/preparation/to-be-decided.md
Markdown | 117 lines | 85 code | 32 blank | 0 comment | 0 complexity | 745d7d5a3af8ded02c1b72903f70c2a1 MD5 | raw file
- # Topics to be discussed before music entry
- This document is a work in progress and is open to modification.
- Particularly encouraged is the removal of obsolete topics ;-) .
- ## Score and part structure
- As the instruments make use of some heavy *divisi* and staff splitting we have
- to make sure before that we can efficiently make use of the entered music.
- As for the cello part I have an idea:
- - create four parts (as that's the maximum)
- - enter music for part 1 first
- - enter music for additional parts,
- adding hyperlinks for all segments where the part doesn't have individual
- music. That should ensure complete parts.
- - The score should also have four cello staves (frenched), but I think we'll
- have to manually prepare them by picking the right segments where necessary.
- ## Cue notes
- We should be clear about the fact that we will have to enter cue notes at a
- later point. I think we should send the finished parts to the orchestra and
- ask someone there to decide about where to enter cue notes in the parts.
- Then we have to be able to take care of that painlessly.
- Maybe that's not an issue at all, but I don't know about it, so I prefer being
- overly cautious ...
- ## Order of music entry
- There are two aspects to that question:
- - we need a MIDI/Audio file until october. So we need to have a complete
- "outline".
- - There is so much music in parallel that it makes sense to work in segments.
- That is: one enterer takes one segment and works his way through the wnole
- score.
- ## Identifying ambiguous elements in the score
- e.g. hairpins not aligned consistently
- ## Use of markup functions
- It's important to define rules and "stylesheets" to make the markups
- consistent. I think we need a) styles for certain types of markups (e.g.
- playing indications) and b) commands to directly enter default stuff (e.g.
- "pizz.").
- ## General coding standards
- We should at least decide about authoritative standards concerning
- - indentation
- - (each measure begins at the beginning of a line,
- everything else should be indented
- - more than one measure on one line is only acceptable
- in the case of multimeasure rests
- - substantial (i.e. long) tweaks or commands should go on a line of its own
- - barchecks after each measure
- - barnumber comments
- (I suggest an empty line and a line with only a barnumber comment
- for *every* measure)
- - an authoritative list of available commands (from `oll` or from
- a project library
- Should we also define more concrete rules? :
- - when to use whitespace and when not: e.g. `e4~ e` vs. ` e4 ~ e`
- - order of attached items, e.g.
- "note(duration)->articulation->dynamic->beam->tie/slur"
- I know this may be intrusive, but I think also that it's important to have
- a consistent coding, first for ourselves to be able to read others' code
- better, and then when we want to use the project as a reference/demo.
- ## Define rules how to fill out the templates
- ## Define how to generate new part templates
- I'm afraid the Python script(s) are somewhat messed up.
- ## Define rules how to deal with issues in the original score
- Anybody who notices an issue in the original score should handle that instance
- somehow. Obvious errors should be fixed immediately, but only when documented
- properly. In doubt one should faithfully copy the original but also document it.
- For this `\annotate` has to be fixed to be at least usable for music entry.
- ## Branching workflow
- I think only peer-reviewed content should ever go into `master`.
- That means any music entry should be done in a branch and finished with a
- pull request. However, work doesn't have to be completely finished for that.
- I.e. if I enter five segments of a part I can ask for review.
- Probably it would be best if there are teams of two that do their reviews
- mutually.
- These branches should be quite short-lived because we want `master` to document
- the progress. I think any push to master should trigger a build of the project
- that is then displayed publicly.
- ## Communication channels
- I think we should use a mailing list and an issue tracker, the issue tracker
- only for real issues (bugs, persistent questions and TODOs) and the mailing list
- for usual questions ("does anybody currently work on the harp part?",
- "how do I achieve this?" or "Can somebody confirm this should be a d flat?" etc).
- Although in general tools like Trello serve such needs well I think a proper
- (and dedicated) mailing list will be good too, without forcing contributors
- to sign up with Trello.
- Or should we (additionally/alternatively) use a forum? I've installed
- [Vanilla forums](http://vanillaforums.org) on https://git.ursliska.de/forum as
- a test