Incubation, healthy communities, and the Apache brand (long)

[Cliff Schmidt <cl...@gmail.com>]

Here are my $0.02 on the reasons for the incubation process:

Thanks to those who came long before me (particularly the httpd
folks), the Apache brand is incredibly strong.  There are literally
millions of users of Apache products who believe that they are making
a good choice by simply building on something from Apache.  This brand
adds value to all ASF projects, which allows them to get increased
exposure and hopefully continue to attract developers to do more great
work here.  However, as the number of projects at Apache grows, we are
faced with a very difficult challenge of keeping that brand strong.

Many of the users who decide to build their businesses on Apache
products are doing so with the belief that the product will "be there
tomorrow", which means that questions will be answered on the mailing
lists, bugs will eventually get fixed, and the product will release
future versions to keep up with the latest user requests (e.g.
interoperability with the latest standard).  This is why community is
more important than code to Apache -- without a strong and diverse
community we can't live up to these expectations.  If development is
driven solely by one person/vendor, Apache is taking a risk that
personal/business changes won't cause all development to come to a
halt.

IMO, the purpose of incubation is to allow new projects to get
familiar with the way things work at Apache, to ensure any legal
issues are adequately addressed, and to build a diverse user and
developer community that minimizes the brand risk of endorsing the
project under the Apache name, which happens only after graduation.

Over the last couple years, the Incubator community has tried to
quantify the standard for an acceptable community.  We have the 3+
committers rule (which really should be from different organizations
-- a major reason for this rule is to allow a veto to prevent a single
person/company from making a self-serving decision that isn't in the
best interests of the overall community); we used to have a rule a
year ago that no more than 50% of the committers could be from any one
organization (which was both unnecessary and could be gamed by a
single vendor if it was the only hard requirement).

My personal belief (and as a single vote on the Incubator PMC) is that
a project is ready for graduation when it appears to be an acceptable
risk to the Apache brand.  Ideally, this means that no single
individual/company could irrevocably damage the project by walking
away; however, if that's not quite the case but the project has grown
and diversified its community and has demonstrated an open process
that continues to invite others to participate, then it might be an
acceptable risk.

Cliff

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org