The interesting thing is that many Americans think the same way (I _need_ a big car), however compare an American suburb to a European one. There isn't much of a difference. So why on earth would you need a SUV to drive to the mall while us euros do fine with our Peugeot or BMW?
You need to haul plywood around. So do we. But we tend to rent a trailer, which is always possible at our Home Depot-lookalikes. And honestly, it's not like you need to haul plywood around every fracking day. The same if we to move: you can always rent a van. Hell o' a lot cheaper, and works better.
The bottom line is that the big American car manufacturers hit upon the SUV-goldmine ten years ago (before that, it's not like everyone in the 'burbs had a pickup). Despite all the disadvantages (awful fuel consumption, a higher center-of-gravity causing an increased risk of tipping over, lack of close-proximity sight due to increased elevation) the SUV was successful because Americans thought they needed one. Afterwards they try to validate their purchase with (IMHO unsound) arguments.
Feel free to mod me down for this, but us euros frown upon these huge vehicles near our schools and children (think of the children!). They might make you feel safer, but the little ones aren't. If you live in the middle of nowhere, drive through the mud all day, sure you have plenty of reason to buy a big truck. But not if you live in a suburb and hit Home Depot twice a year. Get real.
(For the record: I know two people who have a Commander and a Grand Cherokee. I've driven in them, and those SUVs scare the shit out of me)
I've been using Trac for a couple of years now, also for external projects where customers directly submit tickets. It was a bit rough around the edges, but over the last year a number of good improvements have gone in without compromising the ease-of-use.
This is a ticketing system where people don't need an hour of training to understand what's going on. The integrated wiki allows for note-keeping (I tend to use it for tracking my hours too), and the SVN-browser is simple but useful.
If you need something more structured, try some of the CRM-packages. However I've been in your shoes, tried Trac and haven't looked back.
SugarCRM has been high on my list of projects claiming to be "open source" when they really weren't (according to the OSI-definition atleast, feel free to flame me that OSI has nothing to do with this). Free Distribution is #1 in the open source definition, and projects like SugarCRM forbid this (for instance for commercial purposes).
If even their own community started complaining, then it's about time to either go open or go proprietary. Projects that hang in-between just muddle the waters.
The Slashdot title for this article is flamebait, at best. Come on editors, start doing your job. Nowhere does Con mention that "Linux has failed on the Desktop".
What he does mention is much more interesting. The problems of measuring UI-responsiveness, the ivory tower that is kernel development and lkml, and the fact that our computers have gotten slower (for desktop-users) over the last 10 years. A great read and an interesting look into those ck-patches everyone seemed to be using a couple of years ago.
Thanks Con for all that work you put in, and I hope you find Japanese more fulfilling:)
Also, please don't suggest WEP/WPA, because distributing a password/passkey amoung that number of users is as good as not having one at all.
It's not like there are better protocols with proper authentication servers that could be used together with your ID and password to only allow access after you authenticated. This would stop anything but 802.1X-traffic from being accepted at the AP, and no client is able to flood the internal network.
Win XP and OS X have 802.1X support built in. Heck, there's even a Linux client (xsupplicant). Handing out IP addresses to unauthenticated clients is silly in such an environment.
In any case he's done the FOSS project, he's demonstrated he has a genuine interest in programming, additional FOSS work will not make him any more attractive of a candidate.
I understand your point of view, but I still think you're wrong on this. Better grades won't land you a better job (I don't even have my grades on mine, an interviewer doesn't care about my 20X-Calculus grade). Working on a FLOSS-project _and extending it_ (presentations, articles you wrote, conferences) to show you are capable of more than programming, that is something interesting. It shows that you are capable of more than some code-monkey's job and that you are willing to go the extra mile.
Naturally they don't care about a specific project, I wouldn't expect them to, however the interviewer needs something more than good grades in order to figure out if you are made of the right material. And if you can't even talk in non-technical terms about a project that you put a lot of spare time in, well that also says something about the interviewee.
IMHO a good interviewer tries to judge someones character, ability to communicate and technical competence. What you do in your spare time, and your ability to describe that in lay-man terms, helps the interviewer judge those first two at sets you apart from the rest of the crowd. Technical competence? Well, you graduated didn't you?
