I know this belongs as a reply, but there are some 15-20 posters who insist that a judicial review will save basic rights...
Some problems with that:
1. The judge only hears one side, because the prospective "defendant" doesn't know that they're being probed.
2. Check the statistics on judicial decisions and wiretaps/cell taps for terrorism. if you can find a *single* one that was ever turned down, I'd be mightily surprised.
3. Who wants to be the judge that lets another supposed Mohammed Atta through, even at the cost of millions of conversations like "I'm late, on the road, be there in ten minutes"?
1. It's not some wonky, secure, stripped, stable *nix box that can't run GUI IDEs.
2. It can give me multicolored colored visual cues to code, so monochrome terminal sessions are out. Colored terminals are OK, but they lack the same elegance of a full GUI.
3. It can quickly deal with running those additional programs I need when reading laughable client 'specs' in PowerPoint, MS word, HTML email from hell, etc.
4. The code, on my GUI-heavy isolated dev box, can run identically to the code deployed on the stripped, burly, boxen.
5. If I need to test speeds without a GUI, I can ignore it, and use the box in a stripped, clean, mode.
6. It must be stable without being cold and wooden (Red flag of personal preference, I find almost all *nix platforms to be far less comfortable than they could potentially be. I know the KDE/Gnome/Solaris folks are working on it, and have their advantages as well)
7. It must support additional "development necessary tools" such as playing mp3's, accounting for my time with professional accounting software, play mpg's, and run the occasional "break tool" in the form of some game that isn't 5 (or more) years old.
8. It must support running javascript, and be able to test IE, NN, *and* lynx, and be able to run MacOS 9, MacOS X, X windows (and sub-managers), as well as MS windows (and their many variants).
9. The hardware and software should need my personal intervention for tweaking and updates, well, almost never. I am not paid to update my box, I am paid for writing original code.
I run a business, and I use (deep breath) LinuxPPC, Yellow Dog Linux, SuSe (on X86 and PPC), Debian (on PPC and X86), Mac OSX, RedHatX86 (four versions), OpenBSD (PPC and X86), FreeBSD (X86), SunOS (really, some clients still use it), Solaris (all 'of the flock, ugly), Win 3.1, Win 95, Win 98, Win ME, Win NT 3.x, Win NT 4.x, Win2K, Win XP (all).
Of all of the above OS's, which one, do you think, can actually do requirements 1-9? (There's only one, take your time...) I used to do dev work on no less than 3 different boxes a day to meet those requirements. Now I use one.
Of course, if I have a few sites with a few million hits a day (I do), I'll host it elsewhere than my test box, an OS X box.. But I'm not going to develop on that box. I'm going to develop on a box that makes me the most productive, a box where I don't care about IRQ's, drivers, optimizing window managers, running rpm or apt-get or any other time-wasting CLUI tools that interfere with writing code.
For writing code, use a box that meets *your* needs. All platforms prior to OS X meant I was using far too many comps, because I needed multi-platform, multi-client-platform, code. No other platform allows you to test as many platforms at once as PPC/OS 9/OSX on Mac.
For one, Wolfram is lobbing a mind-grenade at the ivory-tower sciences. For years, interdisciplinary communication has been neglected, with each discipline thinking that their area is "unique" or "special" somehow. His idea that "it's all much simpler than that" flies in the face of those who want to believe that their "specialized" knowledge is more unique, or valuable, than oh, a general algorithm. This will offend most scientists, who want to think that, say, astrophysics is more complex than sociology (and vice versa).
Human pride in their past achievements stifles much of new human achievement, and the proponents of major achievements were usually whackos at the time (Galileo, Copernicus, Einstein, Nash, Darwin) who questioned legions of prior thought with much simpler explanations. Ockham's razor indicates that four lines are much more likely than an infinitely complex universe, but, in the meantime, millions of scientists are working on finding more complexity, not simplifying.
He (Wolfram) also directly challenges some other older, widely held, beliefs in science, such as a "wet lab" (say, using human cells) cannot be replaced by a simulation. I find it amusing (as a computer professional), as it is akin to saying that an "actual accountant cannot be replaced by a spreadsheet" (or even a that a car mechanic cannot be replaced by a 'bot).
