Re: Is the eCos Open Source development model at risk?

[Thomas BINDER <Thomas dot Binder at frequentis dot com>]

Michael Checky wrote:
> 
> I don't entirely agree with your assessment.  I do agree that some
> gatekeeper mechanism is needed to keep the eCos sources in a high quality
> state, but I think the current mechanism is starting to show its
> deficiencies.  I base this on what I think are the some of the most
> important reasons for an Open Source development model:
> 
> 1. The eCos community benefits when the users contribute bug fixes,
> improvements, enhanced features, etc. back to the source respository.
> 2. Having these contributions in the source repository increases the
> probability that they will be used and other users will help improve them.
> 
> When you post your contributions to the mailing list, after a few months
> they are effectively lost to the rest of the community, since it takes a
> reasonable amount of effort search the mailing list to find if any
> particular patch is just what the user wants.  The eCos community has lost
> the benefits of your contribution.  Second, if no one is using your
> contribution, you lose the benefit of other users improving your
> contribution.  And the situation will get worse, especially if the
> popularity of eCos increases and we get more user contributions.
> 


I also think that this is an important point especially for companies: An increasing number of users, also increases the probability to discover bugs. This was also the main reason which made us decide to release our extensions (memory protection). 

However, the lack of response from the maintainers (the original mail was sent ~ 2 months ago) is an indication that either, (a) nobody seems to be interested, or (b) the maintainers are simply too busy to respond.

If (a), then the question arises what type of applications eCos *is* typically used for. After all, a working memory protection is usually essential for applications where stability and safety are of major concern. I am not talking about kernel bugs here, the idea is rather to protect the kernel from the application. Commercial operating systems like vxWorks obviously come with memory protection to meet high end requirements. 

One of the reasons we chose eCos over the commercial alternative(s) was the open source development model, where benefits clearly come from the big community that helps in stabilizing the code.

On the other hand, if (b) the maintainers are just too busy, then I agree with Michael, the development model might not be optimal. Many useful contributions will get lost, in which case the benefits for the companies might be harder to find....


T.Binder