Unless your job has nothing to do with software development I'd drop the open source project.
I'd advise on doing something you enjoy next to all those boring classes, and certainly not drop something you are enjoying in order to focus solely on those grades. Any hobby is potentially interesting during an interview, as you simply don't know who is sitting across the table.
Recently had an interview at IBM, the manager doing the interview was very interested in my research/publications and work experience (my company), but the couple of open source projects on page two got quite a bit of attention too. It shows that you have a technical interest, and are willing to put in your spare time to complement that part-time code job with something you enjoy. And a presentation at a FLOSS-conference goes a long way, even if you are only having fun on a small niche project.
If you are looking for a dull job, don't do anything besides those courses and work. If you are interested in a truly interesting job, spice up that resume with side-projects (commercial or not), presentations, publications and hobbies. You never know which one might trigger an interesting conversation, in which the interviewer can get to know you better.
Everyone knows that the US government has been spying on other countries for years, including a couple of high-profile industrial espionage cases. If I pulled the strings in France, I would make proper encryption mandatory and not let a single byte escape from government computers/handhelds to the internet.
... and these surprise me. I've been seeing a lot of "Open Source" products that forbid redistribution, or only allow their source to be used for non-commercial projects. This is simply incorrect/dishonest marketing, but unfortunately the OSI can't go after these people because the term "Open Source" isn't trademarked. Microsoft can call their "Shared Source" license Open Source, and we couldn't do a thing.
The problem here is ambiguity of the term "Open Source", similar to equating "Free Software" to freeware. Lets just ditch both, and call our software "Freedom Software" instead. Should go fine next to those Freedom Fries.
1. Yes, you can do this with a Linux distribution. Note that you are technically selling a Linux distribution (with hardware), so you'd have to either include the distribution sources, or offer them online, or include an offer to supply any source for any GPL'ed binary in your distribution. And Linux probably won't go GPLv3 anytime soon.
2. Legally using a statically (or shared!) linked library with your own code creates a 'combined work'. If you use GPL'ed libraries (they exist), you have to supply the source code to the whole combined work. The LGPL only states that you have to supply the source code of the library you used, not the whole combined work. The LGPL doesn't differ between statically or shared linking, which is a common misconception. See http://www.gnu.org/licenses/lgpl.html for details.
3. Sure you can obfuscate your code. Write it in Perl, and you get the obfuscation for free.
4. No, you won't be forced to publish your code by some 3-d party. Not unless you use a GPL'ed library or code in your combined work, or make changes to GPL'ed code.
5. It depends on what you see as a problem. It is true that you would be able to distribute your hard+software without having to supply any source code if you only use BSD licensed binaries and libraries, but as long as you use BSD/LGPL-licensed libraries you wouldn't have to supply the source code of your own work.
And don't be put-off by the Linux fanboys. Kudos to the people who are trying to make the free software model work for their business and play by the rules.
I've come across MaNGOS a few times. Have people forgotten the bnetd-drama this soon? I assume MaNGOS has adopted a low-profile because of this, Blizzard/Vivendi would stomp them out as quick as you can say "Zug-Zug".
BTW, an incompatible EULA for a GPL-project? Yikes, and small chance it'll stop the onslaught of Vivendi lawyers. We'll see...
The problem is that you can't parallelize everything equally well. Sure, non-professional programmers are able to write perfectly valid MPI code, but I'll bet you that their problem was trivially parallel to start off with (that is: you can split your problem space into independent chunks and have each processor deal with its own section of the problem). This is the case for a lot of problems, and can lead to a dramatic speedup.
When you get into more complex problems (like search algorithms) MPI (or any other method of PP) gets increasingly hard to parallelize properly. Scaling into 100s of processors isn't uncommon, and communication adds up exponentially the further you go, possibly griding your program to a halt.
Having said that, I've seen some amazing code by physicists trying to squeeze every last cycle out of their code. The problem is that they normally don't see pitfalls (deadlocks etc) until their code goes up in flames, but yes physicists are smart people and PP isn't rocket science, no matter what Intel or the like want you to believe.
