LWN.net Logo

Preferred form

Preferred form

Posted Mar 5, 2011 17:14 UTC (Sat) by ewan (subscriber, #5533)
In reply to: Preferred form by corbet
Parent article: Commitment to Open (Red Hat News)

rather than, say, object files or preprocessor output

I think in practice it's usually a bit stronger than that; the 'preferred form' is the actual source, excluding machine generated files. That excludes preprocessor output, as you say, but also plain C code generated from other things, and, I think it's at least arguable, a big single patch that's been generated from something else.

The idea being that you can modify and rebuild the program.

Again, I think it's a bit stronger than that - you could, after all, rebuild something from pre-processor output too, it would just be hard work. It seems to me that it's generally taken to mean that you should have access to the same source as the person distributing the code - if they've written machine code directly with a hex editor, that's the source, if they've generated it from assember, then that's the source, and if it came from a big block of C, then that's the source. And if it came from a VCS, a big set of patches, and scripts to assemble the whole lot, then that is the source.


(Log in to post comments)

This is interesting idea, but...

Posted Mar 5, 2011 20:39 UTC (Sat) by khim (subscriber, #9252) [Link]

as I've already noted before Oracle (or anyone else) will have any hope to prevail in court they should answer one simple question: 20 years ago it was perfectly Ok to keep sources in private VCS and only present tarballs to the world (FSF did that, Cygnus did that, etc - it was normal M.O. back then). If, as you assert, times changed and now it's not Ok then you must explain what exactly happened between then and now to change the equation. VCS is not a new idea, you know: SCCS was released 38 years ago.

Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds