CFV: New Project: CRaC

Anton Kozlov akozlov at
Fri Aug 6 19:36:00 UTC 2021

On 8/6/21 2:53 PM, Jesper Wilhelmsson wrote:
> Obviously the discussion will continue after the project has been created and go on as long as the project is active, but the initial discussion that happens before the project is created has a different purpose than the discussion that follows the creation. It's a matter of determining what the actual goal is, and it's also a matter of making sure everyone who is expected to participate in the project later feels welcome and included already from start. Adding Committers later is obviously possible but there's a process around it that requires the potential Committer to prove her/himself worthy before being accepted into the group. This can be a quite large mental blocker for some people. Being included from the start will encourage people to be more active and devoted to the project.

The committer nomination process provides only a rough guide for the
prerequisites.  I think it won't be any problem for an experienced OpenJDK
developer to become a committer.

Other prototype implementations can be counted as sufficient significant

There is no aim to introduce barriers for contributors.  If someone wants to be
included in committers, I ask to send a message.  So far, the initial set
includes everyone who explicitly expressed the desire.  Contributors are

> Looking at the current project description and the initial set of Committers/Reviewers I can't help but wondering what the purpose of this project is. Is the goal to investigate a generic problem area, or is the goal to get your prototype productized and pushed? Both are perfectly valid goals of a project, but the project description should clearly state what the goal is. The interest from others to join and contribute to the project depends very much on which one of these two goals is the actual goal. How many different approaches are you prepared to investigate within this project? Are you prepared to throw your prototype away if some other team enters the project with a better suited solution? There are already several prototypes in this area trying different approaches but the project description clearly states that this project is about your prototype.
> So far I haven't sen enough evidence that this project is about the generic research of a problem space. If I am to ask my friends at the university to look at this, I first want to make sure it's a research project, not just a finish someone else's implementation project.

The purpose of the project is to research the area.  The CRaC prototype is
rather small, so enhancing may mean rewriting it.  All approaches should be
considered.  I think the first step is so obvious that all prototypes are
similar and differ by names and code organization only :) If there is a better
prototype that covers what is done in CRaC and the another prototype does not
have drawbacks, then it's should be logical to take the better one.  But if two
prototypes have different properties, there should be a way to combine their
strong sides.

The CRaC prototype is only the starting point.  It will be great to have
different ones e.g. in separate branches right at the start.  Are there links
to the other prototypes?


More information about the discuss mailing list