I've read the Forbes article and it doesn't sound like Facebook will become a MySpace soon... but still, do we still care for all these proprietary social networks merely striving to lock-in users? Facebook, MySpace, LinkedIn, Orkut, Friendster, localized networks like Hyves (big over here in.nl)... they all want more eyeballs (== more $$$), it's just that their way to generate more pageviews is different.
Maybe I'm just getting old (queue the GOML-quotes), but I don't see the point in investing time into developing for Facebook. In two or three years, a new YASN will have all the kid's attention. Although you might get a short-term boost in eyeballs now by developing using this API, the drawback is that once Facebook goes south you will be sucked down with it. eBay has been here for >10 years and will still be here over 10 years because it has critical mass and doesn't depend on a third party to survive.
Now an open meta-YASN engine, which gives me one page/feed with all my contacts over all these networks, that might be an improvement.
PS: On a side note, you discuss processing power and implications to AI. Do you know of any titles that would be limited in AI due to the Wii's processing? I can't think of any, and always had my doubts about AI being a bullet behind the pros of better hardware. Better theoretical AI maybe:-P
All I know is that the AI in Wii Sports Tennis still beats the crap out of me. Probably has more to do with me sucking, though.
But seriously, AI in games indeed isn't where all the cycles go into. It's the graphics. Building an engine capable of rendering an additional quadrillion vertexes will always be an easier way to waste extra cycles than coming up with AI that seems remotely intelligent. The Wii proves that nicer graphics != more fun, no matter what your clockspeed is.
It's hard to blame people for something they don't understand if they didn't grow up with it. For us these are trivial concepts, but there are still millions of 50+ people who have never browsed a website.
Certain people actively investigate new appliances, others are busy enough as is that they simply don't care. One of my grandma's has vowed never to touch a computer, even though the family is full of IT-nerds. My other grandma is IM'ing more than I am.
The best thing a judge can do in such a case is acknowledge he doesn't understand the topic, and preferably have a different judge that does know the topic take over, regardless of the topic.
I'm all for the 'Church for the destruction of the Church of Scientology'. Fight fire with fire. Anyone who doesn't agree that the CoS is an evil pyramid-scheme should be sentenced to death (or an atomic wedgie).
We live in a world where we honor, and support the honoring of, intellectual property. We play by the same rules as the rest of the business. Recently it has come to our attention that certain products of Microsoft (among which Windows XP, Vista and Microsoft Office) infringe on no fewer than 236 software patents.
We won't identify the patents, by which you might attempt to avoid our insane software patent licensing costs and Armageddon. We'll just scare everyone into never doing business with you ever again.
I completely agree with you on this one. After trying Trac out a few years ago, I haven't looked back. Besides for personal and university projects, I've been using separate Trac-installs for clients. Everything from support requests to development projects end up as tickets, with wiki pages for additional details and background information.
The main reason why Trac is so successful is indeed the flat learning curve / simple interface. Two sentences along the lines of "If you add a new ticket here, I'll get right on it" are enough. Clients happy because they can easily bug me and have an overview of past tickets, I'm happy because of the decrease of postit's on my monitor.
Oh, Bugzilla you ask? I've had to deal with it in the past for a few OSS projects. I'm still scared of it, and wouldn't let my clients near it even if they begged me. It might be more useful when you have hundreds of technical users working on a single project with needs like reporting and time tracking, but for anything less a more flexible alternative is preferable, IMHO.
While certainly no encryption is unbreakable, I still believe I fare better odds placing my data on hard wires, where individuals would have to be specifically targetting it, rather than letting it flow free and open into the air for all to capture and (attempt to) abuse
Don't forget about Van Eck phreaking. Even by using a computer monitor, you are already sending out data than can be captured without having to tamper with any wires.
The EU is surprisingly un-democratic, actually. There is an elected EU-parliament, but they have no divisive say on anything. The EU-officials making the calls are pushed forward by the member-nations.
The main problem (if you can call it one) is that the EU isn't a country. You may take your share of blame for your NiC because your countrymen put him in power. Who should we blame for a failing EU? People from Poland, Romania or Malta? They don't speak my language, don't read or watch the same media, can't vote on the same parties. There is no such thing as a 'European'.
