Re: Proposed new available and recommended behavior of \relative

[David Kastrup]

Evan Driscoll <address@hidden> writes:

> On 03/10/2013 03:50 AM, David Kastrup wrote:
>
>     The problem I have with talking much about \relative f is that f seems
> arbitrary.  However, maybe an explanation linking both of these concepts
> and explaining how f is arrived at will allow both views to coexist.
>
> That's what I was trying to get at with the second suggestion I was
> making: "Note that when a pitch is written relative to f, the relative
> and absolute representations of the note are the same."
>
> You could even make this a stronger statement: "The reference pitch f
> was chosen because notes written relative to f have the same
> representation as their absolute pitch."
>
>     Quite unlikely.  This conversion rule does not touch code it does not
> understand.
>
> That's certainly believable -- but the problem is that "{ \rhythm g }"
> looks locally like something that it actually does understand, as it
> looks like just a use of a music variable followed by a pitch.

First, it does not try interpreting things like \rhythm as a music
variable.  I was somewhat tempted to make an exception for \global since
\global is used in a rather consistent manner in the docs, and
considering it as neutral material would get quite a few more converted
use cases.

At any rate, take a look at

\rhythm = { d e f g }

\relative c' { \rhythm a }

converting this to \relative { \rhythm a } would be utterly wrong since
it would result in { d e f g a } instead of { d' e' f' g' a' }.  So even
assuming that \rhythm is a music variable does not make for a working
conversion strategy.

> To understand that it doesn't understand that, it would have to do at
> least enough parsing of the definition of \rhythm in order to
> determine that it is a scheme function that takes a single argument.
> Does it do that? I don't know.

It just keeps its hands off.

-- 
David Kastrup