RE: bug fix process question

Ginny, it's not that there is a specific example where the end result was 
undesireable.  As I said, the question was asked, it seems a valid 
question, so I simply realized we may not all be on the same page.  One 
key to a smoothly functioning group is a common understanding of work 
processes, and process maintanance comes along with the chair role. 
I remember contributing to some chaos myself there back in August by 
opening some as editorial (in that case, without prior group discussion); 
even though I trusted my judgement enough to take that "risk", it turned 
out in at least one case that what I believed to be an editorial change 
someone else found to be substantive.  Maybe I'm risk-averse, but that 
experience suggested (strongly) to me that I was better off letting it run 
the normal course so everyone had a full and obvious ("in person") 
opportunity to review it.  We know some members are better able to keep up 
with all the bugzilla emails than others; it's a good thing for consensus 
if no one feels left behind.
So if I'm hearing you right, there are a couple of cases:
We discuss the bug as a group in advance of it being opened, as we did 
last week.  When we do this, we will need to be careful to clearly do an 
explicit consensus check if we plan to feed something directly to the 
editors.  As I couched the Tuesday discussion, several times, I do 
remember saying that its purpose was _only_ to figure out whether or not 
to decide whether a bug was needed, so it would be perfectly reasonable 
(in the absence of an explicit suspension of that assumption) for someone 
to assume we would triage the bugs as a group after they were opened.
If an explicit consensus check is done, then we can open the bug later and 
treat it as already triaged.  Doing so puts slightly more onus on whomever 
is not present to review bugzilla traffic carefully.
In the absence of an explicit consensus check, it is no different than the 
garden variety bug. 
Garden variety bug
If the submittor marks it editorial, 
the editors are free to work on it.
when editorial work is done, 
If the submittor commented that it needs later review
it is marked needsReview
Else
it is closed as fixed (question below targets this path)
Otherwise it gets triaged at the next meeting where bug triage is on the 
agenda
Triage serves as a point of "conscious review" i.e. more than a simple 
email miss is required for it to get by one's attention.

The question I would ask on the cited path is: are we as a working 
comfortable with the existence of this "easy to miss" state?   We have at 
least these alternatives:
1. yes, comfortable with that as is.  other safety nets exist and can be 
used.
2. no, we should force those into needsReview so the full working group 
does a formal consensus check
Note that, even if we took the second approach, it is still possible for 
someone to miss the consensus check/discussion.  In the end, the ability 
of anyone to re-open an old, or open a new, bug on the final content is 
the real safety net.  So it could well and fairly be argued that the first 
is a better trade-off on process weight vs safety net.
I'd suggest folks think it over, and we'll do a quick check tomorrow 
before triage.  If no other options emerge, we decide then and there.  If 
other options do emerge that people wish to consider, we make a decision 
amongst these two tomorrow on how to operate until people have considered 
all options, and supersede tomorrow's decision later.

Best Regards, John

Street address: 2455 South Road, P328 Poughkeepsie, NY USA 12601
Voice: 1+845-435-9470      Fax: 1+845-432-9787



"Smith, Virginia (HP Software)" <virginia.smith@hp.com> 
Sent by: member-sml-request@w3.org
01/29/2008 05:46 PM

To
"member-sml@w3.org" <member-sml@w3.org>
cc

Subject
RE: bug fix process question






John, do you have some specific bugs in mind? I looked through the bugs 
and didn’t see anything untoward. It is my recollection that when we 
reviewed some my spec review issues, we reached consensus on some of the 
minor ones (essentially we triaged before the bug was submitted). 
Therefore,  I think I probably entered some of these ‘spec review’ bugs 
and immediately marked the ones we reached consensus on as ‘editorial’ so 
the changes could be made - and then I normally mark it as ‘needsReview’ 
to bring it back to the attention of the group.
 
So, yes, you are correct about the process but it was my recollection that 
some issues had consensus at the f2f and, therefore, did not need a 
separate triage to put them into ‘editorial’ status. Our process does have 
provision for going directly to ‘editorial’ at the discretion of the 
submitter. 
 
That said, I did not see any bugs for which this would be a problem. 
Please let me know what specific bugs you are concerned about. 
 
I did notice that Kumar marked bug 5384 as fixed but ‘needsReview’. That 
looks like an oversight.
 
--
ginny
 
From: public-sml-request@w3.org [mailto:public-sml-request@w3.org] On 
Behalf Of John Arwe
Sent: Tuesday, January 29, 2008 6:55 AM
To: public-sml@w3.org
Subject: bug fix process question
 

Editors, I was under the impression that things entered the "editorial" 
work queue when the working group triaged the bugs and agreed to mark them 
editorial (implictly meaning that we have some level of consensus on the 
resolution, and allowing anyone on the edge to assert the need for a 
post-editorial wg needsReview stage).  I seem to remember a discussion a 
month or so ago re-affirming this.  Am I remembering incorrectly? 
Someone pointed out to me that some bugs opened last week, but as yet not 
triaged, are getting editorial attention, and this confused them. 
There is a tradeoff between process confidence (transparency, enforcement 
i.e. the agreed-to process is being followed) and throughput, I'll readily 
acknowledge.  If the working group as a whole is now comfortable allowing 
editors to use their judgement to fix bugs in advance of triage by the 
working group, I believe that is a decision within our purview... I'm just 
not sure we've had that discussion.  Alternatively, and I think pretty 
safely, in cases where the editors believe a bug is purely editorial they 
could append the suggested revision (logically, make a proposal) in 
advance of triage.  There might well be other viable alternatives. 

Best Regards, John

Street address: 2455 South Road, P328 Poughkeepsie, NY USA 12601
Voice: 1+845-435-9470      Fax: 1+845-432-9787