The only similarity is that we use the same currency, it ends at that. The EU was founded as an economic union, and became a very successful one. The problem is that it went on to other facets of life, and became a bludgeoning bureaucratic monstrosity with failures like the new Airbus and Galileo.
This is the main reason why we Dutchies, together with the French, voted against the EU 'constitution'. We're perfectly happy without the EU wanting to assimilate us into a 'federation'.
Slashdot Mirror
Minix!
I'll bite!
The interesting thing is that many Americans think the same way (I _need_ a big car), however compare an American suburb to a European one. There isn't much of a difference. So why on earth would you need a SUV to drive to the mall while us euros do fine with our Peugeot or BMW?
You need to haul plywood around. So do we. But we tend to rent a trailer, which is always possible at our Home Depot-lookalikes. And honestly, it's not like you need to haul plywood around every fracking day. The same if we to move: you can always rent a van. Hell o' a lot cheaper, and works better.
The bottom line is that the big American car manufacturers hit upon the SUV-goldmine ten years ago (before that, it's not like everyone in the 'burbs had a pickup). Despite all the disadvantages (awful fuel consumption, a higher center-of-gravity causing an increased risk of tipping over, lack of close-proximity sight due to increased elevation) the SUV was successful because Americans thought they needed one. Afterwards they try to validate their purchase with (IMHO unsound) arguments.
Feel free to mod me down for this, but us euros frown upon these huge vehicles near our schools and children (think of the children!). They might make you feel safer, but the little ones aren't. If you live in the middle of nowhere, drive through the mud all day, sure you have plenty of reason to buy a big truck. But not if you live in a suburb and hit Home Depot twice a year. Get real.
(For the record: I know two people who have a Commander and a Grand Cherokee. I've driven in them, and those SUVs scare the shit out of me)
I've been using Trac for a couple of years now, also for external projects where customers directly submit tickets. It was a bit rough around the edges, but over the last year a number of good improvements have gone in without compromising the ease-of-use.
This is a ticketing system where people don't need an hour of training to understand what's going on. The integrated wiki allows for note-keeping (I tend to use it for tracking my hours too), and the SVN-browser is simple but useful.
If you need something more structured, try some of the CRM-packages. However I've been in your shoes, tried Trac and haven't looked back.
SugarCRM has been high on my list of projects claiming to be "open source" when they really weren't (according to the OSI-definition atleast, feel free to flame me that OSI has nothing to do with this). Free Distribution is #1 in the open source definition, and projects like SugarCRM forbid this (for instance for commercial purposes).
If even their own community started complaining, then it's about time to either go open or go proprietary. Projects that hang in-between just muddle the waters.
What he does mention is much more interesting. The problems of measuring UI-responsiveness, the ivory tower that is kernel development and lkml, and the fact that our computers have gotten slower (for desktop-users) over the last 10 years. A great read and an interesting look into those ck-patches everyone seemed to be using a couple of years ago.
Thanks Con for all that work you put in, and I hope you find Japanese more fulfilling
It's not like there are better protocols with proper authentication servers that could be used together with your ID and password to only allow access after you authenticated. This would stop anything but 802.1X-traffic from being accepted at the AP, and no client is able to flood the internal network.
Win XP and OS X have 802.1X support built in. Heck, there's even a Linux client (xsupplicant). Handing out IP addresses to unauthenticated clients is silly in such an environment.
I understand your point of view, but I still think you're wrong on this. Better grades won't land you a better job (I don't even have my grades on mine, an interviewer doesn't care about my 20X-Calculus grade). Working on a FLOSS-project _and extending it_ (presentations, articles you wrote, conferences) to show you are capable of more than programming, that is something interesting. It shows that you are capable of more than some code-monkey's job and that you are willing to go the extra mile.
Naturally they don't care about a specific project, I wouldn't expect them to, however the interviewer needs something more than good grades in order to figure out if you are made of the right material. And if you can't even talk in non-technical terms about a project that you put a lot of spare time in, well that also says something about the interviewee.
IMHO a good interviewer tries to judge someones character, ability to communicate and technical competence. What you do in your spare time, and your ability to describe that in lay-man terms, helps the interviewer judge those first two at sets you apart from the rest of the crowd. Technical competence? Well, you graduated didn't you?
I'd advise on doing something you enjoy next to all those boring classes, and certainly not drop something you are enjoying in order to focus solely on those grades. Any hobby is potentially interesting during an interview, as you simply don't know who is sitting across the table.
Recently had an interview at IBM, the manager doing the interview was very interested in my research/publications and work experience (my company), but the couple of open source projects on page two got quite a bit of attention too. It shows that you have a technical interest, and are willing to put in your spare time to complement that part-time code job with something you enjoy. And a presentation at a FLOSS-conference goes a long way, even if you are only having fun on a small niche project.
If you are looking for a dull job, don't do anything besides those courses and work. If you are interested in a truly interesting job, spice up that resume with side-projects (commercial or not), presentations, publications and hobbies. You never know which one might trigger an interesting conversation, in which the interviewer can get to know you better.
All work and no play makes Jack a dull boy.
Everyone knows that the US government has been spying on other countries for years, including a couple of high-profile industrial espionage cases. If I pulled the strings in France, I would make proper encryption mandatory and not let a single byte escape from government computers/handhelds to the internet.
*reinforces his tinfoil hat with lead*
... and these surprise me. I've been seeing a lot of "Open Source" products that forbid redistribution, or only allow their source to be used for non-commercial projects. This is simply incorrect/dishonest marketing, but unfortunately the OSI can't go after these people because the term "Open Source" isn't trademarked. Microsoft can call their "Shared Source" license Open Source, and we couldn't do a thing.
The problem here is ambiguity of the term "Open Source", similar to equating "Free Software" to freeware. Lets just ditch both, and call our software "Freedom Software" instead. Should go fine next to those Freedom Fries.
1. Yes, you can do this with a Linux distribution. Note that you are technically selling a Linux distribution (with hardware), so you'd have to either include the distribution sources, or offer them online, or include an offer to supply any source for any GPL'ed binary in your distribution. And Linux probably won't go GPLv3 anytime soon.
2. Legally using a statically (or shared!) linked library with your own code creates a 'combined work'. If you use GPL'ed libraries (they exist), you have to supply the source code to the whole combined work. The LGPL only states that you have to supply the source code of the library you used, not the whole combined work. The LGPL doesn't differ between statically or shared linking, which is a common misconception. See http://www.gnu.org/licenses/lgpl.html for details.
3. Sure you can obfuscate your code. Write it in Perl, and you get the obfuscation for free.
4. No, you won't be forced to publish your code by some 3-d party. Not unless you use a GPL'ed library or code in your combined work, or make changes to GPL'ed code.
5. It depends on what you see as a problem. It is true that you would be able to distribute your hard+software without having to supply any source code if you only use BSD licensed binaries and libraries, but as long as you use BSD/LGPL-licensed libraries you wouldn't have to supply the source code of your own work.
And don't be put-off by the Linux fanboys. Kudos to the people who are trying to make the free software model work for their business and play by the rules.
And everyone who has read the Ender series knows that FTL/Ansible communications leads to FTL transportation, given a large enough "computer" :)
Clusty does something similar. Searching for "Apple" will show categories for OSX and fruit, for instance.
Seems like that is the idea. Would be very neat, no longer any hassle with memory-card readers supporting a zillion different standards.
BTW, an incompatible EULA for a GPL-project? Yikes, and small chance it'll stop the onslaught of Vivendi lawyers. We'll see...
The problem is that you can't parallelize everything equally well. Sure, non-professional programmers are able to write perfectly valid MPI code, but I'll bet you that their problem was trivially parallel to start off with (that is: you can split your problem space into independent chunks and have each processor deal with its own section of the problem). This is the case for a lot of problems, and can lead to a dramatic speedup.
When you get into more complex problems (like search algorithms) MPI (or any other method of PP) gets increasingly hard to parallelize properly. Scaling into 100s of processors isn't uncommon, and communication adds up exponentially the further you go, possibly griding your program to a halt.
Having said that, I've seen some amazing code by physicists trying to squeeze every last cycle out of their code. The problem is that they normally don't see pitfalls (deadlocks etc) until their code goes up in flames, but yes physicists are smart people and PP isn't rocket science, no matter what Intel or the like want you to believe.
