Looking at features around Lonsdale, North Vancouver, I would expect to see both of:
But I only see the federal district mapped. Looking in a few other places around Canada, the same issue arises. Is the absence of provincial electoral boundaries intentional? If not, I could pull them in (e.g. GIS Spatial Data | Elections BC). If so, rationale?
I suspect the license hasn’t been approved. No license is listed at all on that page, but looking at the closest result in the open data portal I could quickly find, I don’t see the “Elections BC Open Data Licence” listed on the wiki page for Canada or the LWG’s Canada page. It looks extremely similar to the approved BC OGL 2.0, to the point I suspect they’re effectively identical, but because it’s more than just changing the name (e.g., it removes a sentence on more specific attribution statements for information providers, swaps “to the maximum extent permitted by law” with “of any kind and on any legal basis”, etc.), and I’m not a lawyer, I can’t make that call.
I see your point! I’ll reach out to legal-questions@osmfoundation.org as per Licensing Working Group - OpenStreetMap Foundation.
Yeah, I would really like to see the provincial- and territorial-level electoral districts added to OpenStreetMap as well. There is going to be a lot of work that needs to be done to get every province and territory onside with such a project, in terms of both data availablity and licencing. But another question is, how would the provincial and territorial electoral districts be tagged to differentiate them from the federal ones? Would we have to use something like “political_division=legislative_const”?
The distinction between “parliamentary” and “legislative” constituency might not be very obvious to many Canadians, let alone others 
Not to mention some provincial legislatures don’t have “Legislative” in their name.
The wiki page Key:political_division - OpenStreetMap Wiki has a few examples of namespaced values (e.g. NL:kamerkieskring, TR:TBMM) - something like that could work here as well. Like political_division=CA:BC:provincial, political_division=CA:NL:provincial? Or maybe just political_division=CA:BC, because I don’t recall different kinds of ridings at provincial level? And the federal ridings would be political_division=CA?
Please also consider the following: OSM generally only maps what is visible (and therefore, verifiable) on the ground.
Exceptions from this general rule require a very good reason.
Administrative boundaries are generally considered to pass this test, because not only are administrative boundaries important in the lives of many people, they also frequently help us in our mapping work.
Other boundaries may also be important enough to warrant inclusion in OSM but this depends on the boundary type and the country. But always remember that while you don’t need a reason to map a tree in your backyard (the tree being there is reason enough), you do need a good reason for mapping any kind of boundary that is not visible on the ground. “A good reason” will usually be “this is an important boundary for many people living here” or “this is important for mapping”.
Religious administration boundaries or small electoral districts, for example, will frequently be met with skepticism and may be deleted depending on local community consensus, and the delivery zones of Paul’s Perfect Pizza will definitely be deleted again - by me, if necessary, even if the local community thinks they’re important 
My suggestion would be to avoid investing a lot of work in importing boundaries that carry a risk of being thrown out later due to not passing the “important enough to have in OSM even though not visible on the ground” test.
Generally such boundary data is a foreign object in OSM because it is not subject to surveying and improvement by the OSM community - we’re just importing that from some source and then carrying it around as a dead weight, we can’t work with it. Therefore think twice before you burden OSM with it.
A strong opening. That is why we have German postcodes as polygons. So continuing on with exceptions to “generally”…
Riding boundaries are important for people living in them. Especially in first-past-the-post parliamentary systems, where your representative is intended to be your contact with legislative matters, and often a representative and appeal capacity when dealing with bureaucratic and policy matters.
In Canada, provincial governments are easily more important than those. They have constitutional status and fund and control directions of health care, education, transportation, land use, and more. And the ridings are the constituents’ connection and the way to influence that government policy. They generally contain about 10 to 100 thousand electors, depending on the province.
I can certainly appreciate an “only map what’s visible” razor. But I see many exceptions: land districts, postal codes, municipal neighbourhoods, bird sanctuary boundaries, etc. I’m definitely not asking for a witch hunt of “these boundaries don’t meet the razor, delete them!”. I’m more curious why provincial electoral boundaries would be excluded, especially when federal electoral boundaries and municipal neighbourhoods are included.
Is there an OSM equivalent of Wikipedia:Notability - Wikipedia to which we might refer?
The most major OSM guidelines are in Good practice - OpenStreetMap Wiki. Major criteria for inclusion include Verifiability - OpenStreetMap Wiki, and mapping what exists now (vs what existed in the past, or what might be planned in the future). There is no notability criterion.
These electoral district boundaries can easily be verified by checking the maps on the websites of each respective province and territory’s electoral agency. The provincial and territorial electoral district boundaries are not currently in OSM because the machine-readable data sets that depict them have not had their licences checked for compatibility, and so the easiest way at this point for us to add these boundaries to OSM would be to use exact textual descriptions—e.g., metes and bounds—of the boundaries of each district, and then use our own already-existing OSM data to paint the boundaries. This way we would not risk directly copying from other maps.
“Small electoral districts will frequently be met with skepticism”, and meanwhile we’re talking about electoral districts that in some cases are bigger than most European countries. 
Just as an aside, since Canada Post still considers postal codes in Canada to be proprietary data, we really should not be adding them to OSM.
Canadian postal codes don’t have geographical boundaries, so we shouldn’t add them to the map as boundaries. However, adding them as addr:postcode tags on buildings and POIs as appropriate is perfectly acceptable, there’s just no licence to wholesale copy them from Canada Post’s database.
Canada Post is the only entity considering postal codes as proprietary. They always lost or abandoned in court when challenged on this. You can use postal code anywhere, these are not protected by any means.
That is true. Even if you buy the postal codes boundaries from Canada Post, you will get errors, especially at street corners, because the postal codes are splitted using artificial boundaires. You should always validate the exact address to get the correct postal code. Many buildings have multiple distinct postal codes associated with them.
I don’t want to get too off-topic, but I had made this comment in another thread about the folly of adding postal code boundaries in Canada:
To try and steer the course back in the initial direction, currently:
Moving forward, should provincial electoral boundaries be included as well? Should federal electoral boundaries be removed? Should I forget I ever opened this can of worms?
(assuming this refers to B.C., because e.g. we have provincial boundaries in place at least in southern Ontario)
IMO they can be. But that doesn’t mean someone will do it. It does mean that you can do it.
IMO they shouldn’t be removed, they’re already there and they’re as verifiable as any other administrative boundary.