stop job on behalf of other student programmers etc. Once one of the other student programmers wrote a utility that would log you out every interactive session wherever you where logged in (which was rather handy when someone asked if you would like to come along for beer). The utility stopped all interactive processes that it could find, but the author remembered to explicitly turn of our group priviliges before doing so, so we didn't accidentially log each other out. Somehow our sysadmin got hold of this utility and since it was throughly tried and tested by us for several months, he trusted it. One crusual difference between a student programmer and a sysadmin is that a sysadmin has world-priviliges and the script didn't turn these of...! He learnt the hard way that as a sysadmin you should trust noone.
When we asked above mentioned sysadmin for more privileges ("can I have oper-privileges so I can restart this print-queue?!?") he always answered "Do you want more responsabilities?" No, we only wanted more privileges. "Well" he said "it is the same thing - do you still want your privileges?" Somehow it wasn't really nessecary with more privileges anyway. And that is perhaps one of the most useful lessons to be learnt.
If the application is located on a high bandwidth network, it is actually possible to execute a java-application with a shortcut to a JVM located on a network drive. This gives zero-install on the clients - network performance might suffer though.
It is true that subscription can be a blessing for a provider. But it can also turn into a curse for both provider and customer.
I once (ca. 10 year ago) worked for a firm that sold a program for a yearly subscription (you didn't own the program - you leased the right to use it). It removed the focus of the management from the product to a degree were it almost wasn't supported anymore. There wasn't any pressure from dismissing sales as we lived almost on subscribtion alone.
But once a year a month or so before next year subscription was due I was told quickly to prepare a new release with the sole purpose of giving the impression that our customers did get something for their subscription. Management didn't care what it contained as long as I didn't take to long.
AFAIK most of our customers didn't use the upgrade because it didn't really add anything worthwhile anyway.
This gives me a bad case of deja vu. I still remember all to well the problems of 16 bit processor and the attempt to expand the address-range when 16-bit addressing became too big a constraint.
First there were the segment/offset addressing, which is bad enough. Then came extended/expanded memory and all its quirks og incompatibilities.
Lets not do that again. For most computers 32-bit linear addressing will still be enough for a while (remember, noone will ever need more than 64MB of RAM *grin*) - and for those who actually needs more than the 32-bit architecture can provide not going for the full monty of 64 bit will not be an issue.
The main point of Castle and Hendersons objection s is the the calculation is based on false assumption about the third world economic growth. The current assumption in the IPCC study would make the South African economy about four times greater than the American by 2100. Also they object against the growth-rates used - some growth-rates are triple figure percentages and the lagest currently known growthrate has been 20% for Japan in the last century.
So even the range of "between 1.5 and 6 degrees" is disputed. And this is based solely on the methodology of the economic/statistical calculations. Please note that I am not discussing the point, that there is also some scientifically based doubt about the causality between CO2 emission and global warning - on this point the jury is still out IMHO, and any conclusions will be premature (and therefore based more on belief).
Castle and Hendersons objections is described in this article in the Economist. IPCC is the "Intergovenmental Panel on Climate Change" and describes it self as:
"The Intergovernmental Panel on Climate Change (IPCC) has been established by WMO and UNEP to assess scientific, technical and socio- economic information relevant for the understanding of climate change, its potential impacts and options for adaptation and mitigation." (see
here ).
WMO is the "World Meteorological Organization" a "United Nations Specialized Agency" (see here ).
UNEP is the "United Nations Environment Programme" (se here ).
The crucial assumption that the earth will become 6 degress warmer within the next century probably stems from a IPCC study. But the IPCC study is being disputed - mainly for grossly overestimating the 3rd world growth. And with a more reasonable estimate of the economic growth, the resulting CO2 emission and therefore also the resulting global warming will be substantially lower.
See for instance here: http://www.kuro5hin.org/print/2003/2/17/15110/5194
The principles used in a sliderule is in use not just in aviation, but also in more down-to-the-earth tasks.
Being a little over average recommeded weigth I use one to calculate the energy contribution from fat in foods in the supermarket. It is to nifty little paperwheels with logaritmic scales joined by a clip, and given away for free. If it wasn't for my granddad giving me a sliderule back when I was playing with Lego for fun, I wouldn't have recogniced this lowtech paper-wheel as a modern incarnation of a sliderule.
