New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
RFC: Rename Future to Task in public facing interface #8028
Comments
+1. In general, this fits nicely into our nomenclature: A We can still call it a |
I actually don't have a very strong opinion here. I'm mildly concerned about confusion internally and about migration efforts (for users). If we believe this is worth it, I'm fine with it. If we choose to rename |
I don't think that we should rely on asyncio as a model of intuitive behavior 🙂 My mental model of a naive user finds "submitting a task" and "creating a task" to be similar in terms of complexity. My gut reaction is to leave submit alone for now.
My sense is that we can probably leave aliases here for a long time. |
cc @jacobtomlinson and @jrbourbeau for when he gets back |
Yeah I like this idea. I spend a lot of time thinking about tasks in an async context and I think Dask could totally adopt the same thing. While I agree we shouldn't use asyncio as a model of intuitive behaviour I do think async programming in general has some common terminology that we could use. I like the I think being able to talk about the "Tasks API" rather than the "Futures API" will be more accessible to new users when teaching too. |
@fjetter is this something that someone on your team can do, maybe early in 2024? Alternatively @jacobtomlinson maybe this is something that people in NVIDIA want to help with? It seems like a pretty accessible task. |
Please don't make too much of this issue. This is mostly to start a conversation.
I spent the week with a bunch of folks who are less sophisticated than our typical users, but who really need something very much like Dask. It was educational.
My sense is that our usual idiom:
... is kinda weird for them. Passing the function as an argument is strange (another proposal for that in #7936). Also, folks mentioned that the term "future" doesn't really make any sense to them. Instead, the term "task" made more sense. This doesn't seem unreasonable. We chose the term future to be consistent with the
concurrent.futures
module, which is mostly just known by devs.I'll leave the decorator stuff for another conversation, but here I'd like to propose the following cosmetic API change:
The text was updated successfully, but these errors were encountered: