The npm Registry recently got crashy, and it was incredibly painful.
The headline on scalenpm.org sounds terrible. It is terrible when your packaging solution goes offline and it’s difficult or impossible to install new pieces of software. When anything goes off line1 it’s a painful experience to deal with. We should work as hard as possible to make distributed systems that can resist failing infrastructure. This means backups, mirroring, and adding fallback systems to the core product.
Building a distributed system is hard. I don’t mean to trivialize it with so few sentences. I think the flawed efforts of nodejitsu are nobel. What npm needs isn’t a corporate backer, it needs a non-profit transparent foundation.
Why do they need \(200,000? How much does npm cost to actually run in terms of time and hardware? Nodejitsu has taken a **\)2,650,0000 in funding.2 They are a for-profit company. They have given zero** transparency other than some fancy looking numbers.
The Python Software Foundation is a 501(c)(3) corporation that has sponsor members. These sponsors allow it to run pypi, the Python package index. This is a fully transparent operation. Why isn’t npm run the same way?