I help GPL violators clean up their act, it's my main business.

Every one has had a total lack of due diligence. I will come in and find that they have violated the licenses of 21 proprietary software companies (this is a real customer example) by integrating their code into their main product, just like the GPL code. Some of them only had an "evaluation" license, some not even that, some wildly violated the terms of any license they got.

Most of them are in silicon valley. They seem to have the attitude that they will clean up their legal problems when they're rich, and nothing but getting their product out of the door matters until then.

They don't ask me to feel sorry for them. I bill them a lot, and in the end, they're clean and legal.

Gift-style licensing like BSD licensing is for when you want everyone to use your code so badly that you don't care what they do with it. If you have an economic reason for that, fine. But it can create harm if you don't have your economics straight. Heartbleed was an economic failure of gift-style licensing. Very wealthy companies used OpenSSL and didn't contribute to its maintenance. There was some astronomical amount of economic damage in result. I think we all would have been better off had OpenSSL been dual-licensed and paid for by some folks, even if it had fewer users that way. And maybe that way its original developers would not have had to go to work for RSA, who prohibited them from ever touching their old code again. That's why we still have Eric Young's old, old license with the attribution clause nobody else uses any longer. He can't touch it.

GPL IMO does work best with dual licensing, because people who just hate the GPL can get what they want, and pay for making more Free Software. But if you don't care about money and don't want to use dual licensing, the growth effect you get from GPL is a lot better than making yourself some very rich company's unpaid employee by giving them all possible rights except for a very limited attribution.

Some people should pay. Some should get stuff for free. They aren't in general the same people, and they self-classify.

You must have missed the joke: "Some of our digital communications use only one digit!"

Yes, I know about the NAVAIDs, but they identify at 5 WPM and the airman's charts print the dots and dashes next to the waypoint. And there might still be runway aids that say a few letters, also at 5 WPM, but it's always the same letters for left and right and the outer, middle, and inner marker. Pilots learn the sounds for each.

When I was a Technician licensee, all of the repeaters were populated mostly by Technician licensees, and identified much faster than any of them could copy. So it was clear the Morse tone (erroneously called a "CW" ID because it wasn't Constant Wave) was there for a legal requirement only. But most of the repeaters could identify in phone, too. Back in NY, we had WR2ACD identify with the voice of the famous CBS news anchor Walter Cronkite, who was of course KB2GSD.

Actually it has QWERTY :-)

I remember in about 1979 having a conversation with someone who was convinced of the superiority of EBCDIC, and really disappointed that ASCII had become the standard. I think he died a long time ago. I wonder what he would have thought of Unicode?

I drive an automobile. It has an internal-combustion engine.

My computer keyboard has QUERTY.

I think there's a century behind both of those things, isn't there?

I am told that Bill Cross, the FCC staffer then in charge of Amateur Radio, had this displayed on his office wall. It was, of course, very different from the usual FCC comment. I was trying to make a point of how antique Morse was.

The Novice license stopped being the path to entry once the no-code Technician licensing started. There was indeed an ITU requirement, but it was at the behest of IARU, not as the requirement of any government. Similarly, FCC actually raised code speed requirements at the behest of ARRL. Shore stations had moved to phone and teletype decades before. Most ships no longer employed radio operators, but left that duty to other staff who only used phone. There was only a token continuing monitoring of Morse ship transmissions, now entirely gone.

There was one pro-code guy who pleaded with me to allow Amateur Radio to "die with dignity". If nothing else did, that convinced me that the pro-code folks could see the end coming and would accept it as long as it came after they died. Amateur licensing was declining fast, operators were dying faster than new ones got licenses, and we could see the end of Amateur Radio would come in a few decades at most..

Now there are more hams than ever, and Amateur Radio is healthy. When I say "We won", it means "Amateur Radio won". It's too bad we had to fight our own old guys.

There isn't really any reason for government agencies and NGOs to use Amateur Radio. They have satellite phones, etc. But if it really bothers you, why not lobby against allowing compensation for operators? I'd join that bandwagon.

The Pixar code base came from Lucasfilm, and went back to the 1970's. Some of that code is still in use.

I lobbied to end the requirement for an examination of the ability to decode Morse code with your ear and brain. Until 2007, the U.S. Federal Government required it before they would license all but the lowest grade of Amateur Radio hobbyists.

As part of my lobbying effort, I successfully passed a test for receiving code at 20 words per minute, and then subsequently refused to use the code on the air. 20 WPM is so fast that you have to decode by the sound of each character, you don't have enough time to pick out the individual dots and dashes.

We won.

The sound triangulated was in cryogenic liquid oxygen at 50 PSI. The speed of sound in that is approximately 1 kilometer per second.This paper is about calculating the exact speed. Elon talked in the conference about reading telemetry with millisecond accuracy. But this would yield only 1 meter resolution.

That's not the way it works. When cross-compiling, configure runs on the host. It tests that things compile. It doesn't test that they run.

Roger,

This is great. It does look like a 1:1 mapping to what we expect autoconf to do, except neater and maintainable.

The only problem with selling this to GNU folks is that it would make CMake a prerequisite to everything. But I think it's worth it. And then there's inertia. And the language isn't as pretty as we'd like.

Can you see any other possible objections?

Thanks

Bruce

Cmake are scons require that they already be installed.

This is assuming that a shell capable of running ./configure is available on all of your platforms.