Many accountants grinding numbers were replaced by automations, but the design components, the accounting concepts, were still writing the code. In short, old school scientists looking for lots of "underlying algorithms" would lose job security if their discoveries were simply macros from another field, and their lab was reduced to replication of old data, rather than new discovery.
After reading both Science and Nature for many years, it starts as funny when you first see this actually happening, and then becomes pathetic.... when it takes 8 years for two fields to use the same two line algo's to describe a behavior (say, CA and Planetary formation). The current scientific mindset does not lead to a social scientist browsing physics journals or vice versa, for some reason, they seem to think that some things in the Universe are unrelated to other things in the universe (maybe the "Uni" part is ill-explained to scientists in training?)
Another notion he questions is the concept of "free-will", which has been an underpinning of western civilization since, well, western civilization. Not because it's a well proven, logical concept, but because the very concept of "self" and "identity" hinge on the ideas that somehow, a person is in control. People won't like this, as they'd rather be the master over a machine than a meat-machine. In bio-ethics, there's an entire war over using or changing out the machine components, hinging on a religious belief that there is a "soul", or something similar, that makes a meat-machine unique. If we are all four lines of code, or even 50000, we are much less "special" or "unique", we are not free will but a product of a program.
For all of those tired 42 jokes on this page, maybe I missed the point, maybe they did, I dunno. The entire 42 theme was that humanity, the planet earth, was just code. That any sufficiently complex system may have underlying simple questions, and simple answers. What those who didn't read the Adams books, it goes like this: The earth is just a computing device to give an answer in the form of a question. Nothing more. Tell this to religious authorities, goverment authorities, those who believe in a value of "will" or "life" and they will recoil. The meta code is a bit like two lines:
initialize $earth;
sleep ($limit);// I forget the limit
print question_of_meaning($earth);
With a runtime of many years, and millions of sub variables ($beer_sip_counter and so on), but a simple code starting base (as compared to the derivative results). This is no biggie, it's a self-modifying codebase ($earth varies itself). The premise that the entire universe could be a self-modifying codebase, however, flies in the face of those who want to find the static question, or a static answer. In a self-modifying codebase, both could change. Those who want a concept of "god" want something else messing with variables, those who want "meaning" want something that provides the meat-machine with meaning... and so on.
Just going this far, I know why it took him ten years. Years of self-editing to modify the above, etc.
Anyways, my four lines for a multiverse, in metacode:
while ($existance) {
matter = matter;
//int is for folks who need to declare things often, like gods!
define function multiverse ( rand(multiverse));
multiverse ($earth);
}
1. No pesky elections. Competitors or dissent will simply be strangled to death, on the spot, with jedi tricks.
2. Dissent about, well, anything can be handled by blowing up an entire planet.
3. None of those silly women in power, hey, it's a guy thing.
4. Capital punishment, as a debate, will fade, because after all, they were "evil-doers".
5. Rather than the tedium of semi-electing the children of prior leaders, we can create eleborate black costumes of life support. Shrub for president in 2341!
6. No new taxes!... er... this week.
7. Yes, the space-trains will run on time, hauling the free-thinkers and dissenters to places like Hoth for "education".
8. That goofy geneva convention thing about physical torture can be eliminated, so folks like leia can be properly brutalized.
9. The universe can be dominated by white, human, anglo-saxon, men. (Ever see a pro-empire wookie in the death star?)
10. Droids have no rights.
Basically, the guy who wrote the article is pro-authority. 40 years ago, this is what was called "facist scum" in the US, because they thought law and order was more important than bowing down and serving a happy people.... forgetting, of course, they they have no right to say what people want, or that their role was to serve the people, not be served by the people.
1. The same kind of options that 2D browsers have for UI preferences.... some people like sorting alpha, some by creation date/time, some by access date/time, some by size, etc. Rather than force one or another on users, why not let users pick a default, with a custom setting per window?
2. Storing state. Okay, I quit the app. I launch it 5 minutes later... where is my carefully crafted view? I have to rebuild it?
