Re: A must-see for anybody on this list

[Joseph Rushton Wakeling]

On 02/14/2013 01:36 PM, David Kastrup wrote:
> a) a reliable and scaleable mechanism to make individual problems go
> away by manual labor.  WYSIWYG systems offer that.  I think that
> Frescobaldi tries offering a bit of that as well.

The really simple way of putting this: "It needs to be as easy as possible to 
tweak stuff where you don't like the automatic results."

From a publisher point of view I think also ease of creating a "house style" 
probably plays a large part.  I don't personally find it very easy to create 
something like a "style sheet" for LP.

> Sibelius has thrown away mechanism b) by dismissing their development
> team.  Its problems are there to stay, but you can fix up a lot of them
> manually.

They closed down the longstanding UK-based development and R&D team, but it's a 
bit of a stretch to say that they will no longer be offering commercial 
development support on Sibelius -- they're forming a new development team in 
Ukraine.  It may not be as high quality as it was previously.

> Without a), however, it is hard to maintain a distinguishing edge over
> competition.  The question is how important that is.  Have we lost
> distinguished book publishing houses with significant focus on classical
> books due to public domain _texts_ being in a much easier malleable form
> than public domain _scores_?  So-so.
>
> They do a lot of publishing of _new_ texts.  But the market for new
> _music_ is a bit shallow.  There is pop music published quite a bit, but
> it does not really overly exercise LilyPond's strengths all that much.
>
> And "modern classics" are not really in circulation to a degree where
> large houses can live from them.
>
> Current publishers are still living from "editions".  If we get reliable
> optical music recognition (like OCR for classic _texts_) to LilyPond,
> this market will be getting strained.

Well, define what you mean by a musical "text" (or for that matter a literary 
one).  The problem of most historical works is that there isn't _a_ text -- 
there are multiple conflicting texts where there needs to be very careful work 
to unravel which of the possibilities is most likely to be the one the composer 
or author really intended.  This is where you need an edition that covers the 
evidence available, the alternative possibilities at various moments, and which 
gives a rationale for individual editorial choices (with enough information for 
you to make a personally informed decision about whether to accept the editorial 
decision or not, and what you might choose instead).

Consider what you get in IMSLP -- an out of copyright score, which in practice 
usually means a 19th-century edition prepared by an editor who most likely had 
limited source material available and who was willing to liberally sprinkle his 
own interpretative preferences all over the music.  Contrast that with what you 
get with a high-quality modern Urtext edition (although even here you have to be 
careful -- I've found scores published as "Urtext" which display a shocking 
level of editorial intervention).

Even if IMSLP gets (say) the first edition of a work, very often that 
publication is multiple steps removed from the composer and may even date from 
after their death -- not to mention the errors that may have crept in during the 
publication process.  IMSLP has very nice high-quality scans of the first 
editions of the Debussy preludes, for example -- where it's readily possible to 
identify a number of errors that stem from Debussy's notational indications 
clashing with the "rules of engraving" that Durand's craftsmen operated by.

Put simply, the ability to easily reproduce a given musical text fails to 
reflect the value of what _good_ music publishers (like Henle) actually do.