This is why you can choose between a permissive license (such as the Apache license) and a restrictive one (like the copyleft GPL). I personally don't know why you should care that someone took your code and sold it to someone else. You contributed to the open source software in the world, making it better for everyone, including that company who was able to then make some money out of it. Why do you begrudge them that? They used that to pay their marketing people, developers, support, etc. And so the world goes around.

If on the other hand, you wanted to keep that right to yourself and prevent anyone except you from making money selling your code, you could contribute under the GPL. That is a very useful license for companies who want to exclude other companies using their code to make money from (a good example is mysql). There are other ways to commercialise this software, but not the obvious path of improving it and selling it.

I speak from experience. I have one GPL project which I hope to commercialise myself one day (and so I try to exclude others from the game by making it GPL). And I am a member of the Apache Foundation where my contributions might be used by companies to make money. Good luck to them...

Since banks regularly charge even small customers less than this, I don't see the attraction. I have customers who turnover as little as $2m per year and with merchant rates down to 0.8% for Visa/MC. I don't understand why any merchant would be paying such high rates. Amex of course is much higher, but in Australia that's easily solved: very few retail merchants accept Amex.

The biggest improvement that happened in Australia was when retailers were allowed by law to charge a surcharge to the customer for processing credit cards. That has been very successful in keeping the merchant rate low since banks are not only competing with each other on credit card fees, but also with other forms of payment such as debit cards, EFT or cash.

Actually the latency from Australia to Amazon's Singapore datacentre is quite a bit higher than to the US West coast datacentre. If you look at a map of undersea cable runs, you'll see that Australia has actually quite a bit more low latency and high bandwidth connections to the USA via Hawaii and Guam than through south-east Asia.

With significant new cable coming online later this year, connectivity to the USA is only going to get better.

No. Point to point microwave technology is not helpful to get connectivity out to lots of people. The dishes need to be perfectly aligned to each other since the signal is deliberately kept within a very narrow beam. Microwave doesn't bounce off things like the lower frequencies used for wifi.

I will not comment on your superior art knowledge and ability to price artwork over the internet. However I rather like his work. This one is particularly interesting: http://www.milanigallery.com.au/artwork/little-johnny (although the context is probably not obvious to anyone outside Australia).

However, you are wrong about the way legal fees are calculated by Australian courts. They have a schedule of fees which can be claimed by the victor in a case. This schedule is fixed and has little to do with the actual costs incurred. Typically 30-50% of actual costs will be recoverable, and in this case I can imagine that he may have engaged specialist international contract and intellectual property lawyers with knowledge of US law.

A large number of companies I know about have policies of not allowing GPLv3 within their organisations, mainly because of the extremely aggressive patent invalidation clauses which are part of GPLv3. Whether you believe it is correct or not, the FreeBSD developers have rejected GPLv3 because many important downstream users of FreeBSD would have difficulty with using it had GPLv3 been allowed in, even in the compiler toolchain.

It actually is complicated. You may wish to believe that many corporations are just wrong in their concern about the GPLv3. A developer far removed from a legal department need only write a small compiler plugin to solve a problem they are having, and suddenly the company is mired in legal problems and have potentially given away a large swathe of their patent portfolio.

So, I've been modded down to troll because I disagree with the prevailing Slashdot opinion that the GPLv3 solves all the world's software licensing problems. Even the FSF explicitly states that "Developing nonfree software is not good for society, and we have no obligation to make it easier." Since GCC is their flagship product, and they have made it slightly harder for people to use it for commercial software, I still believe it is wise to be rather cautious of the post 4.2 license changes.

No, it is not as simple as this. The GCC GPL exceptions are quite complicated. Just outputting some intermediate builds from your computer (flicking a switch in the compile arguments) suddenly makes your entire proprietary product GPLv3. That is a very scary thought for any corporate legal department. http://www.gnu.org/licenses/gcc-exception-faq.html

"Oh, you mean we accidentally released a GPL version of Microsoft Exchange server because of a toolchain change in our developer workflow. Oops...."

The FreeBSD project has already declined to upgrade its GCC implementation to the GPL version 3 release. That is one of the driving forces behind the move to clang/llvm within that community. So this really isn't quite so simple... The GPL v3 is quite scary to many corporations (as it was intended to be) and so they refuse to have anything GPLv3 installed on their machines at any level. I manage a 5 person development team and would always avoid anything GPLv3 without serious consideration of the implications.

The only reason they do instant strep tests at MD offices is that they want to know the results before you walk out the door. The same goes for rapid STD tests: If you walk out the door and have to come back to get the results later, there's a chance you won't return.

A DNA test is the ultimate in non-urgent tests, and is going to remain something you head off to the phlebotomist to get done... right up until the day where they are done routinely at birth and you leave the hospital jugging a baby, birth certificate, and a flash drive containing its DNA sequence.

FreeBSD 9.0-RELEASE machine images for Amazon EC2 are available for m1.large and larger instance types: http://www.daemonology.net/freebsd-on-ec2/

Tarsnap's snapshotting model is a bit more sophisticated than how duplicity works, and its separate keys for writing/reading/deleting archives makes it possible to do some things you can't do with rsync.net (e.g., you can have a server which does daily backups with Tarsnap while still making it impossible for someone who roots the server to tamper with said backups).

But yes, tarsnap and duplicity+rsync.net are certainly more similar than, say, tarsnap and dropbox.

I love my customers...

If all you ever want to compress is files which contain the same byte repeated many times, then yes, that's how you would compress them.

But most compression formats are tuned to produce good results for the sorts of files which people are more likely to need to compress; so even when run-length encoding is used, it will typically have a low limit -- say, 255 -- on the run length.

If you don't want to live in the bay area, then don't work for RethinkDB, I guess? I don't think they're hiring anyone remote (but I might be wrong).

And what's wrong with rsync for backups?

1. rsync isn't backups, it's synchronization. It will happily synchronize corrupted files and (if you use the relevant option) file deletions.
2. rsync requires you to trust the location you're syncing your data to.
3. Because rsync is designed for synchronization, it doesn't store your data compressed.
4. Because rsync works file-by-file, it can't take advantage of duplication between files.

Probably other things too, but those are the first ones which come to mind.

All WIMPs use icons. If they didn't have icons, they would be WMPs instead.