3. Okay, I'm looking at a platter with a set of files... but nowhere on the platter is the name. How do I know if this is tuesday's set of these files or wednesdays set? To see the name of the platter (so to speak), I have to go a level "down" or to my "platter" menu.
4. Maybe a bug, maybe a feature... go to root platter. Click on hard drive. Move out a bit, go back to root. Click on same drive. Move out a bit again. The "Platter" menu now has multiple drive listings.
5. Font selection?
6. Pulling platters to their own locations.... this may be a hard one to explain, but it would be nice if I could move a platter I used frequently "away" from the others... say, halfway across the pool.
7. (Adding on 6), even better, some custom "starting points" beyond the root. Say, two or three starting platters match my working style (/,/Users/username, and/Application...)
8. Ability to adjust Icon "density" for folders with lots of files. Imagine the dock concept, where, the more objects to represent, the smaller each object becomes.
9. Dimming/fading/darkening non-active platters... maybe increasing fog intensity for "distant" objects would do the trick?
10. Some other navigation, non-platter based. How about some keys for 'fly-around" control, so I can get an overhead look, or move around without using platters as the reference point?
Overall, a great piece of work, though. Much nicer for my working style than my usual 25 stacked windows.
If you're kind of person who really wants to know how much is in an "update" (Measly bullet points doesn't do it justice): "lsbom/Library/Receipts/MacOSXUpdate10.1.3.pkg/Contents/ Resources/MacOSXUpdate10.1.3.bom" Will produce a file listing... piped to WC, there's 1832 files/folders involved in this "litte" fix.
Classic bonehead things I did to make openldap installations and directories less-than optimal:
1. I didn't stay up on patches, so I was running versions with known performance bugs. I should have kept up.
2. I did lousy, MySQL-101-db-like, structures (flat as all get out), which made the tree and replication features somewhere between pointless and painful. I should have fanned the tree out to different servers, increasing the ways I could spread the load. I also went too far the other way, and made high-quality, relational, db structures, which made my entry count absloutely absurd. I should have tuned my design under load.
3. I tried storing large objects in a record (say, 10-100K gifs) in each record and then wondered why it too so long to move 500 of those records.... over a 56K line. I should have used pointers or references to external blobs.
4. In speed critical applications, I did non-filtering lookups (the SQL equivalent of "select * from..."), so I was slogging though millions of bits I didn't need. I should have specified the data I needed, when I needed it.
5. I indexed all sorts of things that didn't need indexing, basically wasting tons of CPU for lookups that wouldn't be needed. I should have only indexed fields that were being frequently used, for the way they were being used.
6. I tried to stored transactional information, stuff that changed every few seconds or minutes, in an LDAP implementation that wasn't designed for "row or column" transactions. I should have used the right tool for that job, such as an atomic-transaction database.
7. When I compiled, I didn't use a fast backend, I accepted a slower gdb default. I should have used a tuned, compiled, backend.
8. I ran off of slow disks. I should never have tried to use IDE for a high-speed database.
9. I didn't tune LDAP to cache as much in RAM as possible. I should have tuned the slapd.conf entries.
10. I didn't deploy ldap servers at critical points, and instead, tried to use a few big ones. A tree of smaller servers would have worked faster.
11. I wrote scripts that did _really_ dumb things, like try to read and then edit every record in a live OpenLDAP *dbm database on the fly. I should have broken it into smaller work units,and treat the protocol on its merits.
12. Rather than pull a fast, clean, backup daily (which requires a restart), I was just running it until it died. I should have pulled regular backups, and the restart helps to close any niggling descriptors left open.:-)
There's lots of ways to screw up any program.... I've killed apache with oversized log files and open fds, I've brought postgresql to it's knees by neglecting maintenance, I've brought MySQL to a dead stop because of file sizes (ugh.) and table design.... the list goes on and on.
That doesn't mean that apache, or http, sucks, or that SQL is useless, it means that tools have to be used properly, and learning the limitations of your tools is a painful step in learning how to use them.
With computer code, at least there are a some people out there who might modify your code and make it better. There is nothing similar in the music world.
