Thunderbird mailing-list / NNTP decision rationale

Andrew Sutherland asutherland at asutherland.org
Tue Aug 10 02:00:14 UTC 2010


  On 08/09/2010 06:10 PM, Bryan Clark wrote:
> This isn't an attitude or a political statement and I don't think 
> anyone is out to remove NNTP.  I'm reading your messages as suggesting 
> that and I don't think any of that is true at all.

I think it's worth elaborating on this a little bit.

No one has a problem with NNTP as a protocol or newsgroups as a 
concept.  There are concerns about the complexity of the Thunderbird 
codebase as a whole and our ability to maintain and otherwise support 
the current C++ NNTP implementation given the current level of resources.

If something happened where our options were between sinking a 
non-trivial amount of MoMo man-hours into NNTP or removing the current 
C++ implementation, I believe we would remove it.  We would do this 
because, as noted, NNTP is not a MoMo priority and we need to focus on 
things that are priorities.  If a contributor who deeply cares about the 
C++ news implementation shows up and has the time to help make sure that 
never happens, we can avoid that eventuality.  The other possibility is 
that NNTP can be migrated to a simpler JS-based implementation, but that 
also requires effort that will not primarily come from within MoMo.

Right now, Joshua Cranmer is the news sub-module owner.  I am under the 
impression that he has a limited amount of time available and so the 
continuity of the current news implementation is not completely 
assured.  If someone else magically appears to help out Joshua, the 
current low probability of removal could be further reduced.

The good news is that one of the things Joshua is spending his time on 
is a way for JS extensions to implement account types which would reduce 
the maintenance burden for news.  The bad news is that his mechanism for 
allowing this currently depends on some scary and complex JS and C++ 
stuff which itself carries its own maintenance concerns.  Back in the 
good news category again, bienvenu's back-end road map eventually 
includes work that supports custom account types that is less scary and 
I think he has some ideas for how Joshua's work could be less scary if 
Joshua has the time to follow-up on those.  In the 'good news to some' 
category, my gloda plans will also allow for custom account types 
whereby one could implement NNTP in JS.  I say 'good news to some' 
because it is not clear that the set of people who most love NNTP 
overlaps with the set of people willing to use gloda and all that it 
entails.

Andrew



More information about the tb-planning mailing list