So, even now people who wouldn't know a logaritm even if they saw one, are using sliderule for everyday lowtech tasks.
Slashdot Mirror
Some tasks can't be done in parallel and this is the Achilles heel of massive parallel architectures. See for instance http://en.wikipedia.org/wiki/Amdahl's_law. No parallel hardware and no parallel algoritms - no matter how clever - will help you, if you have a task of sequential nature (and even only help you somewhat no matter how massively parallel it is, if your task is partly sequential).
If one truckdriver can drive a truck 50 miles in one hour, how far can two drivers then drive the truck in the same amount of time?
And yes, I'm bitter, because I had to learn it the hard way...
Excellent timeline. What it misses though is that the original article in Jyllandsposten was posted in response to an incident where the danish writer Kåre Bluitgen couldn't find an illustrator for his childrens book about Muhammed. The illustrators Mr. Bluitgen had contacted politely declined citing fear of personal safety.
At the time I saw the article as a provocation - but in my opinion the provocation was not directed against muslims or the arabic world - it was directed against the danes giving in to the atmosfere of threats and violence apparently originating from orthodox muslims. Fear and self-censorship is very slippery slope indeed.
It aches my heart to see how it all is blown totally out of proportions - how extremes on both sides is provoking each other and how every voice of moderation and reason was ignored. The seamingly calm waters had a strong undercurrent.
Because compromising written passwords do require physical access. If you have physical access to where my written password would have been stored (if I did write them down) then password security would be the least of my problems.
It is called an appeal. That means that the case is found to be of sufficient general interest to be retried at a higher level. As there are a finite number of levels with which to appeal the process will terminate at some point.
It is a very handy construct found in many legal system around the world.
With the risk of spoiling a joke:
OKOKRIM sound to me like a abbreviation of "Okonomisk Kriminalitet" (the first letter should be an "Oslash") which lead me to suspect that it is the prosecutor for economic criminality.
Make that "even more so for governmental institutions". Some years ago I worked at a Danish business school where we had employment data for approx. 10% of the danish workforce (where you had been employed and for how long - it was based on social security data (primarily "ATP-indberetninger")). Data where anonymized both regarding workers and workplaces (and some other dimensions I don't remenber today) but even so available to only a very limited group of scientists.
Even so we weren't allowed to keep the data for very long. The authorities was afraid that we might be able to track information about individuals and therefore afterwards we weren't allowed to have access to anything but aggregated data.
Having access to data that were immidiately crossreferencable was a goldmine. Take for instance the publication "Danish economy May 1988, The Labour Market" from the Danish Economic Counsil: For the first time economist where able to get behind the truth of unemployment rates. They could see that 75% of the workforce (the "A-team") where fulltime employed and 25% of the workforce (the "B-team") walked in and out for employment. If the unemploymentrate was 10% it only ment that the B-team where unemployed for shorter or longer periods on average 40% of their time.
What is my point? Well, there are actually two: 1) Danish laws regarding privacy are being rigorously upheld by the relevant authorities and 2) universially crossreferencable data is a goldmine for researchers (here are a list of publications from the Danish National Center for Register based Research) and used wisely benificial for everyone.
I remember talking to a sysadmin 10 years+ ago, that had a hard time not rolling over from laugther. Someone somewhere on the DECnet in the middle of nowhere in American had by mistake announced their transatlantic routing-cost as zero. Then two things happened:
1) as most machines on both side of the pond automaticly updated their routing tables this unfortunate machine got almost all transatlantic traffic routed to it.
2) as transatlantic traffic virtually stopped then all of this machines immitiate neighbors rather quickly manually updated their routing cost to this machine to close to infinity.
Or this is at least how I remember the story...
Not only do I have OpenOffice installed, but also X-Windows... Oh boy - I must be in deep troubles now...!
How can you have an article about artwork and not having any pictures?
...was to see that users like my parent couldn't care less about security. Anything that hinders them in doing what they want to do is considered a nuisance.
I am also having a hard time explaining to my wife why I have revoked most priviliges for the "Internet zone" in IE (yes, I prefer Phoenix too, but the sad fact is, that there are many sites that won't work unless you use IE) - somehow it is still to much of a bother to add sites that we trust to the "trusted site" list the first time we visit them.
