Open psmulovics opened 1 year ago
Tagged @victorjunlu based on discussions yesterday
Cost of Forking:
1 FORK = E10,000 - 100,000 TCO (@tsteenbe) Unwitting creation of multiple forks of the same project.
Compliance,
Duplication of Work
Lost productivity (releases, merging)
CI/CD for the fork
Sometimes it's impossible to avoid when there is a license issue.
We need to get more insight into SW development costs. Licensing / Compliance are known. Engineering perspective - 20-30% reduction In cloud build costs.
(PH) - Is there a misunderstanding that there is proprietary business logic in these forks, do management believe that's why they can't open source it? e.g.
(RM) - I'd not thought of this before because it's a huge violation of architectural principles.
A lot of managers communicate the need for forks to be proprietary as a moat to protect their workspace. If they could work with upstream directly, this would mean that staff are more fungible. Open Source is used as a defensive shield. Forks protect the engineering budget. This is like bad development Io create job security.
Internal Forks are waste (in the lean language)
However, CxOs are interested in employee mobility. Using proper open source packages aids mobility.
The goal should be to upstream form.
For every fork, you should know where it goes: which projects it ends up in. etc. Otherwise it's unmanaged. Is it being built? is it used in a project?
CI/CD could be per-team, or for the whole organisation. Artifactory or Nexus should be able to tell you about this. A forked binary of Apache Commons. It could be hard to detect these at scale. Tools are bad here.
first step - get an organisational github org.
Gartner route
PS - the only internal alternative to forking is to write your own version
@mimiflynn , let's discuss whether we can put some numbers on this
Making a model to trickle down the executive desire to execution.