first beta of MEI Basic

[kepper]

this adds a draft for MEI basic, based on our discussions in Oxford. It's still not complete and needs significant consideration of:

• how much metadata should be included
• make sure controlevents have one method for placement only (but potentially a different for every type)

We could continue to develop in this branch, or merge it into develop.

[lpugin]

not sure about meterSig being excluded.

[kepper]

The spreadsheet isn’t addressing everything, but we made a statement there about meterSig to be excluded. I think the idea was to encourage people to use scoreDef between measures instead. Basic should exclude meters changing in the middle of a measure (no clue how this would work anyway…)

[lpugin]

The spreadsheet isn’t addressing everything, but we made a statement there about meterSig to be excluded. I think the idea was to encourage people to use scoreDef between measures instead. Basic should exclude meters changing in the middle of a measure (no clue how this would work anyway…)

Makes sense. Is keySig excluded as well? It should be

[kepper]

If I remember correctly (in need of a power outlet to verify…) the argument to keep keySig was that keys may actually change in the middle of a measure. However, we should make sure that it’s not used inside scoreDef…

[ahankinson]

I would like to see us address #582 first before merging this PR. MEI Basic shouldn't land as part of 4.0.1.

[kepper]

We now have a first beta of MEI Basic. It hasn't been tested to great detail yet, though, so please have a look. I'm particularly curious to hear from @rettinghaus and @lpugin if it has everything that's needed for rendering with Meiler and Verovio, respectively.

[kepper]

Dear @music-encoding/core-contributors, dear all, I finally managed to get around with the draft of MEI Basic, as we've discussed in the past. Please have a look and see, I surely missed some aspects. It is pretty restrictive, but is build to allow integration of additional features in the future without causing many disruptions -> very often, we picked elements, but disallowed all advanced attributes on them for the time being. For the @meiversion-attribute, the allowed value is 4.0.1-rc1+basic. I like the +basic part at the end (which is a possible part in semantic versioning), because it allows us to indicate MEI profiles in the schema. We may even have multiple profile combinations (something like cmn + mensural), and control them through a regex, instead of a value list as we have right now. However, that'll be a separate issue to discuss… 

[kepper]

btw., could someone more familiar with Travis look into the build failures? They don't seem to originate from problems in the repo, but in the Travis setup. @ahankinson, are you able to help out?

[martha-thomae]

btw., could someone more familiar with Travis look into the build failures? They don't seem to originate from problems in the repo, but in the Travis setup. @ahankinson, are you able to help out?

I think the Travis issue is the one I fixed on the pull request for the mensural module (see 4fbb831). I was the first one who got the building error, so I added this fix in the PR so that, when accepted, it would build fine for all. (I don't know Travis, this was with the help of someone from my Lab that does. It built ok after that commit.)

[lpugin]

@stem.dir missing on <chord>
@fontname missing on <rend>
@accid is allowed on <note> but not @accid.ges
@vgrp missing on <dynam> (I would still consider basic, but maybe not everybody would agree)
@oloc and @ploc on <rest> (ditto)

[lpugin]

For evaluating it with MNX, we should use the test files provided in their converter https://github.com/w3c/mnxconverter/tree/master/tests
(We could easily have an importer in Verovio for them. To test with MEI basic we should add an option to set a specific schema in the MEI output, but this is straighforward)