osdf-devteam team mailing list archive

[Sam Hart <criswellious@xxxxxxxxx>]

On Tue, Dec 21, 2010 at 8:32 AM, zeto28 <amf5bp@xxxxxxxxxxxxxx> wrote:
> On Mon, Dec 20, 2010 at 02:16:56PM -0800, Andrew Ettinger wrote:
>> I was thinking we should install / use a listserver for the other
>> teams, so they don't have to go through LaunchPad...
>
> Finally.  However, I already see a big problem emerging: we are getting
> too many channels with no clear definition of each one's purpose.
> Having to follow both the forum and the lists will become a pain very
> quickly, and we should resolve this as early as possible.

I actually totally agree with this- I even thought of it after I
submitted my suggested on Google Groups (but was too busy to
follow-up).

I mean, let's re-cap what we have right now (and I may not even have
everything covered, so please someone correct me if I'm missing
something):

* Launchpad, which includes:
- This mailing list
- Some issue tracking
- Source-code management
- A wiki
- Other items? (Did someone mention a forum?)
* The Reddit Subgroup
* A couple of websites including
- http://www.theosdf.org/ is the new hotness
- Didn't we have something else at one point as well?

The way *I* see it, we're pretty close to being saturated with methods
of communication.

Launchpad and its features are probably the best for the
developer-side of things- dealing with code, managing technical
details, etc. I saw Stacy say that she didn't find Launchpad useful
for project management with designers, but (and it's not my intention
to be rude) I'd suggest that's probably not a flaw in Launchpad-
Launchpad is used in over 6,000 projects[1] including and most notably
Ubuntu http://www.ubuntu.com/ . Chances are if we can't find what we
need for project management in Launchpad, we're probably just using it
wrong.

> The only things that speak for a web forum are that less tech-savvy
> people can use it more easily and that moderators can move, split and
> merge threads.  However, it is easy to DDOS, provides only very basic
> search functions (no regexps, rate-limited), is insecure (SMF stores
> passwords in the clear AND mails them around!), and lacks tree threads
> and message-level read management.  Mailing lists OTOH are fully
> decentralized, searchable on the client-side (reducing server load),
> support tree threads, offer far better read management (i.e. one has to
> only hit a key to read the next unread message), and official posts can
> be cryptographically signed so everyone knows they're genuine.  There
> will be some reasons that the Linux kernel developers don't use a forum
> as the primary means of communication.  I for one have developed kind of
> an aversion to web forums.

I agree there probably *should* be some sort of forum/mailing-list
barrier between those averse to forums and those averse to mailing
lists. I am certainly in the "I hate forums" camp (because I'm
definitely from the hardcore, low-level Free-Software development
camp). But I will admit that for some general person who *just* wants
to communicate with the rest of the team (and not really contribute in
any other way other than make a post) a forum is the way to go.

The real question is, what do we get from additional forum software
that we don't already get from the rpac subreddit? E.g., do we
*really* need another bit of technology thrown into an already crowded
pool?

> If MajorDomo is too arcane, GNU Mailman would be the next best
> alternative, I think.  There also exists a software called "Syncom"
> whose purpose it is to sync Mailman mailing lists and a phpBB forum
> using a news server.
<snip>
> Google Groups would probably be technically an option, but only for
> lists where the public can't post
> (<http://ejohn.org/blog/google-groups-is-dead>).  Using Google
> infrastructure would make me feel uncomfortable, though.  We should only
> consider that as a last resort option, I think.

See, the extra complication you detailed above was exactly why I
originally suggested Google Groups as it would come pre-built with
those integrated features (and we all tend to be busy- I'd hate us to
commit to some complicated integration that none of us have the time
to complete).

But I'll admit for the same reasons I was uncomfortable automatically
recommending HSBC here
http://www.reddit.com/r/rpac/comments/elqvi/what_bank_should_we_use/c191ztu
, I'm also hesitant recommending Google.

[1] https://bugs.launchpad.net/

                                      ---Sam