Mark,

This is really good. The bottom line seems to be:

        - if we build it, we will use our own interface stability  
classifications, and
          deliver them via language-specific documentation facilities

        - if someone else built it, and there is IF stability info, we'll  
support it via
          the language-specific documentation facilities

        - if someone else built it, but IF stability info is inadequate, we  
can add a man
          page, and urge the team to add IF stability info to the language- 
specific doc.

Since man pages are the de facto language-specific IF doc for C/C++,  
can we simplify the third option to state:

        - if someone else built it, but there is inadequate IF stability  
info, we urge the team
          to add IF stability info to the language-specific doc.

In other words, add a TCA to the opinion. I don't think we add value  
by requiring a man page in one specific situation.  If man pages are  
not consistently generally available for Java/Python/etc., I don't  
think an occasional page will help anything.
-tdc

On May 19, 2009, at 3:23 PM, Mark A. Carlson wrote:

> During the case discussion today, I took the AI to help Michael  
> Kearney draft
> a minority opinion. There may be other minority opinions, but if  
> this looks
> close to something you would sign on to, I am open to small changes.
>
> -- mark
>
>
> 5. Minority Opinion
>
>     Background
>
>     It is not typical for programmers working with non C/C++/Assembler
>     files, such as Java Jar files, to determine the
>     Exported Interface stability level using the man command.  Java
>     programmers depend on Javadoc, Python programmers
>     depend on pydoc and so forth to document interfaces and the
>     stability would best be indicated there.
>     Approval of OpenSolaris projects have been inconsistent in
>     preferring man pages or native documentation.  This opinion seeks
>     to clarify the issue and define a policy for all such cases going
>     forward.
>
>     Best Practice
>
>     Case A - Sun Developed Components
>
>         1) Sun project team developing a Jar file shall document the
>     ARC interface classification in the native documentation. (i.e.
>     Javadocs)
>
>     Case B - Components imported from external OSS Communities
>
>         1) The OSS Community documents the interface classification in
>     their native documentation
>
>             a) OpenSolaris project team agrees with the classification
>     and supports it
>                 - Javadoc or other native documentation required
>     (unchanged)
>                 - No man page shall be allowed
>
>             b) OpenSolaris project team disagrees with the  
> classification
>                 - Javadoc or native documentation required, but
>     project team must change the OSS documentation
>                 to match the project team's classification
>                 - No man page shall be allowed
>
>        2) OSS Community does not document interface classification in
>     their native documentation
>             a) OpenSolaris project team is strongly encouraged to
>     update the native documentation to reflect the OpenSolaris project
>     team
>             classification.
>                 - Changed Javadoc or other native documentation  
> required
>                 - No man page shall be allowed
>             b) OpenSolaris project team cannot support deltas to the
>     native documentation
>                 - Unchanged Javadoc or other native documentation  
> required
>                 - A man page shall be provided
>
> _______________________________________________
> opensolaris-arc mailing list
> opensolaris-arc at opensolaris.org
>


Reply via email to