Well, depending on how you provide the "source", be it sheet music, MIDI files, sample libraries, or your studio tracks, individuals can:
make re-orchestrations (Kronos Quartet re-recording Hendrix's "Purple Haze" with classical strings or Joan Baez recording the folk song "house of the rising sun" as a rock ballad)
record with a different voicing (Sinead O' Connor recording Prince's "Nothing Compares 2 U")
make new remixes to take the song in completely different directions (Autechure remixing Skinny Puppy's Killing Game)
or even start a symphony (the ultimate cover band...)
Musicians have been using eachother's "source" for much longer than coders have been using and improving eachother's code.
I've been a musician for 21 years, a coder for 16 years, and I've worked in (and with) a variety of licenses in that time. Some good, some bad, some indifferent.
I find music and code to be virtually identical in nature: It is the structured ordering, presentation and manipulation of simple primitives, to create a more complex body of work, so it makes sense to me to have similar licenses.
Is this kind of license needed for music? Definitely. Just like the coding profession, some pieces, indeed, some jobs, are entirely dependant on using free tools, be they code libraries or sample libraries. As noted, every major operating system today now ships with free (GPL, Q, BSD, or similar) tools and components, which can be copied and distributed without any royalties. Some software products have succeeded in obtaining large market $hare, some have not.
Those products that have lined the pocket of IT professionals (including mine) were built both with privately owned IP, and public (GPL, Q, BSD etc.) IP. Those that were built strictly with only private or only public IP did not become successful or profitable based on the IP they used... that's a red herring in this discussion. The nature of the license does not dictate the profitability or success of a software, or music, product.
So what is the impact, if I choose this license, to me? Well, just like the GPL work I've done, it means that others can do what they want.
Does it prevent me from releasing other work? Of course not. I can OPL/GPL some work, BSD other work, and use a restrictive EULA for other work. Since I'm not a one-hit-wonder kind of guy (I'm a professional, I "publish" a lot), this isn't a problem.
Is my ability to capture revenue altered by my choice of license, be it in code libraries, or in songs? Of course it is. But it's my choice.
Slashdot Mirror
I know this belongs as a reply, but there are some 15-20 posters who insist that a judicial review will save basic rights...
Some problems with that:
1. The judge only hears one side, because the prospective "defendant" doesn't know that they're being probed.
2. Check the statistics on judicial decisions and wiretaps/cell taps for terrorism. if you can find a *single* one that was ever turned down, I'd be mightily surprised.
3. Who wants to be the judge that lets another supposed Mohammed Atta through, even at the cost of millions of conversations like "I'm late, on the road, be there in ten minutes"?
1. It's not some wonky, secure, stripped, stable *nix box that can't run GUI IDEs.
2. It can give me multicolored colored visual cues to code, so monochrome terminal sessions are out. Colored terminals are OK, but they lack the same elegance of a full GUI.
3. It can quickly deal with running those additional programs I need when reading laughable client 'specs' in PowerPoint, MS word, HTML email from hell, etc.
4. The code, on my GUI-heavy isolated dev box, can run identically to the code deployed on the stripped, burly, boxen.
5. If I need to test speeds without a GUI, I can ignore it, and use the box in a stripped, clean, mode.
6. It must be stable without being cold and wooden (Red flag of personal preference, I find almost all *nix platforms to be far less comfortable than they could potentially be. I know the KDE/Gnome/Solaris folks are working on it, and have their advantages as well)
7. It must support additional "development necessary tools" such as playing mp3's, accounting for my time with professional accounting software, play mpg's, and run the occasional "break tool" in the form of some game that isn't 5 (or more) years old.
8. It must support running javascript, and be able to test IE, NN, *and* lynx, and be able to run MacOS 9, MacOS X, X windows (and sub-managers), as well as MS windows (and their many variants).
9. The hardware and software should need my personal intervention for tweaking and updates, well, almost never. I am not paid to update my box, I am paid for writing original code.
