Open mrocklin opened 1 year ago
cc @scharlottej13 we have to update blog posts if we rename the repository
To be clear, this is mostly a question first. We can figure out operational stuff later if we want
(and we could also have redirects for a while, we don't necessarily have to go through and update old posts)
On Fri, Oct 27, 2023 at 8:47 AM Patrick Hoefler @.***> wrote:
cc @scharlottej13 https://github.com/scharlottej13 we have to update blog posts if we rename the repository
— Reply to this email directly, view it on GitHub https://github.com/dask-contrib/dask-expr/issues/372#issuecomment-1782950274, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACKZTBVPMIQGDMWKFTIUUTYBO3OXAVCNFSM6AAAAAA6S2XQBSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTOOBSHE2TAMRXGQ . You are receiving this because you authored the thread.Message ID: @.***>
We tend to forget things like this, so I'd rather make others aware of the possibility early on
I'm certainly open to a change while the project is still young. Did you have certain ideas in mind? I can imagine something like "dask/data-primer".
I can imagine something like "dask/data-primer".
I'm not sure I understand this. Can you elaborate?
I'm not sure I understand this. Can you elaborate?
If it sounds confusing then it's probably not a good choice :)
The "primer" part is just suggesting that the optimization functionality is analogous to the primer of a combustion engine ("a small pump for pumping fuel to prime an internal combustion engine, especially an aircraft"). So the idea is just that the project is intended to "prime" a data workflow for optimal dask execution.
Did you have certain ideas in mind?
My default expectation was dask-dataframe
I don't like this name because it makes this project specific to dataframes (it is today, but maybe it won't be in the future). I like it though because it's simple.
Probably that choice would push us to think about splitting up this repo into parts in some expr-array future, which has its own challenges.
I don't like this name because it makes this project specific to dataframes (it is today, but maybe it won't be in the future). I like it though because it's simple.
I also worry that it would be a bit confusing for the name to be the same as the existing dask.dataframe
module. It would become more difficult to distinguish the two projects in a conversation. I know dask.dataframe
will likely become synonymous with dask-expr in the future, but it will take time for that to happen.
See https://github.com/dask-contrib/dask-expr/issues/374 , maybe that's really the conversation to be having then.
maybe that's really the conversation to be having then.
Sounds good. These two issues seem tightly entangled. I agree that the answer here depends a lot on https://github.com/dask-contrib/dask-expr/issues/374. It also depends a lot on the possible plan for array support (is there an issue for that).
If this repository is going to go semi-public then it should probably have a more user-focused name.
expr
corresponds to an internal detail. It's useful for us but doesn't mean anything to a user.Any thoughts?