I've read the Forbes article and it doesn't sound like Facebook will become a MySpace soon... but still, do we still care for all these proprietary social networks merely striving to lock-in users? Facebook, MySpace, LinkedIn, Orkut, Friendster, localized networks like Hyves (big over here in .nl)... they all want more eyeballs (== more $$$), it's just that their way to generate more pageviews is different.
Maybe I'm just getting old (queue the GOML-quotes), but I don't see the point in investing time into developing for Facebook. In two or three years, a new YASN will have all the kid's attention. Although you might get a short-term boost in eyeballs now by developing using this API, the drawback is that once Facebook goes south you will be sucked down with it. eBay has been here for >10 years and will still be here over 10 years because it has critical mass and doesn't depend on a third party to survive.
Now an open meta-YASN engine, which gives me one page/feed with all my contacts over all these networks, that might be an improvement.
All I know is that the AI in Wii Sports Tennis still beats the crap out of me. Probably has more to do with me sucking, though.
But seriously, AI in games indeed isn't where all the cycles go into. It's the graphics. Building an engine capable of rendering an additional quadrillion vertexes will always be an easier way to waste extra cycles than coming up with AI that seems remotely intelligent. The Wii proves that nicer graphics != more fun, no matter what your clockspeed is.
Certain people actively investigate new appliances, others are busy enough as is that they simply don't care. One of my grandma's has vowed never to touch a computer, even though the family is full of IT-nerds. My other grandma is IM'ing more than I am.
The best thing a judge can do in such a case is acknowledge he doesn't understand the topic, and preferably have a different judge that does know the topic take over, regardless of the topic.
I'm all for the 'Church for the destruction of the Church of Scientology'. Fight fire with fire. Anyone who doesn't agree that the CoS is an evil pyramid-scheme should be sentenced to death (or an atomic wedgie).
All hail lord Xenu!
Dear Microsoft,
We live in a world where we honor, and support the honoring of, intellectual property. We play by the same rules as the rest of the business. Recently it has come to our attention that certain products of Microsoft (among which Windows XP, Vista and Microsoft Office) infringe on no fewer than 236 software patents.
We won't identify the patents, by which you might attempt to avoid our insane software patent licensing costs and Armageddon. We'll just scare everyone into never doing business with you ever again.
Sincerely,
Darl McBride, SCO
No. Next question.
I completely agree with you on this one. After trying Trac out a few years ago, I haven't looked back. Besides for personal and university projects, I've been using separate Trac-installs for clients. Everything from support requests to development projects end up as tickets, with wiki pages for additional details and background information.
The main reason why Trac is so successful is indeed the flat learning curve / simple interface. Two sentences along the lines of "If you add a new ticket here, I'll get right on it" are enough. Clients happy because they can easily bug me and have an overview of past tickets, I'm happy because of the decrease of postit's on my monitor.
Oh, Bugzilla you ask? I've had to deal with it in the past for a few OSS projects. I'm still scared of it, and wouldn't let my clients near it even if they begged me. It might be more useful when you have hundreds of technical users working on a single project with needs like reporting and time tracking, but for anything less a more flexible alternative is preferable, IMHO.
*runs to grab his tinfoil hat*
The EU is surprisingly un-democratic, actually. There is an elected EU-parliament, but they have no divisive say on anything. The EU-officials making the calls are pushed forward by the member-nations.
The main problem (if you can call it one) is that the EU isn't a country. You may take your share of blame for your NiC because your countrymen put him in power. Who should we blame for a failing EU? People from Poland, Romania or Malta? They don't speak my language, don't read or watch the same media, can't vote on the same parties. There is no such thing as a 'European'.
The only similarity is that we use the same currency, it ends at that. The EU was founded as an economic union, and became a very successful one. The problem is that it went on to other facets of life, and became a bludgeoning bureaucratic monstrosity with failures like the new Airbus and Galileo.
This is the main reason why we Dutchies, together with the French, voted against the EU 'constitution'. We're perfectly happy without the EU wanting to assimilate us into a 'federation'.