TCP/ IP, USB stacks and ThreadX

The point is that far more people complain about the restrictions in GPL3 than do in the software source we supply.

The point is that "open" source may have far more licensing restrictions than commercial source. This is something few OS devotees will admit.

Incidentally it is wrong to suggest that Opens Source is used on "virtually all" computers world wide.

It amuses me that OS people make these wild claims and then complain and nit-pick when taken to task.

The fact is GPL license is controversial and not accepted by many in it's own community. Other commercial licenses for source are often far more simple and less restrictive. Numbers of users is irrelevant

... snip ...

Now this is a sensible attitude, rather than the violent anti-open-source one too often displayed here. This makes the major problem a suitable library license, that will allow library (etc.) use without opensourcing the caller.

... snip ...

No, the difference is that people can see what is available under GPL, and express their wishes for further access to it. How can you object to the author freely releasing his code and its source, under the license of his choice, while the normal commercial release a) requires money and b) forbids further promulgation and c) often forbids alteration. Or more.

That code is copyright by the author, and s/he has the complete right to determine the conditions of release. Many GPL authors also release material under public domain. For some examples, just explore my pages (see sig for URL). The code for nmalloc, ggets, stripovr, strlcpy is all in public domain, for example. Likewise everything in the cpm/ directory, whether or not so marked. Also note the paucity of re-releases, which are primarily triggered by bug reports.

That may well be true (since I don't know what software you supply, and how many complaints you get). But also remember that the GPL license itself, and projects using it, use an open development model - they encourage feedback (positive or negative). Thus there were huge numbers of discussion threads, blogs, articles, etc., around the drafting process for GPL3. People complain publicly about the GPL, because it can make a difference - complaints have helped shape the GPL, and influence the choice of licence in open source projects. Contrast this with, say, the license of Windows Vista. Some people complain publicly, but most people who read it (thus discounting the masses who ignore all licenses) just swear quietly to themselves - there is absolutely nothing they can do about the license, so they must accept it or not use the software.

Open source *may* have more, less or different restrictions from commercial source code. There are a fair number of popular open source licenses, and endless varieties of closed source licenses, and thus every case should be looked at individually. It is certainly true that many OS devotees gloss over this issue - but then, it is particularly relevant for us in the world of embedded development (key points being static linking, and "closed box" systems), while most OS devotees, like most programmers, live in the world of PC software development.

Obviously a hefty proportion of servers and other internet infrastructure computers either run an open source OS directly, or use major OS software such as web servers, php/perl engines, etc. And obviously the huge majority of web connections and emails pass through open source servers, routers, email transports, etc.

The majority of PC clients use some variety of windows (the Linux, BSD, MacOS, etc., machines are all heavily open source dependant). I'd expect that the majority of these have some third-party open source software (Firefox, Thunderbird, Bittorrent clients, graphics programs, pdf writers, anti-spyware programs, zip utilities, open office, etc.). Then there are the closed source programs that use open source components or libraries (programs written in languages like python, software that handles graphics formats such as png, pretty much everything that can open a zip file, etc.). For the most part, you won't know that these are using open source libraries without doing a lot of investigation. Then there are the closed source programs that use parts of BSD-licensed open source code (which requires a copyright notice, but you don't have to release the source code) - Windows itself famously used the BSD TCP/IP stack in earlier versions, and I don't think anyone would believe claims that there are no BSD-licensed open source code in the Windows code. Finally, there are the closed source programs that illegally make use of open source code without following the licensing requirements - there is no way to judge the level at which this is done, but you'd be naïve in the extreme to think it is uncommon.

Of course, not all the above cases are GPL'ed software, rather than other open source licenses.

The GPL license is not suitable for all software - people know this, and they choose appropriate licenses for their software. It's not "controversial" - developers have a choice, and they use that choice. Sometimes different developers on open source projects disagree about the best license for their project - that would happen in the closed source world too if developers (rather than their PHB) had a say in the matter.

While it is certainly true that some commercial source licenses are simpler and less restrictive than the GPL, it is by no means a black and white issue. Which license is most restrictive - one that says you must release your own source code, or one that says you can keep your code but you can only compile it on the one specified computer? Which is simpler - understanding the GPL, or figuring out if version 1.1 of your product counts as a separate "product" and requires a separate "single product license" from your RTOS vendor? The questions are meaningless - as with everything in embedded development, the answer is "it depends", and each case must be considered individually.

For what its worth, since settling on the modified GPL license for FreeRTOS.org I have had very little time taken up with license discussions on the subject (prior to that there was quite a lot).

Our commercially licensed OpenRTOS product on the other hand is a different matter. The license is often scrutinised by the customers legal department, then ping ponged back and forth. If we took this time into account in our pricing we should double the price!

With respect to GPL 3, as you point out in your article this month, only 10% of projects have moved to GPL 3.

BTW Chris, when was the last time your column did not mention open source software in some way? One could draw the conclusion you were obsessed ;o)

I've given up reading his column because of the bias displayed.

Raising 'pints' seems a lot more fun than talking about licensing, especially on a Friday.

Who was that?

In message , FreeRTOS.org writes

You have my number. It would embarrass them If I mentioned it here.

Besides they have enough problems as it was noticed by two other companies that were exhibiting who claim support for Do178B and are actually working on Do178C (whereas the one mentioning Do178C does not ) It was also noticed by quite a few people who are in Aerospace and involved in writing Do178C.

So I will leave them to shoot themselves in the foot. When challenged by some one who is working on Do178C they said it was a miss print. Difficult as the two lines above said 61508 and Do178B. Apparently one person on the stand then claimed it was a "car registration number" and they were trying to suggest they were in the automotive market. Hmmmm doesn't look that way in the photo. No one else was convinced either.