UNIX admin writes:
> > Despite assertions to the contrary, Sun does break
> > certain
> > compatibility from time to time and has guidelines
> > for doing so.
> 
> Might you be kind enough to point me to that document, so that I can, uh, 
> "warm up the chair"?

See:

  http://www.opensolaris.org/os/community/arc/policies/interface-taxonomy/

The short answer is that (a) it depends on the type of release (bigger
changes are allowed in a "major" release, such as SunOS 4.1 -> Solaris
2.0, versus a "minor" or a "patch") and (b) it depends on transition
plans and documentation.

In order to make an incompatible change to an interface we'd otherwise
promised not to change, the typical process is:

  - The interface is marked Obsolete first.

  - Documentation is sent to the field indicating that the interface
    may be going away, and what users can do about it to prepare.

  - The interface may not be touched until at least one year has
    elapsed since the time the documentation was published.

Obviously, it requires ARC approval to do something like this.
Depending on the effect of the change, it's possible the ARC may say,
"sorry, you still can't do that no matter how much notice you give or
what release binding you have."  We'd probably say that if someone
wanted to break printf(3C).

A lot of it depends on the transition plan (is there something
reasonable the user can do?) and the circumstances around the change.
For another example, if we were legally enjoined from shipping
something or if there were some unfixable security or data integrity
problem in an interface, then compatibility just doesn't matter.  We'd
break it on purpose, and probably quite quickly at that.

I don't think there are *simple* answers here, because it depends on
context, but if you care about the issue deeply, then pay attention to
what's going on in the ARC community, because that's one of the key
places where changes must be reviewed.

(Or maybe it's just helpful to know that there are some mostly
sensible people involved, and it's not a free-for-all.  :->)

-- 
James Carlson, Solaris Networking              <[EMAIL PROTECTED]>
Sun Microsystems / 35 Network Drive        71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Reply via email to