> Side note, I've noticed a steady decline in quality of Apache projects for at least the past decade or so.
That's a weird trend to notice, given that none of the same developers are working on "Apache projects" generally. The ASF isn't like e.g. Mozilla; it isn't a monolithic org with its own devs. It's a meta-bureaucracy that takes ownership of projects for legal (copyright + patent assignment) reasons, and then offers project-management charters to member projects, giving them process and structure for contribution without a canonical owner.
An ASF project is sort of like a "working group" (e.g. Khronos, WHATWG), except that it's usually one or two large corporations and a bunch of FOSS devs, rather than being exclusively a bunch of corporations.
---
On the other hand, if there is a trend, it might be because of the increasing reliance on the "open core" strategy for corporate FOSS contributors to make money.
My own experience is with Cloudant, who contribute to Apache CouchDB, but also have their own private fork of CouchDB running on their SaaS. Originally, Cloudant did what they liked in their own fork, and then tried to push the results upstream. The ASF didn't like this very much, since these contributions weren't planned in the open, so Cloudant increasingly tried instead to mirror its internal issues as ASF Bugzilla issues and get public-member sign-off before going forward on private solutions. Which is kind of a shadow play, since many of the founding members of the CouchDB ASF project either have always worked for Cloudant, or have since been hired by Cloudant, so it's Cloudant employees (in their roles as ASF project members) signing off on these Cloudant issues. But it still slows things down.
A good comparison that people might be familiar with, with the same dynamics, is WebKit. The WebKit project has its own separate development organization, but most of the developers happen to also work for Apple.
Previously, WebKit was Apple and Google, but even two corporate contributors were too big for the small pond. Which, to me, shows that they were each there expecting to dominate the decision-making process, rather than find consensus with the FOSS contributors; and having an equally-powerful player that they had to form consensus with was too much for them.
How's that dispositive of a declining average code quality? ASF has influence over which projects get incubated, and even some over how projects are managed.
That's a weird trend to notice, given that none of the same developers are working on "Apache projects" generally. The ASF isn't like e.g. Mozilla; it isn't a monolithic org with its own devs. It's a meta-bureaucracy that takes ownership of projects for legal (copyright + patent assignment) reasons, and then offers project-management charters to member projects, giving them process and structure for contribution without a canonical owner.
An ASF project is sort of like a "working group" (e.g. Khronos, WHATWG), except that it's usually one or two large corporations and a bunch of FOSS devs, rather than being exclusively a bunch of corporations.
---
On the other hand, if there is a trend, it might be because of the increasing reliance on the "open core" strategy for corporate FOSS contributors to make money.
My own experience is with Cloudant, who contribute to Apache CouchDB, but also have their own private fork of CouchDB running on their SaaS. Originally, Cloudant did what they liked in their own fork, and then tried to push the results upstream. The ASF didn't like this very much, since these contributions weren't planned in the open, so Cloudant increasingly tried instead to mirror its internal issues as ASF Bugzilla issues and get public-member sign-off before going forward on private solutions. Which is kind of a shadow play, since many of the founding members of the CouchDB ASF project either have always worked for Cloudant, or have since been hired by Cloudant, so it's Cloudant employees (in their roles as ASF project members) signing off on these Cloudant issues. But it still slows things down.
A good comparison that people might be familiar with, with the same dynamics, is WebKit. The WebKit project has its own separate development organization, but most of the developers happen to also work for Apple.
Previously, WebKit was Apple and Google, but even two corporate contributors were too big for the small pond. Which, to me, shows that they were each there expecting to dominate the decision-making process, rather than find consensus with the FOSS contributors; and having an equally-powerful player that they had to form consensus with was too much for them.