I run a business, and I use (deep breath) LinuxPPC, Yellow Dog Linux, SuSe (on X86 and PPC), Debian (on PPC and X86), Mac OSX, RedHatX86 (four versions), OpenBSD (PPC and X86), FreeBSD (X86), SunOS (really, some clients still use it), Solaris (all 'of the flock, ugly), Win 3.1, Win 95, Win 98, Win ME, Win NT 3.x, Win NT 4.x, Win2K, Win XP (all).
Of all of the above OS's, which one, do you think, can actually do requirements 1-9? (There's only one, take your time...) I used to do dev work on no less than 3 different boxes a day to meet those requirements. Now I use one.
Of course, if I have a few sites with a few million hits a day (I do), I'll host it elsewhere than my test box, an OS X box.. But I'm not going to develop on that box. I'm going to develop on a box that makes me the most productive, a box where I don't care about IRQ's, drivers, optimizing window managers, running rpm or apt-get or any other time-wasting CLUI tools that interfere with writing code.
For writing code, use a box that meets *your* needs. All platforms prior to OS X meant I was using far too many comps, because I needed multi-platform, multi-client-platform, code. No other platform allows you to test as many platforms at once as PPC/OS 9/OSX on Mac.
-Bop
For one, Wolfram is lobbing a mind-grenade at the ivory-tower sciences. For years, interdisciplinary communication has been neglected, with each discipline thinking that their area is "unique" or "special" somehow. His idea that "it's all much simpler than that" flies in the face of those who want to believe that their "specialized" knowledge is more unique, or valuable, than oh, a general algorithm. This will offend most scientists, who want to think that, say, astrophysics is more complex than sociology (and vice versa).
Human pride in their past achievements stifles much of new human achievement, and the proponents of major achievements were usually whackos at the time (Galileo, Copernicus, Einstein, Nash, Darwin) who questioned legions of prior thought with much simpler explanations. Ockham's razor indicates that four lines are much more likely than an infinitely complex universe, but, in the meantime, millions of scientists are working on finding more complexity, not simplifying.
He (Wolfram) also directly challenges some other older, widely held, beliefs in science, such as a "wet lab" (say, using human cells) cannot be replaced by a simulation. I find it amusing (as a computer professional), as it is akin to saying that an "actual accountant cannot be replaced by a spreadsheet" (or even a that a car mechanic cannot be replaced by a 'bot).
Many accountants grinding numbers were replaced by automations, but the design components, the accounting concepts, were still writing the code. In short, old school scientists looking for lots of "underlying algorithms" would lose job security if their discoveries were simply macros from another field, and their lab was reduced to replication of old data, rather than new discovery.
After reading both Science and Nature for many years, it starts as funny when you first see this actually happening, and then becomes pathetic.... when it takes 8 years for two fields to use the same two line algo's to describe a behavior (say, CA and Planetary formation). The current scientific mindset does not lead to a social scientist browsing physics journals or vice versa, for some reason, they seem to think that some things in the Universe are unrelated to other things in the universe (maybe the "Uni" part is ill-explained to scientists in training?)
Another notion he questions is the concept of "free-will", which has been an underpinning of western civilization since, well, western civilization. Not because it's a well proven, logical concept, but because the very concept of "self" and "identity" hinge on the ideas that somehow, a person is in control. People won't like this, as they'd rather be the master over a machine than a meat-machine. In bio-ethics, there's an entire war over using or changing out the machine components, hinging on a religious belief that there is a "soul", or something similar, that makes a meat-machine unique. If we are all four lines of code, or even 50000, we are much less "special" or "unique", we are not free will but a product of a program.
For all of those tired 42 jokes on this page, maybe I missed the point, maybe they did, I dunno. The entire 42 theme was that humanity, the planet earth, was just code. That any sufficiently complex system may have underlying simple questions, and simple answers. What those who didn't read the Adams books, it goes like this: The earth is just a computing device to give an answer in the form of a question. Nothing more. Tell this to religious authorities, goverment authorities, those who believe in a value of "will" or "life" and they will recoil. The meta code is a bit like two lines: // I forget the limit
initialize $earth;
sleep ($limit);
print question_of_meaning($earth);
With a runtime of many years, and millions of sub variables ($beer_sip_counter and so on), but a simple code starting base (as compared to the derivative results). This is no biggie, it's a self-modifying codebase ($earth varies itself). The premise that the entire universe could be a self-modifying codebase, however, flies in the face of those who want to find the static question, or a static answer. In a self-modifying codebase, both could change. Those who want a concept of "god" want something else messing with variables, those who want "meaning" want something that provides the meat-machine with meaning... and so on.
