Right, I hadn't had my coffee yet. I was thinking about software for some reason.
The general point still stands though. There are very few people who pirate because they're not going to buy it for any price. They pirate because it's the easiest option for them. Either because they lack the money (but would buy for a cheaper price if they could), or because the pirate provides better service (no DRM, easy downloads, etc).
Kids, unemployed people maybe. Adults much less so. At some point you realize: "Hmm, I can either pay for this with 15 minutes of work, or spend 2 hours looking for a crack that might be loaded with spyware. Then I'll probably have to spend a day to reinstall Windows."
If you don't have money, you just "pay" with your time instead. Give me MP3 at $0.1 per song and pirating will be completely pointless. Movies I'd be willing to buy at about $5, with no DRM, as a DVD image.
Personally, I don't buy DVDs. Why? Expensive, insane industry, forced ads in content I'm paying for, DRM interfaces... like hell I will pay for that. Remove all that insanity, offer it with an easy to access system that works from Linux, and I'll happily start buying.
I've never seen a program that needs to be built differently depending on whether the host is SMP or not. If it does exist, that's just horrible design. If it supports an optimization with multiple CPUs, it should detect that at runtime instead.
Ok, so how do you propose an user sees a difference between PNG, GIF and JPG?
Thing is, the file format IS important:
PNG: Photos would be awfully large. GIF: Photos would look like crap. JPG: Text, webcomics, etc would look like crap.
Programs are capable just fine of handling the difference without the extension. Now tell me, without an extension, how does the user figure out which format it is, without using some specific tool or opening each image?
Sure, grandma doesn't think she cares whether it's JPG or PNG, but it's going to be rather puzzling that a picture of the cat from the webcam can be sent by mail just far, but the one from the scanner is for some reason huge and takes ages on dialup, or that text on a pretty postcard is ugly.
Running a server on Debian can be a pain. Say, I have a mail server. Before Sarge was finally released I was stuck with an ancient version of exim without support for exiscan-acl. Sure, I could get it from backports, but that's a pain, and implies trusting an external source. I could also compile it, which is also a pain for reasons you already mentioned.
The thing is that for a mail server, exim is THE thing I want to run, and the rest is pretty much auxiliary. It makes no sense to have the one thing I have the machine for come from some outside source.
So, on my part, Debian is getting relegated to the firewall, which does pretty much nothing but iptables. There the glacially slow release schedule is not a problem. And the servers are getting switched to Gentoo.
Nonsense. Adapt to the new times, or die. Especially in computing, where things change every year.
I work on in-house software and I assure you I don't give a damn if every piece of software that can be bought off the shelf becomes free or disappears. The task I have to accomplish won't change anyway.
Sheesh, then I repeat: Don't license your software under the GPL!
First, all I'm saying is that under the GPL it's *possible to make money*. Perhaps you don't like the way that works, but there's a very easy solution to that: Don't use the GPL for your work.
And BTW, if you want to be paid, you ALWAYS write what other people want you to write. If you write something for which there's no demand, you get zilch as a reward. It's called capitalism.
And what does that have to do with this discussion?
The article is about some people who got for some reason surprised at seeing RMS charge for a service. Well, I posted and explained that he never said he had anything against charging for a service.
And that's it, the economics of basing a business on selling GPLd software have absolutely nothing to do with what we're discussing here.
RMS never said he had anything charging for services being wrong. What he doesn't like is the producer being able to hold him hostage. That's his opinion of how things should work, if you don't like it that way, then just don't release your work under the GPL.
Besides, earning money under the model of selling it to one person is perfectly possible. It's called contract work. Not all software is made to be sold on the shelf.
If you'd ever read the GPL, you'd notice that source only needs to distributed to the people who got the binary, and the binary can be charged for. I never heard Stallman say that services like duplication, tech support, etc should be free. IIRC, Stallman has a webpage somewhere detailing his requirements if you want to have him give a talk, which sounds completely fine to me.
Stallman was always about freedom in the political sense, not in the lack of economic compensation one.
So, somebody asks Stallman to sign their badge. Stallman realizes he could be stuck there for hours signing badges instead of doing something more useful. So he asks for a donation for the FSF (not even for himself!) to get something out of it, and hopefully reduce the size of the queue. Sounds completely reasonable to me.
It's not like Stallman ever had anything against charging money, from what I heard, he sold Emacs tapes.
