Merge pull request #51 from protocol/add-problems

Add problems.md and modify issue template

Author: Mike Goelzer

.github/ISSUE_TEMPLATE/project-proposal.md

@@ -16,5 +16,6 @@ detail and PR discussion(s).
 Project name: [insert name]
 Proposal: [link to markdown file in this repo]
 Initial PR: [reference PR number with the first proposal draft]
+Problem Addressed: [reference a problem from `/problems.md` in this repo]
 
 This issue is for project management of a proposal, deciding, scheduling etc. Discussion about the relative merits of the proposal is welcome here, but discussion about the content of the proposal is better in a PR to the proposal document.

problems.md

@@ -0,0 +1,57 @@
+# Top Problems from Product Team
+
+## P0
+
+ - The fact that storing and retrieving data with the stack (IPFS+Filecoin) can ever fail is a problem.
+
+## Stack Ranked P1
+
+ - We currently do not always understand why storage deals (in the wild) fail so frequently.
+
+ - Data that has been stored successfully is not guaranteed to be retrievable.
+
+ - Filecoin does not support small file storage well, even though small files are the most common use case.
+
+ - IPFS users can’t pin their data to Filecoin.
+
+ - New developers have to spend too much time and money to go through a simple storage/retrieval workflow.
+
+ - Clients find it extraordinarily difficult to discover reliable miners who will meet their storage needs (SLA, geography, miner cost quotes).
+
+## Stack Ranked "Probably P1" (but need investigation)
+
+ - Clients in certain cases can lose data that they have stored to the Filecoin network.
+
+ - Clients want to store redundant copies of their data, but cannot do so without paying for multiple data transfers.
+
+ - Clients today don’t know how much Filecoin storage will actually cost them.
+
+ - Clients cannot renew DataCap, making Filecoin Plus significantly less useful.
+
+ - Garbage collection in IPFS blocks normal node operations.
+
+## Stack Ranked P2s
+
+ - Clients with massive datasets do not know what best practices to follow when storing those datasets on the stack.
+
+ - Users don’t have a "standard" cloud configuration to deploy without manual configuration of Lotus.
+
+ - Highest-value IPFS users (pinning services) do not advertise their pins publicly
+
+ - The total cost of integration+storage on the stack is higher than using AWS
+
+ - Clients that want to extend the amount of time their data is stored on Filecoin cannot do so in a reasonable way.
+
+## Stack Ranked P3s
+
+ - Developers cannot rely on our stack to provide stable APIs (eg, semver). Instead, they have to contend with unexpected breaking changes.
+
+ - A user with millions of pins should be able to store them on the stack
+
+ - Developers cannot build complete browser-based and desktop dapps with our JS libraries.
+
+ - Developers cannot find reliable information about our stack when doing Google searches.
+
+ - Filecoin is currently unsuitable as the default storage solution for NFTs.
+
+