Just going this far, I know why it took him ten years. Years of self-editing to modify the above, etc.
Anyways, my four lines for a multiverse, in metacode:
while ($existance) { matter = matter;
//int is for folks who need to declare things often, like gods!
define function multiverse ( rand(multiverse));
multiverse ($earth); }
2. Dissent about, well, anything can be handled by blowing up an entire planet.
3. None of those silly women in power, hey, it's a guy thing.
4. Capital punishment, as a debate, will fade, because after all, they were "evil-doers".
5. Rather than the tedium of semi-electing the children of prior leaders, we can create eleborate black costumes of life support. Shrub for president in 2341!
6. No new taxes!... er... this week.
7. Yes, the space-trains will run on time, hauling the free-thinkers and dissenters to places like Hoth for "education".
8. That goofy geneva convention thing about physical torture can be eliminated, so folks like leia can be properly brutalized.
9. The universe can be dominated by white, human, anglo-saxon, men. (Ever see a pro-empire wookie in the death star?)
10. Droids have no rights.
Basically, the guy who wrote the article is pro-authority. 40 years ago, this is what was called "facist scum" in the US, because they thought law and order was more important than bowing down and serving a happy people.... forgetting, of course, they they have no right to say what people want, or that their role was to serve the people, not be served by the people.
DMA contacts (such as webmaster@the-dma.org)
Your local congresspeople/parliment officers/etc (such as John_McCain@McCain.senate.gov )
Those fine doubleclick people(such as publicrelations@doubleclick.net)
Don't be greedy, share the love with those who want to help companies share their fine product information with us!
1. The same kind of options that 2D browsers have for UI preferences.... some people like sorting alpha, some by creation date/time, some by access date/time, some by size, etc. Rather than force one or another on users, why not let users pick a default, with a custom setting per window? /Users/username, and /Application...)
2. Storing state. Okay, I quit the app. I launch it 5 minutes later... where is my carefully crafted view? I have to rebuild it?
3. Okay, I'm looking at a platter with a set of files... but nowhere on the platter is the name. How do I know if this is tuesday's set of these files or wednesdays set? To see the name of the platter (so to speak), I have to go a level "down" or to my "platter" menu.
4. Maybe a bug, maybe a feature... go to root platter. Click on hard drive. Move out a bit, go back to root. Click on same drive. Move out a bit again. The "Platter" menu now has multiple drive listings.
5. Font selection?
6. Pulling platters to their own locations.... this may be a hard one to explain, but it would be nice if I could move a platter I used frequently "away" from the others... say, halfway across the pool.
7. (Adding on 6), even better, some custom "starting points" beyond the root. Say, two or three starting platters match my working style (/,
8. Ability to adjust Icon "density" for folders with lots of files. Imagine the dock concept, where, the more objects to represent, the smaller each object becomes.
9. Dimming/fading/darkening non-active platters... maybe increasing fog intensity for "distant" objects would do the trick?
10. Some other navigation, non-platter based. How about some keys for 'fly-around" control, so I can get an overhead look, or move around without using platters as the reference point?
Overall, a great piece of work, though. Much nicer for my working style than my usual 25 stacked windows.
If you're kind of person who really wants to know how much is in an "update" (Measly bullet points doesn't do it justice): /Library/Receipts/MacOSXUpdate10.1.3.pkg/Contents/ Resources/MacOSXUpdate10.1.3.bom"
"lsbom
Will produce a file listing... piped to WC, there's 1832 files/folders involved in this "litte" fix.
Classic bonehead things I did to make openldap installations and directories less-than optimal:
:-)
1. I didn't stay up on patches, so I was running versions with known performance bugs. I should have kept up.
2. I did lousy, MySQL-101-db-like, structures (flat as all get out), which made the tree and replication features somewhere between pointless and painful. I should have fanned the tree out to different servers, increasing the ways I could spread the load. I also went too far the other way, and made high-quality, relational, db structures, which made my entry count absloutely absurd. I should have tuned my design under load.
