----- Original Message -----

From: Joe O'Farrell

To: csound@lists.bath.ac.uk

Sent: Thursday, August 23, 2007 7:44 AM

Subject: Re: [Csnd] cpspch updates

On 23 Aug 2007, at 05:15, Anthony Kozar wrote:

No orchestras will be broken but some might give slightly different output.

You mean they'll actually do what most of us probably thought they already did? ;-)

One more vote in favour of the change

Joe

One further point to throw in to this discussion - archived pieces could
always be rendered with archived versions of Csound, couldn't they (if
for some bizarre reason an archived recording is unavailable)?

Even in the relatively unlikely scenario that an older version of Csound
wouldn't compile/run, the code could still be inspected by any 23rd
century musicologists hunting for the authentic sound of the mid 1990s...

Yours truly,
David

Art Hunkins wrote:
>  
> ----- Original Message -----
> *From:* Art Hunkins 
> *To:* csound@lists.bath.ac.uk 
> *Sent:* Thursday, August 23, 2007 10:20 AM
> *Subject:* Re: [Csnd] cpspch updates
> 
> I also vote for the change.
>  
> Csound is simply *executing* a set of composer instructions. If Csound
> can improve on the accuracy of its execution, it should. (Interpretation
> is not Csound's job; in this medium, as a basic rule, interpretation is
> done by the composer.)
>  
> The only possible alternative I see to the "update issue" is adding an
> optional flag to existing opcodes, one that specifies higher resolution.
> Whether this would be appropriate/possible here I leave to the
> developers to determine. Such a solution, which has been successfully
> applied elsewhere, would certainly be ideal from a user perspective.
>  
> Art Hunkins
> 
>     ----- Original Message -----
>     *From:* Joe O'Farrell 
>     *To:* csound@lists.bath.ac.uk 
>     *Sent:* Thursday, August 23, 2007 7:44 AM
>     *Subject:* Re: [Csnd] cpspch updates
> 
> 
>     On 23 Aug 2007, at 05:15, Anthony Kozar wrote:
> 
>>     No orchestras will be broken but some might give slightly
>>     different output.
>>
> 
>     You mean they'll actually do what most of us probably thought they
>     already did? ;-)
> 
>     One more vote in favour of the change
> 
>     Joe
> 

Hummmmmmm  - getting that *Classis* 90's or 70's mainframe computer  
music sound...

i will tell you all that when I composed trapped in convert in the  
summer of 1979 at MIT on
a PDP11 with this custom built set of *floating-point* converters  
built by Analog Devices,
I have NEVER been able to get the ending to sound the way it did on  
that machine - using
any other version of Csound - even versions running on a PDP11  
computer at the University
of California in San Diego.

I revised and revised and revised the ending for about 15 years - I  
do have some old
analog tapes of the original.... but still.... I consider that we  
HAVE achieved backward
compatibility because we can still render this piece.

And I would vote to fix any and every bug we find in Csound - even 30  
year old bugs!
Congratulations Anthony for finding it!

-dr.B.


On Aug 23, 2007, at 12:57 PM, David Ogborn wrote:

> One further point to throw in to this discussion - archived pieces  
> could
> always be rendered with archived versions of Csound, couldn't they (if
> for some bizarre reason an archived recording is unavailable)?
>
> Even in the relatively unlikely scenario that an older version of  
> Csound
> wouldn't compile/run, the code could still be inspected by any 23rd
> century musicologists hunting for the authentic sound of the mid  
> 1990s...
>
> Yours truly,
> David
>
> Art Hunkins wrote:
>>
>> ----- Original Message -----
>> *From:* Art Hunkins 
>> *To:* csound@lists.bath.ac.uk 
>> *Sent:* Thursday, August 23, 2007 10:20 AM
>> *Subject:* Re: [Csnd] cpspch updates
>>
>> I also vote for the change.
>>
>> Csound is simply *executing* a set of composer instructions. If  
>> Csound
>> can improve on the accuracy of its execution, it should.  
>> (Interpretation
>> is not Csound's job; in this medium, as a basic rule,  
>> interpretation is
>> done by the composer.)
>>
>> The only possible alternative I see to the "update issue" is  
>> adding an
>> optional flag to existing opcodes, one that specifies higher  
>> resolution.
>> Whether this would be appropriate/possible here I leave to the
>> developers to determine. Such a solution, which has been successfully
>> applied elsewhere, would certainly be ideal from a user perspective.
>>
>> Art Hunkins
>>
>>     ----- Original Message -----
>>     *From:* Joe O'Farrell 
>>     *To:* csound@lists.bath.ac.uk 
>>     *Sent:* Thursday, August 23, 2007 7:44 AM
>>     *Subject:* Re: [Csnd] cpspch updates
>>
>>
>>     On 23 Aug 2007, at 05:15, Anthony Kozar wrote:
>>
>>>     No orchestras will be broken but some might give slightly
>>>     different output.
>>>
>>
>>     You mean they'll actually do what most of us probably thought  
>> they
>>     already did? ;-)
>>
>>     One more vote in favour of the change
>>
>>     Joe
>>
> -- 
> Send bugs reports to this list.
> To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk

Hello again,

I think Dr. B's story is a good example of the point Art Hunkins made
earlier about responsibility for achieving a certain result always
necessarily resting with the artist.

Any more or less elaborate scheme for making old pieces "render the same
as they used to" essentially shifts (or attempts to shift) that
responsibility onto Csound developers and onto Csound itself.

I think the cost of any such scheme (as they all MUST amount to some
form of code fork this is potentially quite large) needs to be weighed
primarily against the cost of simply saying "if you truly care about
archiving the code from your piece, archive the relevant Csound codebase
and binaries also."

In fact, I have done the latter for a number of years now, and it is
pretty easy.  Just a few clicks and drags...

(...and this has saved me on a couple of occasions when returning to an
old piece to rework it...)

Yours truly,
David

Dr. Richard Boulanger wrote:
> Hummmmmmm  - getting that *Classis* 90's or 70's mainframe computer
> music sound...
> 
> i will tell you all that when I composed trapped in convert in the
> summer of 1979 at MIT on
> a PDP11 with this custom built set of *floating-point* converters built
> by Analog Devices,
> I have NEVER been able to get the ending to sound the way it did on that
> machine - using
> any other version of Csound - even versions running on a PDP11 computer
> at the University
> of California in San Diego.
> 
> I revised and revised and revised the ending for about 15 years - I do
> have some old
> analog tapes of the original.... but still.... I consider that we HAVE
> achieved backward
> compatibility because we can still render this piece.
> 
> And I would vote to fix any and every bug we find in Csound - even 30
> year old bugs!
> Congratulations Anthony for finding it!
> 
> -dr.B.
> 
> 
> On Aug 23, 2007, at 12:57 PM, David Ogborn wrote:
> 
>> One further point to throw in to this discussion - archived pieces could
>> always be rendered with archived versions of Csound, couldn't they (if
>> for some bizarre reason an archived recording is unavailable)?
>>
>> Even in the relatively unlikely scenario that an older version of Csound
>> wouldn't compile/run, the code could still be inspected by any 23rd
>> century musicologists hunting for the authentic sound of the mid 1990s...
>>
>> Yours truly,
>> David
>>
>> Art Hunkins wrote:
>>>
>>> ----- Original Message -----
>>> *From:* Art Hunkins 
>>> *To:* csound@lists.bath.ac.uk 
>>> *Sent:* Thursday, August 23, 2007 10:20 AM
>>> *Subject:* Re: [Csnd] cpspch updates
>>>
>>> I also vote for the change.
>>>
>>> Csound is simply *executing* a set of composer instructions. If Csound
>>> can improve on the accuracy of its execution, it should. (Interpretation
>>> is not Csound's job; in this medium, as a basic rule, interpretation is
>>> done by the composer.)
>>>
>>> The only possible alternative I see to the "update issue" is adding an
>>> optional flag to existing opcodes, one that specifies higher resolution.
>>> Whether this would be appropriate/possible here I leave to the
>>> developers to determine. Such a solution, which has been successfully
>>> applied elsewhere, would certainly be ideal from a user perspective.
>>>
>>> Art Hunkins
>>>
>>>     ----- Original Message -----
>>>     *From:* Joe O'Farrell 
>>>     *To:* csound@lists.bath.ac.uk 
>>>     *Sent:* Thursday, August 23, 2007 7:44 AM
>>>     *Subject:* Re: [Csnd] cpspch updates
>>>
>>>
>>>     On 23 Aug 2007, at 05:15, Anthony Kozar wrote:
>>>
>>>>     No orchestras will be broken but some might give slightly
>>>>     different output.
>>>>
>>>
>>>     You mean they'll actually do what most of us probably thought they
>>>     already did? ;-)
>>>
>>>     One more vote in favour of the change
>>>
>>>     Joe
>>>
>> -- 
>> Send bugs reports to this list.
>> To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>