MSE feature ideas

mse-feature-ideas.md

Here's my ideas on what could be added to the Magic Set Editor now that we have a way to build it. Obviously these don't necessarily have to happen in this exact order, so for example a bugfix could also land in 2.0.3, 2.1.1, 3.0.1, etc.

MSE 2.0.2 (bugfixes)

• fix off-by-one error when working with card in the Console tab (instead of using the card selected in the Cards tab, it uses the one below it)
  • there's a possibly related error where you can't get planeswalker cards as card, that should be fixed as well
• fix length not working with maps
• better RTL language support (see MightyD33r#5246's issues with making a Hebrew template)
• fix substring("foo", begin: 3) returning true instead of ""
• fix Cards tab card render not updating when styling data is changed

MSE 2.1 (backwards-compatible features)

• allow empty map literal [:] (like in Swift)
• give styles access to the flavor bar position (the position of the separator in a combined_editor field)
• add a type built-in function that returns the type of its input as a string
• allow access to the current date/time
• allow to_image(card) and passing cards as input to crop etc (this removes the need for an external DFC splitter tool)
• add a function/operator to check whether a map has a field
• allow calling an export template from the command line (this is already in the svn_history branch)
• allow searching the style list
• “shiny” (gradient) borders (for gold/silver/bronze) (these might be doable already, not sure)
• console history with arrow keys
• monospace font and syntax highlighting in console
• tooling for external IDEs (syntax highlighting, lints, jump-to-definition, etc.)
• tooltips on template list for long titles/subtitles
• allow “scale down to” on single-line text fields, to make “shrink type line” etc redundant
• option to sync text size on multiple fields, to make explicit planeswalker text size redundant
  • possibly redundant with ability to locate the flavor bar?
• import scripts (equivalent to exporters, to replace MSG)
• don't require MSE reload after export template changes
• self-updater
• macOS and Linux builds
• allow rotating the image in the image importer
• support for bulleted lists (with proper indentation after soft line breaks)
• allow access to a card's styling data in scripts

MSE 3 (backwards-incompatible features)

• make .mse3set a directory file extension, allow reading .mse3set.zip as well
• use card names and proper file extensions for image files
• remove unused images
• suffix .mse-{type}/{type} file with .mse file extension for editor support (syntax highlighting etc)
• allow versioning packages
  • notify when updates are available
  • better dependency management (something similar to what Rust's Cargo.lock does maybe)
  • allow upgrade paths including handling of new defaults (e.g. Ancestral generic mana, brown legendary Vehicle P/T)
  • bundle a link to a default package repository with MSE, so that it works as a standalone .exe without a data folder
• rename “HTML” export menu item to something clearer like “other”
• allow space indentation in data files
• maybe: modernize the scripting language
  • remove weak typing features in favor of better explicit type conversion tools
  • document what on earth is going on with \<, remove/replace it if it's not useful
  • improve debugging and error handling
  • allow importing MSE 2 code from MSE 3 code and vice versa, like with Rust editions
  • drop the special case that allows = instead of := when unambiguous with == due to precedence
• rework exporters to generate either a file or a directory, not both
• generate a lockfile to prevent simultaneous edits, e.g. when collaborating over Dropbox or Google Drive
• instead of saving a .bak file, persist edit history on disk (and add an option to export without edit history)
  • maybe some way to visualize edit history
  • give exporters access to edit history and allow revision options, to facilitate writing a changelog exporter
• real-time collaboration
  • requires a server component where sets can be hosted
  • login to tag edits with your username (could use the same accounts as a package repository, maybe just use Discord OAuth)
  • edit locking on a per-card basis instead of per-set
• invisible field support (to replace card notes commands)
  • link related cards so the link is displayed on both cards
  • this is different from just implementing them as style options since they ignore the “options specific to this card” toggle and are always specific
• allow assigning card fields explicitly in styles, e.g. card.name := "some other string"

• script functions: Right now I'm looking for some way of, given a card value (rather than the global card), either getting the style name, e.g. m15-altered, or checking whether the style has a given styling option (a function for checking whether a map has a field would be useful here).
• I'm not sure how much sense [:] as a separate concept from [] would make, but having that (in combination with type_name) would allow me to make the code for the JSON export template I'm writing right now much more readable.
• I would prefer moving to file extensions that don't have - in them since macOS Finder doesn't recognize file extensions that contain - as such, adding extra keystrokes each time you rename a file.

• You should be able to do somecard.stylesheet.name, except that this crashes MSE right now.
• For checking if a field exists there is currently the or else keyword. So you try to use it and if that fails you do something else, But that is not ideal.
• Right now vectors and maps that come from a script are internally the same type, so [:] doesn't make sense. If the types are split it would.
• Should I add a to_json function?
• How often do you have to rename a file? We could move to .mseset, .msescript, .msedata, etc., but I am not sure if it is worth the effort.

• A to_json builtin would only make sense with the map and vector types separated, otherwise you won't be able to generate both {} and []. Besides, I don't think it's the sort of thing that needs to be built in (but I suppose that depends on the overall approach to how much is provided as builtins)
• I rename set files quite often, though I admit that's part of my rather unique workflow as the maintainer of Lore Seeker.