BTW, the recent discussions about process made me go look at the swiki at what is actually visible, and actually, if you're a new Squeaker browsing through, starting from /1, it is hard to get to the good documentation, which is a pity, because people are probably missing it all the time.
I'll be adding to it and refactoring it soon.
Daniel
Doug Way dway@riskmetrics.com wrote:
Daniel Vainsencher wrote:
I agree generally, but 25% of ones harvests seems like something that is too hard to keep track of. How about simply restricting it to very simple stuff, that is under 1k of code?
This fits with a feeling I have that trivial cleanups should be easy to insert, because they hold little risk, and big stuff should get more reviews because they hold more risk.
That sounds like a good rule.
The 25% figure doesn't have to be a strict rule... it could be more of a guideline. The general idea being that we want to encourage reviewing/approving other people's submissions as much as possible.
- Doug
Doug Way dway@riskmetrics.com wrote:
Daniel Vainsencher wrote:
I know why isSymbol is needed, and don't personally mind your self approving it. ...
I've been thinking that the rigid "you can't approve your own submission" rule is probably too strict. When we started up the harvesting group a couple of years ago (with SqC still in charge), harvesters were allowed to harvest their own submissions, as long as they mostly reviewed other people's.
What we could do is allow approving one's own submissions if they're reasonably simple fixes. But we should have some rule of thumb such as: no more than 25% or so of the things you harvest should be your own items, because we don't want the harvesters just working on their own stuff.
This would have the added benefit that if harvesters *really* wants to get their own submissions in, they'll also review a few items from other people, so that they meet the ~25% rule.
So something like this might encourage more harvesting. It won't necesssarily solve all of our current issues, but it's a start...
- Doug
I don't know why Number>>extend: is needed, and probably won't miss it very much for the next 3 years.
And more to the point, those extensions Anthony wrote that are generally useful (asBit, header information printer, around half the OC protocol mentioned), deserve to be presented to Squeakers so they'll get used, instead of becoming more browser-filler.
Daniel
Marcus Denker marcus@ira.uka.de wrote:
On Sun, Jun 22, 2003 at 03:55:44AM +0200, Daniel Vainsencher wrote:
[Additions to various classes...] Note that my problem with these is not that I argue with the usefulness of specific items - I simply don't think that it's appropriate to include them as part of a fix. Why don't you post them one by one to the mailing list for discussion, where people may argue whether they are appropriate. You're saying that something should be part of the class library, defend your opinions on squeak-dev, just like Richard does, method by method, idea by idea.
I just self-approved a two-liner I posted six weeks ago. So I would estimate that this procedure will take maybe 3 years.
Marcus
-- Marcus Denker marcus@ira.uka.de -- Squeak! http://squeak.de
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
squeakfoundation@lists.squeakfoundation.org