Re: proposal: drop bugzilla@jessica.w3.org emails from public-html human lists

When we first set up the bug notification, we wanted to watch how it works out and gather feedback. I'd be interested in hearing more input on how it's working out.

My personal perspectives:

A. I think removing the notifications on keyword change has been helpful - those were 99.9% noise.

B. I think adding notifications for all new bugs has on net added value. More WG members are aware of incoming issues. When I follow the links in bug notification emails (which I often do), I nowadays often see that there are already multiple useful comments in relatively new bugs. I expect even more people are reading bugs of interest than just the people commenting. I think this has had two significant benefits:

   1. WG members are much more aware of what is going on with the group. Empirically, even the most active and engaged WG members were not able to follow what was going on nearly as well, solely using bugzilla and opt-in tools.

   2. Initial dispositions of bugs by editors tend to be higher quality, as they benefit from additional information provided by comments.

C. We do have a bit of an ongoing spam/junk bug problem, largely via the comment forms on both the WHATWG and HTML WG drafts of the spec. These junk bugs have always existed, but broadcasting them increases their annoyance value. If we can find a sufficiently lightweight means of moderation, that might help.


On some level, there's a mechanical equivalence between sending bug notifications to the list, and asking those who wish to skip them to opt out via filtering, and having a separate opt-in mechanism to get them. However, the default also has a big impact on creating social norms. Nowadays, when you see a bug and comment on it, you can assume that interested parties have likely seen the email notification. So there is less need to drag discussions into the mailing list, and less grounds to worry that key people will entirely miss a particular bug. That's generally a safe assumption even though some people may opt out via mail filtering. But with an opt-in system, even one that a considerable number of people used, you couldn't make these assumptions.

After seeing how it has played out in practice, I think new bug notification is the right default. An opt-out system seems to be more effective for this group than opt-in, and we get higher-quality discussion and work product, with less noise.

Still, I'd be glad to hear feedback from others on how things are working for them.

Regards,
Maciej


On Nov 15, 2010, at 1:22 PM, Tantek Çelik wrote:

> HTML WG:
> 
> I propose that we drop the cc'ing from bugzilla@jessica.w3.org
> to public-html lists that are used by/for humans for discussion.
> 
> At a minimum: public-html@w3.org and public-html-a11y@w3.org.
> 
> I've occasionally looked at the bugzilla@jessica emails and found that
> they're basically noise.
> 
> In the interests of providing less noise/pain to people on the lists
> in general, let's stop cc'ing the lists with bugzilla@jessica.w3.org
> and any other robot emails.
> 
> Anyone that wants to keep up with specific bug edits (the special
> case) can simply use the bugzilla interface directly and be notified
> of changes to particular bugs that they're interested in.  No need to
> burden everyone with this.
> 
> If there are a non-trivial number of people that prefer to get robot
> emails, perhaps we can setup public-html-robots for such automatic
> notifications.
> 
> Thanks for your consideration,
> 
> Tantek
> 
> -- 
> http://tantek.com/ - I made an HTML5 tutorial! http://tantek.com/html5
>

Received on Monday, 15 November 2010 23:12:21 UTC