Since the 'scientific process' (not the outcome) can contain subjective elements one can state that 'the wisdom of groups' (aka concensus) can hinder it. I would think that especially revolutionary thoughts are hindered by this mechanism.
PS the story is _exactly_ the reason why I did not choose to become a researcher.
PS2 Currently, universities (in the Netherlands) are more school than university.
I second that. And mod the guy up for being extremely insightful.
I do have some wishes for your plan9-like VM: I would like to do software partitioning with namespaces (some memory adresses map on VM/system/resource X, other on Y, some opcodes executed on resource X other on Y) . Second: I would like late namespace determination, driven by the application. I'm no plan9 expert so maybe it can be done already.
Some people keep wondering what we can do with all those 1000's of cores on our future processors. Now we know.
It would be bad to _not_ know C. It's like driving a car and not knowing about car mechanics. Maybe it won't be your last or most used language. I would be weird to only know perl, java or c++. And while you are at it somewhere in your future, start with it so you won't 'doh' yourself too much. The only drawback is that it won't learn you good style unless you get a good course/book (c can get ugly).
My final goal would be to know a half a dozen languages and be prepared to learn a few of the other two dozen. Call it job security/flexibility. If that is your goal, it is good to understand the hardware (since all programming languages cannot do more than the hardware allows). C is pretty close to the hardware and gives you a better feeling for what hardware can do (and I can't think of alternatives that are a better choice). Although I love assembler, I cannot recommend it. I can recommend to learn about compilers though.
Plugging myself (again): I wrote a small research article about heterogeneous architectures which proves there is a method to determine which part of the job requires fast processing (either raw speed or functional optimalization by the processor). It required the job to be partitioned (to Kahn process networks) and profiled but you can think of more dynamic optimalisations.
By the way: priority is not a term to use. Its premise is that there is a shortage of processors to schedule on. If you have a heterogenenous environment with a lot of processors, you never have a shortage (just a lot of bad choices to make).
Even if they have valid research, they must have some assumptions to arrive at that conclusion. Since they have a vested interest, those assumptions will always be the subject of conflict. That makes your point irrelevant (it doesn't make the research irrelevant though: big difference).
Therefore his question is valid: how do others value the 'Linux ecosystem'.
Fucking my mother messed up your brain. Did you get tested? [/cynicism]: i'll bite.
Since you don't know nothing about the OP's situation it is hard to discuss ethics and drawing valid conclusions.
You let yourself get angry without thinking if that is the best response. What do you gain by calling him names? It won't make him reconsider his job. What would you want the OP to do? Is this the best way to achieve that goal? And no, your previous answer doesn't count.
No: thank him. He could have shut up. Maybe the ethics are questionable but at least he provides the public with the situation at hand (so they too can form an opinion).
That depends on the number of users / server. Suppose a server costs 2000 dollar/year to run. You would need 2000/50 = 40 concurrent users/year (excluding managing the systems).
Well, 40 is not a lot. A decent application can service 1000 users/system. Maybe a user is online for 8 hours a day. This means you can service 3000 users per server. Maybe you need a db server and a failover ratio of 2. That still serves you 750 users/server on average.
Suppose you have a 90% ratio of demo to payed subscriptions. That still earns you 35*50 dollar per server (=1750). Decent margin. The real question becomes: can you manage your servers efficiently?
1) Did I forget to mention that it mandatory to use a mode-S transponder here? It is coupled to an aircraft registration.
2) The transponders are mandatory, even if not flying under ATC. As I mentioned, it would be nice to receive back the radar information.
3) True, however most IFR here flies above the transition level. That means they use flightlevels instead of QNH. It would be nice to just know what is around (even if the info is not 100% accurate).
4) True, but this is the case with all electronics in an aircraft (the pilot has a master power switch). In the NL a transponder is more or less mandatory (so you are not allowed to turn it off). All aircraft types, including balloons. I think there is an exception for parachutists and maybe paragliders.
Drivers are the real terrorists of America/sarcasm
But you are completely right. If all anti-terror-efforts were put in the elimination of deaths in society (medical treatments, road safety, psychological treatments) then we could save many more people.
Cruise control is done by a computer too. All throttle related issues in my engine are done by engine management system. The only way I can intervene is via the clutch (but I've never heard of a problem with those 'computers').
It is called transponder. It is mandatory in larger parts of Holland above 2500ft (even for gliders).
However, it would be nice to be able to receive the radarscope (without having a radar on board). ATC would be able to provide that on another frequency.
Slashdot Mirror
Since the 'scientific process' (not the outcome) can contain subjective elements one can state that 'the wisdom of groups' (aka concensus) can hinder it. I would think that especially revolutionary thoughts are hindered by this mechanism.
PS the story is _exactly_ the reason why I did not choose to become a researcher.
PS2 Currently, universities (in the Netherlands) are more school than university.
would the gpl be enforcable?
I second that. And mod the guy up for being extremely insightful.
I do have some wishes for your plan9-like VM: I would like to do software partitioning with namespaces (some memory adresses map on VM/system/resource X, other on Y, some opcodes executed on resource X other on Y) . Second: I would like late namespace determination, driven by the application. I'm no plan9 expert so maybe it can be done already.
