Agreed. The GPL seems to be a solution, only if you accept that the actual problem is one worth solving.

If you're using a commercial version, then you have a cash hammer. Demand that they update their sources and release you an upgrade, or send your dollars to a consulting company to create the extra one or two features you need as an apache module.

The GP isn't talking about taking a piece of GPL code and removing the GPL, he's talking about licensing the code BSD, or other free use license. Comparing GPL favourably to vanilla copyright is not only an easy comparison, it's way off topic.

How is this payment to the original author? Notwithstanding if I provide a service instead of a product, then I never have to release anything... but even if I release a product in a niche market, say controllers in medical hardware, using a hypothetical modified linux kernel. I need to make the source for the modified kernel available to my customers, which they are free to redistribute, but the greater linux kernel community has no such right to the source unless they purchase one of the machines. Unless I voluntarily give back the code, or someone wants to pay me for my time by buying one of my machines, the original author never sees any of my changes.

The number of times I see comments saying that the GPL means people have to give back. This is not, and has never been the case.

So, how much can you trust open source software? Can you be sure it isn't infringing someone's copyright? The answer is usually no, unless you wrote it yourself.

As a developer, I will never release software under the GPL, but I will release software under actually free licenses like BSD or Apache. Of course, this means that I need to avoid using GPL libraries, because they do not give me the freedom to do what I require with the result of my own effort. Now in your world, this seems to be getting something for nothing, but in my world, it means I am being unreasonably restricted in a way that that would be unimaginable under any other license. Of course, if I find and fix a bug in the non-GPL library that I use, then I will naturally contribute the patch back to the project, and be happy that I could be of help. The GPL library gets no such support from me because I won't use it (if I have any choice).

Just to sum up, I find it completely unreasonable for a GPL library author to request me to release my entire application just because I make use of a supposedly free library. Even the "evil" closed-source commercial proprietary world would rarely put such a restriction on use of a library, and they wouldn't be as brazen as to say they're doing you a favour at the same time.

Now, as someone who moans about the GPL, do I still fit into your world as someone who wants something for nothing?

You missed *rock* bottom. I've had an Australian contract for the past three years while not living in the country, to keep my number when I go back. I'm on a post-paid contract with a minimum spend of $0... although I originally had some issues with charging me for roaming voicemail until I got roaming turned off.

Haven't had to pay anything for most of three years, although that's of course for zero minutes. Only costs 0.5c per second (AUD) when I make calls though, so 10 minutes per month would be AUD3.

You don't really get plans like that any more, that's the other reason I've been holding on to my plan even while out of the country!

Good idea regarding population density and distance that coverage requires, but it doesn't explain how Australia fared so much better.

For that matter, in both the high usage and low usage charts, Australia even beat out Japan! For some reason not the medium usage, however.

You're comparing small apples to big apples. You can't claim your plan is cheaper (better value) just because it costs less. What would your plan cost if it included roughly 2500 minutes of talk time? What about if you add data? Unlimited data? Personally, I'll be moving back to Australia soon, and the cheapest I can find for data is AUD20 for 1Gb of data (roughly EUR10) per month.

I still remember having a roll of electrical tape confiscated going on to a plane in Australia... apparently tape was only acceptable if it is clear tape. Because it's harder to tie someone up with clear tape than opaque?

Isn't that about how sure we were that the sub-prime bubble wasn't going to burst? At least, depending on which particular model we are discussing. Not that I disagree with you, but questioning the mainstream is not the same as disagreeing with the mainstream. If the truth doesn't stand up to discussion, then I'm not sure whose truth it is...

With all due respect, as someone who works in the music business you're not exactly an impartial commenter. Fair enough, as someone who doesn't know the figures, or even sufficient knowledge of statistics to understand them even if I had them, I'm not either.

It's almost certainly a fact that some people are pirating music instead of buying it. However, it's also almost certainly a fact that if the piracy avenue was not available, then some (most?) of those people would simply do without music. The music industry is not "competing" with free, because it's not a competition. The offerings are not "like for like".