You're missing the "return 0;" statement, which means that the return value will be random junk. This in turn means you can't count on it to determine whether the program ran correctly, so it's not reliable.
Tape drives do have random access, they can rewind. Of course, seek time is about 60 seconds on my DDS4 drive, but it's still a whole lot better than the time it takes to read the whole tape.
If you have an index, and don't write your backups as one huge.tar.bz2 file, you can just tell the tape drive to seek to the right position.
Granted, it's not nearly as nice as a hard disk, but not as bad as you seem to imply either.
Tape. LTO3 can hold 400GB of data, uncompressed. Of course, it's expensive as heck, but it does exist.
I recently got myself a DDS4 tape drive. I thought about my backups and came to a few conclusions:
I'll never keep it up if it involves messing with a stack of DVDs. It's got to fit all my backups in one piece of media. It won't last if it interferes with my normal usage. I need to backup daily to get into the routine (too easy to forget if it's once a week or month)
So, I looked at how much I had to backup, and I had about 15GB of data that I backup daily. Lots of data is static (music) and can be archived on DVDs once. I looked around, and DDS4 had the right capacity (20GB per tape).
The tape backups include my root partition, which allows recovery of the whole system. I just insert a tape in the morning, cron takes care of starting the process, and spits it out when done.
Unix directory layout basics:/bin: Binaries needed to boot the system./lib: Libraries needed to boot the system./etc: Config files/usr: Stuff for use after booting the system/usr/bin: See above
Why is it that way?/usr can be located on NFS (network). You can't mount NFS without going through some minimal boot process where you start the network, get an IP address from DHCP, etc. You also need the mount program. All that is in/bin.
This is all meaningless for a normal user, but really handy for corporate systems. Got 200 machines? Why spend money on disk space when they can all mount it from the network and get all upgraded at once?
Why is stuff in say,/usr/bin/perl instead/programs/perl/perl? Because then the PATH would be awfully long, and each time you typed "perl" the system would need to search a LONG list of possible places where it could be. GUIs don't care about that because every icon simply links to the full path, but command line interfaces don't work that way.
Certainly not, I'm taking about the specific case here: Headlines for google.
This stuff makes perfect sense for local newspapers, where 90% of the audience can be assumed to be native, and to know what's going on in that specific area. But online is a different world, where content is equally available to everybody.
Also, things like RSS allow very quick scanning of headlines without seeing pictures or text, which almost guarantees that many people will ignore something they don't get at first glance. Reducing your audience in this way seems counterproductive.
They're, IMO, just different worlds. In a newspaper you try to get some extra audience by getting people interested, trying to get a reaction of "that's clever", or "wonder what is it about". Online you have vast amounts of people that are actively searching and filtering information, looking for something specific. You'll get them by being direct and to the point.
English is my third language. I'm capable of talking to people online without problems, and can read technical manuals and books just fine. However, I wouldn't get even half of the witty headlines.
To put a few examples from the comments here:
"Super Calligraphic Raster SPX Projectors": This would take a while to sink in, but I'd eventually get it, having seen part of the movie. Have in mind though that not all movies are translated so literally that you can tell a reference to something in English when you've only seen the Spanish dub.
"Foot Heads Arms Body": This would be completely puzzling, as I've never heard of Michael Foot. Just that stops me from getting at all what is it about, even though I have heard about the IRA. Changing it to "Michael Foot Heads Arms Body" would give me a clue, but still require thinking about what arms body they're talking about, as the IRA isn't in the news often here.
"Headless body found in topless bar": Easy.
"Sick Gloria in transit Monday": Don't get it.
"Schizophrenic man rapes woman and flees.": Interpreted literally, I assume there's more there than I can see.
"CLOSE BUT NO CIGAR": Don't get it. Heard the expression but never found the meaning.
"COPS MAKE BUTT-ER KNIFE CON SPREAD 'EM": Don't get it.
Despite any atrocious grammar mistakes that could be above, my knowledge of English is very good by local standards. Most people coming out of school here wouldn't be able to finish even the Harry Potter books, as it'd take them months to read one of the large ones, with constant use of a dictionary.
A very few people will go "Yay! It doesn't look so confusing now. I'll just ignore/system. For those few, it'll be good
Other people will initially be like above, but will then go on to download something like the Linux version of Neverwinter Nights, and be really puzzled that it doesn't work. Then they'll go scream at Bioware, and complain that Linux is crap and why can't just it be like Windows where you click on setup.exe.
Application developers everywhere will be annoyed that now they have to take into account things like that scripts beginning with #!/bin/bash will now fail, and that commonly used tools are suddenly entirely somewhere else. Clever scripts that search common paths for stuff will fail miserably.
Some admins will really hate having them on the network, as the Linux FS layout is well thought for partitioning and NFS.
It'll be a nightmare for IRC channels, now they have to consider that telling people to boot with init=/bin/bash may not work, and some will need to figure out where everything is on this particular distribution. Not only annoying, it looks bad when a supposedly expert user can't figure it out.
People with experience will probably ignore the distribution, why use something made in a way that's incompatible with almost everything else?
As a result of the above, help in IRC and similar will be hard to get
Resuming, IMHO, this means a precious few happy people, a lot of annoyance for many others, and a lot of time spent on development and patching that could be spent a lot better on something more important.
I saw a TV program about these implants some time ago. For that particular patent, vision was in a 16x16 matrix (IIRC), black/white only, and not all the pixels worked.
The effective vision obtained from that wasn't much. Got the ability to notice sources of light, and when standing on the street the ability to determine whether there's a building in front. On the show the patient commented she hoped to be able to see her husband's outline and was disappointed she didn't see it.
Of course, it's an amazing development anyway, which will hopefully be developed further into something actually usable, but at least the one I saw wasn't good for that much.
Re:Why I don't use gmail
on
Gmail vs Pine
·
· Score: 1
By subject, in the current folder, pretty much instantly. Or the whole account, without indexing, which is of course going to be fairly slow.
Of course, this doesn't work for everybody. It works just fine with my particular usage pattern, though. I get large amounts of mailing lists and automated mail from various servers. All this is sorted into folders. Mail from work for instance, would be automatically sorted into Work/<company>/Servers/Mail/Cron or Work/<company>/Departments/<department>. The total amount of mail is quite high, but the amount in each folder is quite low. Having to grep the whole account for something is a very rare ocurrence, as keeping it sorted avoids the need in the first place.
It's indexed by google. Thanks, but I like my private stuff staying private.
KMail can search too, and works faster.
I don't like things changing without my control, I'll update when I want.
I hate HTML email and graphics, and never get any mail that uses it.
I prefer having my data safely at home, on my RAID-1, backed up daily to tape, rather than to depend on an external service for which I'm one among millions.
It doesn't support gpg.
It doesn't have imap.
Weak filtering. No procmail.
It's not possible to filter out parts of messages, such as stripping the ads added by various free mail services.
It's not possible to implement cron jobs that archive/remove old junk.
Konqueror isn't fully supported. I almost never use Firefox.
My mail doesn't stop being available if my connection goes down.
Slashdot Mirror
At this point, the Linux kernel is GPLd forever.
So if you don't like it, just don't use it. If you want FreeBSD or Windows, you know where to find it.
Right, I hadn't had my coffee yet. I was thinking about software for some reason.
The general point still stands though. There are very few people who pirate because they're not going to buy it for any price. They pirate because it's the easiest option for them. Either because they lack the money (but would buy for a cheaper price if they could), or because the pirate provides better service (no DRM, easy downloads, etc).
Nonsense.
Kids, unemployed people maybe. Adults much less so. At some point you realize: "Hmm, I can either pay for this with 15 minutes of work, or spend 2 hours looking for a crack that might be loaded with spyware. Then I'll probably have to spend a day to reinstall Windows."
If you don't have money, you just "pay" with your time instead. Give me MP3 at $0.1 per song and pirating will be completely pointless. Movies I'd be willing to buy at about $5, with no DRM, as a DVD image.
Personally, I don't buy DVDs. Why? Expensive, insane industry, forced ads in content I'm paying for, DRM interfaces... like hell I will pay for that. Remove all that insanity, offer it with an easy to access system that works from Linux, and I'll happily start buying.
What are you compiling, I wonder?
I've never seen a program that needs to be built differently depending on whether the host is SMP or not. If it does exist, that's just horrible design. If it supports an optimization with multiple CPUs, it should detect that at runtime instead.
Ok, so how do you propose an user sees a difference between PNG, GIF and JPG?
Thing is, the file format IS important:
PNG: Photos would be awfully large.
GIF: Photos would look like crap.
JPG: Text, webcomics, etc would look like crap.
