Use Case: Recover what is drawn on the client by the user as (georeferenced) markup / feature objects

[prushforth]

Say we have a map, for which we provide a feature drawing API. The canvas on which that drawing occurs should be on the client, but it should be 'recoverable' as marked up features or even just plain GeoJSON, via a client API. To me, this suggests something like the following as a way to request user input:

<input type="location" subtype="point" (for example)>

Such an input would generate content on the map canvas with the assistance of the user. It might or might not be submitted, but the content could be recovered as georeferenced feature geometry content at least, perhaps via an API call on the map canvas, and perhaps also as an event from the input when user input was completed.

[cmhodgson]

Existing frameworks typically treat map drawing/editing tools as controls. Such editing controls work on editable layers, these layers could be either client side only, or you could allow editing of features retrieved from the server (as mapml vectors, or geojson?) The controls and/or layers provide event hooks that you connect into to perform more advanced functionality, eg. onChange, onCreate, etc. The layer in which the feature is created/edited would be used to access the feature objects.

[AmeliaBR]

This sounds like a very broad use case, it might be worth breaking it down.

Focusing just on pin-point features, what you're asking is an API for using a map as an input for selecting a location. I think there are two ways I could see that working (which both have their use cases):

• a form element <input type="location"> which works like a date picker or a color picker & triggers a browser-built pop-up that allows the user to select locations (maybe directly from a map, or from search, or from a list of saved locations that the browser has access to but the website doesn't). The website just gets the final location data (e.g., longitude & latitude). This all works completely separately from any maps drawn on the web page, although maybe you could associate it with custom map data in a manner similar to how you associate a text input with a <datalist> for auto-complete.

• an API for translating a point on a web map (e.g., from a user click or from cursors controlled with arrow keys) to longitude & latitude data, so that a web author could write a custom location picker

Drawing shapes and paths and so on, is more complicated. I can't see that being a built-in feature to start, but it should definitely be a capability that an author could build from lower-level APIs, generating features point by point, and then exporting them in an interchangeable format.