On Thu, Oct 23, 2003 at 11:37:01AM -0800, Iain Duncan wrote:
> > We can't just change the output value as that can be sure to break an
> > important orch somewhere. Unless of course we can be sure nobody
> > actually uses that opcode!
> 
> Well, I wonder whether Csound5 is a big enough overhaul that it is really
> such a bad thing if a small percentage or orchestras need a small change
> made to run. I know this is not supposed to happen but  a change log file
> somewhere could make it clear what the changes are, so that changing is
> simple, and csound is small enough that it would not be hard for people to
> either find the right build if indicated with the older piece or to include
> the right build with the piece in any kind of historical archiving.
> 
> I just question whether perhaps we are hampering the development too much in
> order to *never* violate this historical principle . . . Personally as a
> user I would have no problem editing orchs to run with Csound 5 or simple
> using older csound to run ones I can't be bothered to edit.

In my opinion you're always better off preserving backwards
compatability by doing one of the following things:

* Making OpCode2 which does the same thing but has a different return
value - this is unattractive because you're duplicating a lot of code.

* Making an extra argument which sets which format the return value
should be in.

I guess it's another option as you say to make it publically known 
that Csound5 will break old orchestras so that people can publish 
their pieces as "pre-csound5" and "post-csound5".

More cents,

Chris.
_________________________________
chris@mccormick.cx
http://www.mccormick.cx
http://www.hypercube.com.au
http://www.sciencegirlrecords.com