Programs are capable just fine of handling the difference without the extension. Now tell me, without an extension, how does the user figure out which format it is, without using some specific tool or opening each image?
Sure, grandma doesn't think she cares whether it's JPG or PNG, but it's going to be rather puzzling that a picture of the cat from the webcam can be sent by mail just far, but the one from the scanner is for some reason huge and takes ages on dialup, or that text on a pretty postcard is ugly.
Thank you, this exactly what I've been saying.
Running a server on Debian can be a pain. Say, I have a mail server. Before Sarge was finally released I was stuck with an ancient version of exim without support for exiscan-acl. Sure, I could get it from backports, but that's a pain, and implies trusting an external source. I could also compile it, which is also a pain for reasons you already mentioned.
The thing is that for a mail server, exim is THE thing I want to run, and the rest is pretty much auxiliary. It makes no sense to have the one thing I have the machine for come from some outside source.
So, on my part, Debian is getting relegated to the firewall, which does pretty much nothing but iptables. There the glacially slow release schedule is not a problem. And the servers are getting switched to Gentoo.
Nonsense. Adapt to the new times, or die. Especially in computing, where things change every year.
I work on in-house software and I assure you I don't give a damn if every piece of software that can be bought off the shelf becomes free or disappears. The task I have to accomplish won't change anyway.
Sheesh, then I repeat: Don't license your software under the GPL!
First, all I'm saying is that under the GPL it's *possible to make money*. Perhaps you don't like the way that works, but there's a very easy solution to that: Don't use the GPL for your work.
And BTW, if you want to be paid, you ALWAYS write what other people want you to write. If you write something for which there's no demand, you get zilch as a reward. It's called capitalism.
And what does that have to do with this discussion?
The article is about some people who got for some reason surprised at seeing RMS charge for a service. Well, I posted and explained that he never said he had anything against charging for a service.
And that's it, the economics of basing a business on selling GPLd software have absolutely nothing to do with what we're discussing here.
True, but what about it?
RMS never said he had anything charging for services being wrong. What he doesn't like is the producer being able to hold him hostage. That's his opinion of how things should work, if you don't like it that way, then just don't release your work under the GPL.
Besides, earning money under the model of selling it to one person is perfectly possible. It's called contract work. Not all software is made to be sold on the shelf.
If you'd ever read the GPL, you'd notice that source only needs to distributed to the people who got the binary, and the binary can be charged for. I never heard Stallman say that services like duplication, tech support, etc should be free. IIRC, Stallman has a webpage somewhere detailing his requirements if you want to have him give a talk, which sounds completely fine to me.
Stallman was always about freedom in the political sense, not in the lack of economic compensation one.
So, somebody asks Stallman to sign their badge. Stallman realizes he could be stuck there for hours signing badges instead of doing something more useful. So he asks for a donation for the FSF (not even for himself!) to get something out of it, and hopefully reduce the size of the queue. Sounds completely reasonable to me.
It's not like Stallman ever had anything against charging money, from what I heard, he sold Emacs tapes.
So, the "land of the free", huh?
Sometimes I'm really glad I don't live in America.
I wonder why people keep bringing the "nova" example when there are much better ones.
Mitsubishi Pajero: "Pajero", in Spain means literally "wanker". No need for weird interpretations.
Mazda Laputa: Will be heard as "Mazda la puta", or "Mazda the whore". "Mazda" also sounds like a female name.
It's got a bug.
You're missing the "return 0;" statement, which means that the return value will be random junk. This in turn means you can't count on it to determine whether the program ran correctly, so it's not reliable.
Tape drives do have random access, they can rewind. Of course, seek time is about 60 seconds on my DDS4 drive, but it's still a whole lot better than the time it takes to read the whole tape.
.tar.bz2 file, you can just tell the tape drive to seek to the right position.
If you have an index, and don't write your backups as one huge
Granted, it's not nearly as nice as a hard disk, but not as bad as you seem to imply either.
Tape. LTO3 can hold 400GB of data, uncompressed. Of course, it's expensive as heck, but it does exist.
I recently got myself a DDS4 tape drive. I thought about my backups and came to a few conclusions:
I'll never keep it up if it involves messing with a stack of DVDs.
It's got to fit all my backups in one piece of media.
It won't last if it interferes with my normal usage.