Maybe you just need to get burned once to respect fire (and of course understand what happened). Within the first 48 hours of my job as a student programmer I managed to wipe all files of several projects - my current directory wasn't what I thought is was and I had become more priviliged than I should be. That day I learnt not to invoke all priviliges in the login-script but only as needed. I also learnt something about proper backup routines that day - the nightly backup really saved me (thank you dear sysadmin for saving my from the wrath of my colleges...!).
Oh - that reminds me of another story. As student programmers we were given group-privileges (this was VAX/VMS). It was very practical to be able to start and stop job on behalf of other student programmers etc. Once one of the other student programmers wrote a utility that would log you out every interactive session wherever you where logged in (which was rather handy when someone asked if you would like to come along for beer). The utility stopped all interactive processes that it could find, but the author remembered to explicitly turn of our group priviliges before doing so, so we didn't accidentially log each other out. Somehow our sysadmin got hold of this utility and since it was throughly tried and tested by us for several months, he trusted it. One crusual difference between a student programmer and a sysadmin is that a sysadmin has world-priviliges and the script didn't turn these of...! He learnt the hard way that as a sysadmin you should trust noone.
When we asked above mentioned sysadmin for more privileges ("can I have oper-privileges so I can restart this print-queue?!?") he always answered "Do you want more responsabilities?" No, we only wanted more privileges. "Well" he said "it is the same thing - do you still want your privileges?" Somehow it wasn't really nessecary with more privileges anyway. And that is perhaps one of the most useful lessons to be learnt.
If the application is located on a high bandwidth network, it is actually possible to execute a java-application with a shortcut to a JVM located on a network drive. This gives zero-install on the clients - network performance might suffer though.
It is true that subscription can be a blessing for a provider. But it can also turn into a curse for both provider and customer.
I once (ca. 10 year ago) worked for a firm that sold a program for a yearly subscription (you didn't own the program - you leased the right to use it). It removed the focus of the management from the product to a degree were it almost wasn't supported anymore. There wasn't any pressure from dismissing sales as we lived almost on subscribtion alone.
But once a year a month or so before next year subscription was due I was told quickly to prepare a new release with the sole purpose of giving the impression that our customers did get something for their subscription. Management didn't care what it contained as long as I didn't take to long.
AFAIK most of our customers didn't use the upgrade because it didn't really add anything worthwhile anyway.
This gives me a bad case of deja vu. I still remember all to well the problems of 16 bit processor and the attempt to expand the address-range when 16-bit addressing became too big a constraint.
First there were the segment/offset addressing, which is bad enough. Then came extended/expanded memory and all its quirks og incompatibilities.
Lets not do that again. For most computers 32-bit linear addressing will still be enough for a while (remember, noone will ever need more than 64MB of RAM *grin*) - and for those who actually needs more than the 32-bit architecture can provide not going for the full monty of 64 bit will not be an issue.
So even the range of "between 1.5 and 6 degrees" is disputed. And this is based solely on the methodology of the economic/statistical calculations. Please note that I am not discussing the point, that there is also some scientifically based doubt about the causality between CO2 emission and global warning - on this point the jury is still out IMHO, and any conclusions will be premature (and therefore based more on belief).
Castle and Hendersons objections is described in this article in the Economist.
IPCC is the "Intergovenmental Panel on Climate Change" and describes it self as:
WMO is the "World Meteorological Organization" a "United Nations Specialized Agency" (see here ).
UNEP is the "United Nations Environment Programme" (se here ).
The crucial assumption that the earth will become 6 degress warmer within the next century probably stems from a IPCC study. But the IPCC study is being disputed - mainly for grossly overestimating the 3rd world growth. And with a more reasonable estimate of the economic growth, the resulting CO2 emission and therefore also the resulting global warming will be substantially lower.
4
See for instance here: http://www.kuro5hin.org/print/2003/2/17/15110/519
Being a little over average recommeded weigth I use one to calculate the energy contribution from fat in foods in the supermarket. It is to nifty little paperwheels with logaritmic scales joined by a clip, and given away for free. If it wasn't for my granddad giving me a sliderule back when I was playing with Lego for fun, I wouldn't have recogniced this lowtech paper-wheel as a modern incarnation of a sliderule.
So, even now people who wouldn't know a logaritm even if they saw one, are using sliderule for everyday lowtech tasks.