FreeLists / openbeos / [openbeos] Re: binary middle ground

["Michael Phipps" <mphipps1@xxxxxxxxxxxxxxxx>]

This is probably the one really outstanding reason *FOR* breaking 
compatability. But the fact is that we have no real test process other
than BeOS as it is today. If I could plug in, say print_server_OBOS and run 
print jobs
from (say) Productive (SHAMELESS PLUG), then I can be fairly sure that 
we have done well. 

There are several things that people would like to do that would break this
plug-and-play architecure. Improving APIs, changing compilers, etc. While none
of those things are bad, since we don't have Be's test suites, it would be hard 
to 
prove that ANYTHING is right. :-/

Binary compatability is good for the end users. It allows them to use tons of 
apps that
are no longer being developed. It is good for us (as developers) because we can 
test. 


>Another issue with binary compatibility is the compiler.  Right now 
>x86 BeOS is locked into a specific version of GCC.  A few months ago 
>there were discussions of what would be required to move to the 
>latest and greatest, and a Be engineer seemed to indicate that all 
>involved C++ code would have to be recompiled with the new compiler.
>
>So, by going for R5 binary compatibility we're also saying we won't 
>be upgrading the compiler any time soon.
>
>Right?
>
>-- 
>Scott Lindsey <wombat@xxxxxxxx>
>
>