I need to backup daily to get into the routine (too easy to forget if it's once a week or month)
So, I looked at how much I had to backup, and I had about 15GB of data that I backup daily. Lots of data is static (music) and can be archived on DVDs once. I looked around, and DDS4 had the right capacity (20GB per tape).
The tape backups include my root partition, which allows recovery of the whole system. I just insert a tape in the morning, cron takes care of starting the process, and spits it out when done.
Unix directory layout basics: /bin: Binaries needed to boot the system. /lib: Libraries needed to boot the system. /etc: Config files /usr: Stuff for use after booting the system /usr/bin: See above
/usr can be located on NFS (network). You can't mount NFS without going through some minimal boot process where you start the network, get an IP address from DHCP, etc. You also need the mount program. All that is in /bin.
/usr/bin/perl instead /programs/perl/perl? Because then the PATH would be awfully long, and each time you typed "perl" the system would need to search a LONG list of possible places where it could be. GUIs don't care about that because every icon simply links to the full path, but command line interfaces don't work that way.
Why is it that way?
This is all meaningless for a normal user, but really handy for corporate systems. Got 200 machines? Why spend money on disk space when they can all mount it from the network and get all upgraded at once?
Why is stuff in say,
Certainly not, I'm taking about the specific case here: Headlines for google.
This stuff makes perfect sense for local newspapers, where 90% of the audience can be assumed to be native, and to know what's going on in that specific area. But online is a different world, where content is equally available to everybody.
Also, things like RSS allow very quick scanning of headlines without seeing pictures or text, which almost guarantees that many people will ignore something they don't get at first glance. Reducing your audience in this way seems counterproductive.
They're, IMO, just different worlds. In a newspaper you try to get some extra audience by getting people interested, trying to get a reaction of "that's clever", or "wonder what is it about". Online you have vast amounts of people that are actively searching and filtering information, looking for something specific. You'll get them by being direct and to the point.
I completely agree.
English is my third language. I'm capable of talking to people online without problems, and can read technical manuals and books just fine. However, I wouldn't get even half of the witty headlines.
To put a few examples from the comments here:
"Super Calligraphic Raster SPX Projectors": This would take a while to sink in, but I'd eventually get it, having seen part of the movie. Have in mind though that not all movies are translated so literally that you can tell a reference to something in English when you've only seen the Spanish dub.
"Foot Heads Arms Body": This would be completely puzzling, as I've never heard of Michael Foot. Just that stops me from getting at all what is it about, even though I have heard about the IRA. Changing it to "Michael Foot Heads Arms Body" would give me a clue, but still require thinking about what arms body they're talking about, as the IRA isn't in the news often here.
"Headless body found in topless bar": Easy.
"Sick Gloria in transit Monday": Don't get it.
"Schizophrenic man rapes woman and flees.": Interpreted literally, I assume there's more there than I can see.
"CLOSE BUT NO CIGAR": Don't get it. Heard the expression but never found the meaning.
"COPS MAKE BUTT-ER KNIFE CON SPREAD 'EM": Don't get it.
Despite any atrocious grammar mistakes that could be above, my knowledge of English is very good by local standards. Most people coming out of school here wouldn't be able to finish even the Harry Potter books, as it'd take them months to read one of the large ones, with constant use of a dictionary.
Resuming, IMHO, this means a precious few happy people, a lot of annoyance for many others, and a lot of time spent on development and patching that could be spent a lot better on something more important.
The algae would probably be MUCH better.
I saw a TV program about these implants some time ago. For that particular patent, vision was in a 16x16 matrix (IIRC), black/white only, and not all the pixels worked.
The effective vision obtained from that wasn't much. Got the ability to notice sources of light, and when standing on the street the ability to determine whether there's a building in front. On the show the patient commented she hoped to be able to see her husband's outline and was disappointed she didn't see it.
Of course, it's an amazing development anyway, which will hopefully be developed further into something actually usable, but at least the one I saw wasn't good for that much.
By subject, in the current folder, pretty much instantly. Or the whole account, without indexing, which is of course going to be fairly slow.
Of course, this doesn't work for everybody. It works just fine with my particular usage pattern, though. I get large amounts of mailing lists and automated mail from various servers. All this is sorted into folders. Mail from work for instance, would be automatically sorted into Work/<company>/Servers/Mail/Cron or Work/<company>/Departments/<department>. The total amount of mail is quite high, but the amount in each folder is quite low. Having to grep the whole account for something is a very rare ocurrence, as keeping it sorted avoids the need in the first place.