As other people have pointed out, it's more than price, as well. A high quality MP3 is more valuable than a high quality AAC with DRM, because you can use it however you want without having to repurchase the same track. Most companies now seem to be going DRM-free, but for myself it's still not a good option because I don't trust digital-only storage... recently losing several years of digital photographs (and the on-site backup) in a home robbery has made me even more wary.

For me, the CD is still the best value option available... which means that if there aren't enough good tracks on the CD to justify the cost, I just do without. Funnily enough, I haven't bought many mainstream CDs in recent years... purchases have been limited to bands I particularly like (which coincidentally happen to be small-label, but I'd buy them even if they were mainstream non-Sony/BMI), musical scores and compilation CDs.

Finally, you mention in another post that "For every successful artist you know of, there are 10-15 failures." I don't believe this sufficiently quantifies the vast explosion of artists producing work at the moment, documented elsewhere in this story. I would put money on the fact that independent and small labels do not have anything approaching this failure rate, since there are such low cost distribution and advertising avenues that even a "failure" manages to break even, and simply doesn't become a success. There is no compelling need for one success to offset so many failures if you aren't losing so much money on the failures, and the double whammy supporting a huge ecosystem of non-artists inside the label itself... unless they are the failures you refer to ;-)

Giant meteor flying at the world, much? It's natural to peak near the end, that's what they call a finale.

That's ok, it's your own patent

One more note between the X25-E and X25-M... The X25-E is branded as an enterprise-level disk, and is correspondingly much more expensive. It is a SLC design (single layer controller), where the X25-M is a MLC design (multi-layer controller). SLC chips are expected to last considerably longer than MLC, because fewer bits are stored in a single cell in an SLC chip, each cell is expected to be written fewer times when writing or modifying data on the disk. But still, as mentioned in numerous comments here, the X25-M should* last long enough for most consumer uses. *- time may prove us wrong, but I hope not.

Sequential reads are certainly sequential reads, but that might not mean exactly what you think it means. In a spinning disk HDD, a sequential read means move the head to the correct location, and then stream the data as the disk spins under the head. This is roughly as fast as it gets with spinning media, ignoring cache etc. In a SSD, you can actually benefit from parallelisation. SSDs can (if supported by the controller) read simultaneously from different chips, which means either reading the file in a number of parallel sequential reads simultaneously, or being able to satisfy multiple read requests simultaneously, as long as they are stored in different memory chips. So defragmentation in the classical sense may not actually be a desirable goal in the new world order.

Why does everyone insist on piping grep to awk when awk can also perform the regex? Actually, there are more reasons than you might expect...

• Familiarity. More people are likely to recognise grep regular expressions as search terms than awk regular expressions, even though they are often identical. Grep just means search, at least to a geek.
• Substitutability. You can replace grep with any other tool that can process stdin, without worrying about which argument is actually the file providing the data. This can be very useful if you're logically building up the command.
• Efficiency. Cat does nothing but open a file and read it. It probably does that extremely well. Grep is primarily a search algorithm, and its file handling routines might not be optimal; it might hold file descriptors open longer than necessary, or whatever. In grep's case, it probably doesn't, but without looking at (and understanding) the source of what you're actually using you don't actually know. Of course, if you use the right tool for each part of the job, then you don't even have to care.
• Permissioning. Linux/unix/posix doesn't actually do this, but you could run grep with a very restricted permission set; it is allowed to read from stdin, write to stdout, and allocate a certain amount of memory, and that's it. Even if your grep was somehow compromised, it couldn't do anything but print bad stuff to stdout. The cat process has a higher level of trust, and is allowed to read from the file system, and write to stdout. If you happened to be doing a multi-file grep, then you could invoke grep with a higher level of permissions so it can read the filesystem itself. If you're running an object permissioned OS (eros, keykos, coyotos, or a number of other OSes that aren't actually ready for use yet) then cat | grep is actually the recommended way to do things.

But mostly, it's just what they're used to. There's nothing wrong with it, it's roughly the same efficiency either way, so do it the way that makes sense to you, and for the problem you're solving.

And don't forget words that cannot be transcribed without context, such as their/there/they're and which/witch... "I'm going over there, which is useful because I need to talk to whoever anyway" vs "I wouldn't give their witch any credibility".