Some people keep wondering what we can do with all those 1000's of cores on our future processors. Now we know.
It would be bad to _not_ know C. It's like driving a car and not knowing about car mechanics. Maybe it won't be your last or most used language. I would be weird to only know perl, java or c++. And while you are at it somewhere in your future, start with it so you won't 'doh' yourself too much. The only drawback is that it won't learn you good style unless you get a good course/book (c can get ugly).
My final goal would be to know a half a dozen languages and be prepared to learn a few of the other two dozen. Call it job security/flexibility. If that is your goal, it is good to understand the hardware (since all programming languages cannot do more than the hardware allows). C is pretty close to the hardware and gives you a better feeling for what hardware can do (and I can't think of alternatives that are a better choice). Although I love assembler, I cannot recommend it. I can recommend to learn about compilers though.
Plugging myself (again): I wrote a small research article about heterogeneous architectures which proves there is a method to determine which part of the job requires fast processing (either raw speed or functional optimalization by the processor). It required the job to be partitioned (to Kahn process networks) and profiled but you can think of more dynamic optimalisations.
By the way: priority is not a term to use. Its premise is that there is a shortage of processors to schedule on. If you have a heterogenenous environment with a lot of processors, you never have a shortage (just a lot of bad choices to make).
Atlassian got a pretty big userbase...
Well, they lost the battle so please convert to the next standard (and I like emacs).
Respect is earned, not enforced.
true.
You must have forgotten to check slashdot itself where plenty of comments contain the words 'vendor lockin'.
I second that. I work at such a place too: 25% profit margin on a 'cheap' product (for European standards).
True, but I think the point is that it didn't restrict the designer in question.
Yeah, like google... Oh wait, they were not the first. Or
Windows, wait... wrong again. Or
Gopher, ... [/cynicism]
The first one gets their customers quickly but these people are also the early adaptors of any better system.
I had the same problem with java 6. I had to add some lines to my eclipse.ini:
-XX:CompileCommand=exclude,org/eclipse/core/internal/dtree/DataTreeNode,forwardDeltaWith
-XX:CompileCommand=exclude,org/eclipse/jdt/internal/compiler/lookup/ParameterizedMethodBinding,
Google some more if you want to find out why (I forgot).
Yeah!
And they use the sun and a magnifying glass to burn their DVD's.
For BluRay you need to filter the light with a prism.
Even if they have valid research, they must have some assumptions to arrive at that conclusion. Since they have a vested interest, those assumptions will always be the subject of conflict. That makes your point irrelevant (it doesn't make the research irrelevant though: big difference).
Therefore his question is valid: how do others value the 'Linux ecosystem'.
That is how I read it.
He does have a point: they have a vested interest.
For that, you have the http://en.wikipedia.org/wiki/MQ-9_Reaper#Specifications
Fucking my mother messed up your brain. Did you get tested? [/cynicism]: i'll bite.
Since you don't know nothing about the OP's situation it is hard to discuss ethics and drawing valid conclusions.
You let yourself get angry without thinking if that is the best response. What do you gain by calling him names? It won't make him reconsider his job. What would you want the OP to do? Is this the best way to achieve that goal? And no, your previous answer doesn't count.
No: thank him. He could have shut up. Maybe the ethics are questionable but at least he provides the public with the situation at hand (so they too can form an opinion).
That depends on the number of users / server. Suppose a server costs 2000 dollar/year to run. You would need 2000/50 = 40 concurrent users/year (excluding managing the systems).
Well, 40 is not a lot. A decent application can service 1000 users/system. Maybe a user is online for 8 hours a day. This means you can service 3000 users per server. Maybe you need a db server and a failover ratio of 2. That still serves you 750 users/server on average.
Suppose you have a 90% ratio of demo to payed subscriptions. That still earns you 35*50 dollar per server (=1750). Decent margin. The real question becomes: can you manage your servers efficiently?
1) Did I forget to mention that it mandatory to use a mode-S transponder here? It is coupled to an aircraft registration.
2) The transponders are mandatory, even if not flying under ATC. As I mentioned, it would be nice to receive back the radar information.
3) True, however most IFR here flies above the transition level. That means they use flightlevels instead of QNH. It would be nice to just know what is around (even if the info is not 100% accurate).
4) True, but this is the case with all electronics in an aircraft (the pilot has a master power switch). In the NL a transponder is more or less mandatory (so you are not allowed to turn it off). All aircraft types, including balloons. I think there is an exception for parachutists and maybe paragliders.
Drivers are the real terrorists of America /sarcasm
But you are completely right. If all anti-terror-efforts were put in the elimination of deaths in society (medical treatments, road safety, psychological treatments) then we could save many more people.
You also need a response that takes care of cars that break down (wheel breaking off, flat tire, cargo that falls off...).
For the rest, I agree (and I've given it a lot of thought too, driving 2/3 hours a day).
Cruise control is done by a computer too. All throttle related issues in my engine are done by engine management system. The only way I can intervene is via the clutch (but I've never heard of a problem with those 'computers').
It is called transponder. It is mandatory in larger parts of Holland above 2500ft (even for gliders).
However, it would be nice to be able to receive the radarscope (without having a radar on board). ATC would be able to provide that on another frequency.