Proposal: JS libraries are trim and modular (build a plan of attack) #47
Conversation
|
|
||
| _How much would nailing this project improve our knowledge and ability to execute future projects?_ | ||
|
|
||
| #### Confidence |
There was a problem hiding this comment.
As a data point to increase confidence here, Textile did not use our libs, but built their own ipld+libp2p variant.
@Gozala was looking into creating something like this a while ago and ended up transfering the project from Textile to our org.
It now lives at https://github.com/ipfs/js-dag-service
There was a problem hiding this comment.
There was some great work there, this proposal would definitely take those learnings on board.
|
Looks great. We should also take into account some of the proposed work for Gateway and Pin APIs since they open opportunities for building small and pretty dumb clients that do a lot. |
|
I really like this idea! I would love if it would~ could~ make other filesystem abstractions like ones implemented by fission and textile team a first class option for anyone to pick up as opposed to making them a second tier. |
|
@carsonfarmer raised this today in the IPLD meeting and suggested that the DagService is a key part of composability for the work they're doing (although composability across the stack would be nice too!). Being able to extract that piece and/or swap it out for alternatives would be really helpful. There's some prioritisation work that can be achieved here that lets us chew off pieces of it that have impact while not committing to the whole lot. |
Co-authored-by: Oli Evans <oli@tableflip.io>
This builds on various people's recently expressed thoughts, including @achingbrain, @vasco-santos, @mgoelzer, @jacobheun, @mikeal and others. I've seen this same sentiment expressed in different ways but @achingbrain wrote it up fairly nicely yesterday and I've captured some of that here. I'd love for any of you who have additional thoughts on this to hack on my branch here to enhance this proposal if you think it's lacking.
Currently this is a proposal to build a plan of attack, not to do the work. I'm conscious of the need to keep the scope manageable so this is just initial work and would lead to further proposals to do the work but likely in pieces, ranked by impact and dependence.