Re: MNX Proposal now Available [via Music Notation Community Group]

Hi Rafa, Joe, Dominik,

Remember that MusicXML is primarily a /graphics/ standard, designed for 
sheet music.
The first pass defining MNX should define it purely as a /graphics/ 
standard, ignoring any (temporal) meanings. What the elements /mean/ is 
a completely separate question that can be dealt with later.

In this first pass, MNX can borrow as much or as little as it likes from 
MusicXML. Its just not allowed to deal with anything temporal. A <tempo> 
is just a legible (graphical) string. A <tuplet> is a graphical object 
that brackets <event> symbols on the page.
<tuplet> graphics are important for /legibility/ in CMN. They tell the 
reader how the bracketed <event> symbols align with other <event> 
symbols on the page. But not all notation systems have <tuplet> 
brackets, so they should be optional. So should <tempo> markings.
There should probably be an enumeration describing the graphical form of 
the <event> symbols. They could be CMN <event> symbols, Asian 
tablatures, neumes of some particular type etc. If they are CMN symbols, 
for example, they have noteheads, augmentation dots, stems and flags, 
ornament decorations, beams etc.

My advice would be not to get bogged down in tangential arguments, but 
to get on with the job of defining the graphical structure of MNX. I 
have, myself, too little experience with MusicXML to help much there, 
but its very nice that Dr. Dominik Hörnel from capella-software has 
joined us! :-)

All the best,
James


Am 31.03.2017 um 03:10 schrieb Rafael López García:
>>
>>     I bet that MNX stands for Music Notation XML, and then what you
>>     want to save is the notation, not the music itself (otherwise you
>>     would use an approach more similar to MIDI).
>>
>> Not bad, and you didn't even have to read the spec!  :-)
> But I read the one of MusicXML. And I wrote a MusicXML parser and a 
> Java-2-MusicXML converter in Android. And I may write another for MNX 
> in the future.
>>
>> Yes, this is about saving a score, which is instructions on how to 
>> perform music, as distinct from music itself.
>>
>>     And in my experience as a programmer I also think that XML is not
>>     a very good format to support "operations" (move a note to
>>     another measure, etc.), only serialize and de-serialize. I think
>>     the XML should be transformed to a object structure in memory and
>>     the operations should be performed to those objects, then saved
>>     again. And here you can choose to save all the XML again or
>>     replace only the one referred by the objects that have changed,
>>     but that is a matter to decide by every software developer (for
>>     example, my Kunkunshi Editor saves everything again as the XML
>>     parser I use is very simple).
>>
> From an informational point of view they are they are not completely 
> the same. A human-readable music score is more than the instructions 
> really needed to perform music. There are elements that are purely 
> organizational (e.g.: the measures) and others that are merely 
> aesthetic. You don't really need them to perform music, and they are 
> actually introducing a certain complexity that could be avoided, that 
> is why Dr. Hörnel argued the possibility of not having them.
>
> And actually... that is my case. When I wrote the aforementioned 
> MusicXML parser, I did it for Kunkunshi Editor, an editor of Okinawan 
> music. Okinawan music is NOT represented in the western fashion with 
> staves and measures, but with a different system called Kunkunshi 
> (工工四). So I also found all those organizational and aesthetic elements 
> completely unnecessary for my use case. And when I write the parser 
> for MNX in the future, it will happen again. But I understand that 
> this tiny and infrequent use case, as medieval music notation, will be 
> out of the scope of your standard.
>> As you say, one shouldn't use XML itself to perform these operations 
>> (and I hope no one does).
>>
>> Standard modern practice is to parse XML into memory using a Document 
>> Object Model (DOM) parser and manipulate the in-memory structure via 
>> the DOM API -- a web standard, see 
>> https://developer.mozilla.org/en-US/docs/Web/API/Document_Object_Model. 
>> It can then be serialized back out.
>>
>> This practice is so ubiquitous that web programmers like me will talk 
>> in terms of manipulating the document, when what we really mean is 
>> manipulating the DOM in memory!
>>
>> ...Joe
> Yes I know what a DOM parser is, I am tired of programming them. But 
> when I read Dr. Hörnel's email it looked like somebody was opening the 
> door to directly perform operations there. I am glad you clarified 
> that is not the case.
>
> Best regards,
>
> Rafa