Hi Matti, Matti Karnaattu wrote on Wed, Sep 24, 2014 at 09:44:52PM +0300:
> And that reason, why not select single component to be as a reference? It's not really needed. If you want good examples, just look for new programs or library functions recently written from scratch by experienced developers, for example signify(1), identd(8), cu(1), mfii(4), tftp-proxy(8), ldomctl(8). There are no doubt many more. The release announcements tell you what is new. > And the reference component: > > "This is finest peace of code at the moment", That's in part a matter of taste, and it doesn't matter that you get *the* finest (whatever that is), you just need very good ones. > "Make rest of the code looks like this", > "Everything that is diffent to reference is wrong way to do", Not quite. Even in very good code, there is stil considerable room for personal style, and i don't doubt you will see that when looking at the above examples (though i admit i didn't check). > Some of the improvements may require enormous amount of effort to > apply whole tree, and requires also that when applying changes, have > to continuously learn/keep in own knowledge different source components. > > I don't expect that everyone have interest to all of the components, > so it doesn't motivate to hack all components. That must be taken into > account Yes. If something is found that is really important, others will help. And if anyone gets bored, they move on to something else - though, as a rule, if something is really important, someone will hopeful show the required persistence. It need not necessarily be the developer who originally found the issue. Yours, Ingo
