On 08/05/2017 10:20 AM, Klemens Morgenstern via Boost wrote:

...

So for some reason I do not trust Microsoft. And I was wondering if the Boost license protects us from an idea behind a library we wrote. Otherwise this makes the Boost license not very useful. Protects us from? You mean protects the idea from being copied right? I don't think a license can do that whatsoever, since it is based on copyright not on a patent. A patent might be able to protect an idea, copyright can only protect the actually written words, i.e. the code itself not the concepts.

You mean it can only protect us from our code being copy & pasted around.

Also: have you reached out to Herb Sutter? I actually don't get your concern, do you want to be credited? His library is published under the MIT license, that's still open-source.

These days when I present root_ptr then I am being told: "Oh your library is like deferred_heap" when it should be the other way around. I also learned the hard way not to contact Microsoft in any way otherwise even my complaint will be plagiarized (sarcasm here). Everybody here knows root_ptr was once called block_ptr and shifted_ptr so my previous attempts came to life long before deferred_ptr. I just don't want to have Microsoft in the way anymore. Regards, -Phil

On 08/05/2017 11:53 AM, Jens Weller via Boost wrote:

Hello Phil,

...

Gesendet: Samstag, 05. August 2017 um 16:04 Uhr Von: "Phil Bouchard via Boost" An: boost@lists.boost.org Cc: "Phil Bouchard" Betreff: [boost] Microsoft vs The Boost License

Hi,

Just to put everybody in context:

1) Microsoft had invested into Corel so that they get rid of Corel Linux back in 2000 when I was working there: https://www.forbes.com/2000/10/03/1003corel.html

2) Microsoft once again copied my Fornux PowerCalc after I had presented it to them using some web interface: http://www.fornux.com/

With their Microsoft PowerToys: https://en.wikipedia.org/wiki/Microsoft_PowerToys#Included_applications

So this is personal?

No I just want to point out their strategy to the community so that it doesn't happen to anyone else.

...

3) And now here I am with Microsoft trying to copy the idea behind root_ptr's node_proxy: https://github.com/philippeb8/root_ptr

With their "deferred_ptr heap": https://github.com/hsutter/gcpp

So for some reason I do not trust Microsoft. And I was wondering if the Boost license protects us from an idea behind a library we wrote. Otherwise this makes the Boost license not very useful.

You should have some more evidence, that there is a relation between Herbs gcpp, and your root_ptr.

Btw. Herb presented this last year in his CppCon keynote: https://www.youtube.com/watch?v=JfmTagWcqoE

So thats why its fairly well known in the community.

Well we can look at the logs of block_ptr / root_ptr which date back to February 2016 on Github and Mr. Sutter's first commit ironically was done later in September of the same year. Regards, -Phil

On Sat, Aug 5, 2017 at 12:30 PM, Phil Bouchard via Boost wrote:

On 08/05/2017 12:11 PM, Phil Bouchard via Boost wrote:

The evidence in this case is probabilistic because what are the chances we come up both at the same time the exact same idea after 60 years of research on the subject?

High, actually. It happens all the time. There is a math thesis about it - predicting the odds of a thesis idea being simultaneously worked on independently by someone else. And just as the thesis would suggest, there is actually more than one paper on the subject. Tony

On 5 August 2017 at 19:11, Phil Bouchard via Boost wrote:

Well we can look at the logs of block_ptr / root_ptr which date back to February 2016 on Github and Mr. Sutter's first commit ironically was done later in September of the same year.

For a similar idea https://en.wikipedia.org/wiki/Hazard_pointer a patent was filed in 2002, by IBM and Michael Maged (and it seems Andrei Alexandrescu was involved as well). As you can read, the application was abandoned in 2010, which says all about how difficult/usefull this approach is. I think it's best to forget about it. I surmise that any production ready M$ code is highly unlikely to look anything close to your code (no-body writes code like that :-)), god it's ugly). and prior art can possibly be demonstrated. degski -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798

On 5 August 2017 at 20:54, Phil Bouchard via Boost wrote:

Lol and thank you all for your advices. I guess my best defense is present my work to the CppCon 2018 and keep working on it. Meanwhile I am presenting it to the ISO C++ committee.

Yes, just (more easily said than done) work towards the best implementation possible and you *will* be acknowledged. degski -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798

On 5 August 2017 at 21:21, Phil Bouchard via Boost wrote:

On the promotion side to ISO I know the procedures are strict (like a thesis you have to defend in person) and takes some time but it looks like I reached to point of no return ;)

Maybe the Microsoft *vs* root_ptr way of looking at things is wrong. Herb Sutter presented an idea (never stated it was all his) and wrote in his words a "Toy" implementation, with no promiss of going beyond this, to demonstrate that idea. It sounds to me like an invitation for some-one to implement the real deal. That some-one could be you. So why don't you rename root_prtr (again :-) ) to deferred_ptr and incorporate any (other) ideas in HS's, that you might have missed, and I guess you will have an early supporter. degski -- "*Ihre sogenannte Religion wirkt bloß wie ein Opiat reizend, betäubend, Schmerzen aus Schwäche stillend.*" - Novalis 1798

On Sat, Aug 5, 2017 at 10:39 AM degski via Boost wrote:

For a similar idea https://en.wikipedia.org/wiki/Hazard_pointer a patent was filed in 2002, by IBM and Michael Maged (and it seems Andrei Alexandrescu was involved as well). As you can read, the application was abandoned in 2010, which says all about how difficult/usefull this approach is.

Hazard pointers relate specifically to concurrency and not general memory management. Efforts to standardize them are ongoing: https://wg21.link/p0233r4

On 08/05/2017 12:19 PM, Paul A. Bristow via Boost wrote:

Boost license does not protect *you* much.

What it does do is protect *everyone* from any attempt to by Microsoft or others to patent or claim copyright on the code or idea, and thus preventing or restricting its use.

This means unless the MIT license is compatible with the Boost license then they are infringing the copyright agreement.

Boost provides excellent evidence of prior art, not just from being in a release, but also being the subject of public discussions on Boost.

(If Boost had published the SHE exception handling, for example, the World would have been a Better Place ;-) And it would have been even better place if software patents had been very much more restricted - I blame the US and UK and EU patent offices, and too many lawyers in parliaments.)

That is no excuse for Microsoft not giving due credit to your work. You can't claim any recompense but it is very bad manners, in this case bad corporate manners.

I'd ask for your work to be recognised and referenced. I believe Herb to be an honourable man, so I'd try him first.

Contacting popular people is a very hard thing to do as he probably receives tons of emails already. Regards, -Phil