Add mapsfrom and mapsfrom.long

[fee1-dead]

Refs: mapsfrom mapsfrom.long

Since this was easy enough I thought I'd just open a PR. Will discuss on Discord as well!

[Enivex]

I would prefer mapsfrom and mapsfrom.long for semantic reasons.

[fee1-dead]

Ah, it also matches stmaryrd's LaTeX command. Would there be any value in keeping mapsto.rev as an alias though? Since mapsto is much more wellknown and my first instinct is to type mapsto.rev.

[Enivex]

I don't think an alias pulls its own weight, especially because it's already an alias for arrow.l.bar. That's three names for the same symbol.

As an alternative, how about

maps
  .to ↦
  .to.long ⟼
  .from ↤
  .from.long ⟻

(There are bunch more arrows too)

Now you get discoverability, and less duplication. The cost being a period for the full name, but you would in fact be able to use the shorter maps and maps.long for the most common use case (or the shorthand).

[MDLC01]

My main concern with this is that mapsto was added as an alias because it is a very common symbol. But in my experience, I don't think I've ever seen ↤ in the wild. For this reason, I think keeping it as arrow.l.bar is okay.

Regarding @Enivex's proposal, "maps.to" is more annoying to write than "mapsto" (on my AZERTY keyboard, that's two additional key presses) and "maps" alone feels less readable. But if we decide to add a special name for ↤, I don't have a strong opinion on any on the options that were mentioned in this thread.

[knuesel]

I also don't have a strong opinion but one downside of x maps y is that it looks semantically incorrect: it reads as if x is mapping y while it's being mapped to y.

[fee1-dead]

I think after considering the comments above, the consensus roughly lies on mapsfrom since the downside of maps/maps.to have been given. So I have tentatively changed this to add mapsfrom.

[T0mstone]

Personally I like the symmetry here and think it's probably worth the added redundancy.

[dccsillag]

I also vote in favor of redundancy. I also strongly agree with @knuesel that x maps y looks very wrong.

[fee1-dead]

I'm not sure what proposals you two were referring too ^^ Let me know if I should change anything, as I am open any proposal as long as we add a reasonable shorthand for it.

[MDLC01]

Since I seem to be the only one arguing in favor of not adding this name, I'll say that if we add it I'd prefer it to be as mapsto.rev. ↤ is not common enough to warrant an entirely new name imo.

[fee1-dead]

My original proposal was mapsto.rev and mapsto.rev.long - I'm ready to revert it back to that if we do end up with a consensus for that.

[Enivex]

For the record, I'm strongly opposed to mapsto.rev. It's the worst of both worlds. Being neither semantic, nor descriptive.

↤ is not common enough to warrant an entirely new name imo

Less common than the one pointing right, but not uncommon. And it's not an entirely new name, just a corresponding semantic name to the already existing mapsto.

[MDLC01]

not uncommon

Do you have examples of documents where this is used? I genuinely don't think I have ever seen this symbol used in real life documents.

[fee1-dead]

for the record, I proposed this because my first year undergrad algebra course used it.

the original usage i forgot, but I imagine this to be quite useful when defining multiple maps, e.g. two sets mapped to the same target:

A1 -> B <- A2
a1 |-> b <-| a2

or one set mapped to two targets:

B1 <- A -> B2

[Enivex]

not uncommon

Do you have examples of documents where this is used? I genuinely don't think I have ever seen this symbol used in real life documents.

So as to not dox myself I'll refrain from posting any of my own papers, but if you do a google search for arxiv "\mapsfrom" you'll find a number of matches