3. I tried storing large objects in a record (say, 10-100K gifs) in each record and then wondered why it too so long to move 500 of those records.... over a 56K line. I should have used pointers or references to external blobs.
4. In speed critical applications, I did non-filtering lookups (the SQL equivalent of "select * from..."), so I was slogging though millions of bits I didn't need. I should have specified the data I needed, when I needed it.
5. I indexed all sorts of things that didn't need indexing, basically wasting tons of CPU for lookups that wouldn't be needed. I should have only indexed fields that were being frequently used, for the way they were being used.
6. I tried to stored transactional information, stuff that changed every few seconds or minutes, in an LDAP implementation that wasn't designed for "row or column" transactions. I should have used the right tool for that job, such as an atomic-transaction database.
7. When I compiled, I didn't use a fast backend, I accepted a slower gdb default. I should have used a tuned, compiled, backend.
8. I ran off of slow disks. I should never have tried to use IDE for a high-speed database.
9. I didn't tune LDAP to cache as much in RAM as possible. I should have tuned the slapd.conf entries.
10. I didn't deploy ldap servers at critical points, and instead, tried to use a few big ones. A tree of smaller servers would have worked faster.
11. I wrote scripts that did _really_ dumb things, like try to read and then edit every record in a live OpenLDAP *dbm database on the fly. I should have broken it into smaller work units,and treat the protocol on its merits.
12. Rather than pull a fast, clean, backup daily (which requires a restart), I was just running it until it died. I should have pulled regular backups, and the restart helps to close any niggling descriptors left open.
There's lots of ways to screw up any program.... I've killed apache with oversized log files and open fds, I've brought postgresql to it's knees by neglecting maintenance, I've brought MySQL to a dead stop because of file sizes (ugh.) and table design.... the list goes on and on.
That doesn't mean that apache, or http, sucks, or that SQL is useless, it means that tools have to be used properly, and learning the limitations of your tools is a painful step in learning how to use them.
HTH,
-Bop
Well, depending on how you provide the "source", be it sheet music, MIDI files, sample libraries, or your studio tracks, individuals can:
make re-orchestrations (Kronos Quartet re-recording Hendrix's "Purple Haze" with classical strings or Joan Baez recording the folk song "house of the rising sun" as a rock ballad)
record with a different voicing (Sinead O' Connor recording Prince's "Nothing Compares 2 U")
make new remixes to take the song in completely different directions (Autechure remixing Skinny Puppy's Killing Game)
or even start a symphony (the ultimate cover band...)
Musicians have been using eachother's "source" for much longer than coders have been using and improving eachother's code.
I find music and code to be virtually identical in nature: It is the structured ordering, presentation and manipulation of simple primitives, to create a more complex body of work, so it makes sense to me to have similar licenses.
Is this kind of license needed for music? Definitely. Just like the coding profession, some pieces, indeed, some jobs, are entirely dependant on using free tools, be they code libraries or sample libraries. As noted, every major operating system today now ships with free (GPL, Q, BSD, or similar) tools and components, which can be copied and distributed without any royalties. Some software products have succeeded in obtaining large market $hare, some have not.
Those products that have lined the pocket of IT professionals (including mine) were built both with privately owned IP, and public (GPL, Q, BSD etc.) IP. Those that were built strictly with only private or only public IP did not become successful or profitable based on the IP they used... that's a red herring in this discussion. The nature of the license does not dictate the profitability or success of a software, or music, product.
So what is the impact, if I choose this license, to me? Well, just like the GPL work I've done, it means that others can do what they want.
Does it prevent me from releasing other work? Of course not. I can OPL/GPL some work, BSD other work, and use a restrictive EULA for other work. Since I'm not a one-hit-wonder kind of guy (I'm a professional, I "publish" a lot), this isn't a problem.
Is my ability to capture revenue altered by my choice of license, be it in code libraries, or in songs? Of course it is. But it's my choice.