Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 169391

Summary: Allow change to committer username
Product: Community Reporter: Scott Lewis <slewis>
Component: CommitterToolsAssignee: Eclipse Webmaster <webmaster>
Status: RESOLVED WONTFIX QA Contact:
Severity: major    
Priority: P3 CC: bjorn.freeman-benson, remy.suen, sharon.corbett, twagner, ward.cunningham
Version: unspecified   
Target Milestone: ---   
Hardware: All   
OS: All   
Whiteboard:

Description Scott Lewis CLA 2007-01-02 20:15:43 EST
An existing committer (rsuen) would like to change username to (rcjsuen).  Some facility should be present to allow this.
Comment 1 Denis Roy CLA 2007-01-02 22:07:04 EST
(In reply to comment #0)
> Some
> facility should be present to allow this.

I'm not sure I understand why this is needed.
Comment 2 Scott Lewis CLA 2007-01-02 22:31:07 EST
(In reply to comment #1)
> (In reply to comment #0)
> > Some
> > facility should be present to allow this.
> 
> I'm not sure I understand why this is needed.

In this case, Remy's current committer account (rsuen) is inconsistent with a number of other Eclipse-related accounts...e.g. IRC account, etc (rcjsuen).  He would like them to be consistent.


Comment 3 Remy Suen CLA 2007-01-03 08:19:18 EST
If such a facility is created, it is probably in the best interests of everyone to enforce a limit on how often you can change it, I don't know, a couple of months or maybe once a year? *shrugs*

When I was first accepted as a committer I had told Wayne that I'd like my username to be rcjsuen. I'm guessing either a) he forgot it, or b) he ignored it because of the first initial/surname policy, or c) the webmaster chose to ignore it because of the policy.

There are of course cases where the policy is not used, but I'm guessing that might be the case because back then there was no policy?
Comment 4 Remy Suen CLA 2007-01-03 10:48:38 EST
I'd like to note that I was not aware that there was actually a policy in place when I noted it to Wayne. So if it can't be helped or the webmasters perceive this to be abused, we can just close this.
Comment 5 Denis Roy CLA 2007-01-05 10:06:36 EST
While I agree that a committer's ID is not always the one they would prefer, our IT policy is to follow the convention of first letter of given name + last name, up to 12 characters. Some of the old-school committers have IDs that don't meet this convention, but they were created before the Foundation's existence.

The purpose of this convention is to offer a faster response to the community; with over 800 committers sending us numerous requests for passwords, permissions, etc, it's easier for us to determine the committer ID with a fixed naming convention rather than having to do a lookup every time.  From what I can tell, most applications that require you to supply an ID allow you to save the ID, hence not requesting it for each connection.

The Committer ID is also tied to many things: committer records, CVS commit history, IP trail, etc. so to change it is not trivial.

I'd like to recommend we close this as WONTFIX. Isn't the Eclipse Committer ID the coolest of them all?
Comment 6 Scott Lewis CLA 2007-01-05 11:06:31 EST
Hi Denis,

Thanks for the reply. 

(In reply to comment #5)
> While I agree that a committer's ID is not always the one they would prefer,
> our IT policy is to follow the convention of first letter of given name + last
> name, up to 12 characters. Some of the old-school committers have IDs that
> don't meet this convention, but they were created before the Foundation's
> existence.
> 
> The purpose of this convention is to offer a faster response to the community;
> with over 800 committers sending us numerous requests for passwords,
> permissions, etc, it's easier for us to determine the committer ID with a fixed
> naming convention rather than having to do a lookup every time.  From what I
> can tell, most applications that require you to supply an ID allow you to save
> the ID, hence not requesting it for each connection.
> 
> The Committer ID is also tied to many things: committer records, CVS commit
> history, IP trail, etc. so to change it is not trivial.

Yes, that's understood and appreciated.  Do you think that some changing of IDs is doable with current staff without running you and others ragged?  That is, would it be possible to do 5 a year, 10 a year or something on that order?  Perhaps it could then be limited to a few 'extreme' cases a year (my guess is that it wouldn't be that many folks anyway).

> 
> I'd like to recommend we close this as WONTFIX. Isn't the Eclipse Committer ID
> the coolest of them all?
> 

Indeed it is :)...at least to those of us Independents :).

Before closing this bug, could we get some idea from you/IT staff whether doing some of these would be possible?  If so, how many would be reasonable?  If staff were added to do this and/or other things reprioritized how would that change the situation (if at all)?

Thanksinadvance.
Comment 7 Denis Roy CLA 2007-01-05 11:58:05 EST
> Yes, that's understood and appreciated.  Do you think that some changing of IDs
> is doable with current staff without running you and others ragged?

Well, yes, it's not like changing a few IDs will automatically run us in the ground, but it does mean we'll be investing time in performing Name -> id lookups. It's not like we can afford to, as the community expects us to be responsive  :)

There are two instances that I know of where a committer ID has changed: 

- a committer's name had changed, so we changed the ID to match the new name. This was at the committer's request, it also maintained our naming convention, so this is absolutely warranted.

- a committer's ID, according to our standard, translated to an obscene word in the German language. As an example, Frank Uckerman's ID would be an obscenity that would certainly warrant a non-standard ID.

My only constraint to ID change requests is that they be requested only when necessary. It appears to me that Remy's ID change request is not stemmed from strict necessity, but rather convenience.  We do appreciate this need for convenience, but, for the benefit of the community, a good naming convention is also convenient for Foundation IT  :)
Comment 8 Denis Roy CLA 2007-02-06 16:08:01 EST
Scott, do we agree to disagree with Comment 7 and close this as "worksforme", or should we investigate further?
Comment 9 Scott Lewis CLA 2007-02-06 16:31:15 EST
Hi Denis,

(In reply to comment #8)
> Scott, do we agree to disagree with Comment 7 and close this as "worksforme",
> or should we investigate further?
> 

I'd like to understand how much resources you would expect something like this to take, if, say 20 of such requests (necessity by your definition in #7) were to come in yearly.  Any additional resources needed to address that mag of load?



Comment 10 Denis Roy CLA 2007-02-06 16:47:27 EST
Adding Bjorn and Ward to this for comment, and Sharon, as she maintains all the legal and paperwork when a committer's ID changes.

Another thing to consider is historical data, such as that on the Dash project:
http://dash.eclipse.org/dash/commits/web-app/
Comment 11 Denis Roy CLA 2007-02-06 16:51:06 EST
One possible compromise would be to allow new committers to choose their ID before we provision their account.
Comment 12 Scott Lewis CLA 2007-02-06 16:54:30 EST
(In reply to comment #11)
> One possible compromise would be to allow new committers to choose their ID
> before we provision their account.

Agreed.  I thought this was already the case, but I haven't been provisioned recently.

Comment 14 Ward Cunningham CLA 2007-02-07 01:02:28 EST
I appreciate the accomodation expressed in Comment #4. 
Comment 15 Scott Lewis CLA 2007-02-26 15:44:14 EST
My suggestion is that we:

1) As per comment #11, we allow committers to choose/review their usernames when they are initially provisioned.  
2) We convey to committers the need/reasons for care/stability in choice of username.
3) Only under extreme circumstances (as judged jointly by committer and EMO and using guidelines outlined in comment #5), the foundation goes through the changes described under comments #10 and #13.

If people don't agree with this approach, then please reopen bug.