Re: Proposal for a comments system

[Shaun McCance <shaunm gnome org>]

On Tue, 2010-04-06 at 18:53 +0530, Anirudh wrote:
> Hi,
> 
> What is the exact motivation for using a comments system? I can see it
> serve two purposes - one to point out any errors, inconsistencies or
> ambiguous documentation, and the other to ask questions. If it's the
> latter, you'll end up having a lot of people asking questions from the
> related to unrelated \- as sometimes people are desperate to reach out
> and find some help.
> 
> On an article about Desktop resolution settings, someone might post
> "Help! My sound doesn't work" and get disappointed as nobody looks or
> is willing to help.

No doubt about it. People email me all sorts of support requests
right now, for every random application under the sun, because
my name is in the About dialog of the help viewer.

The unfortunate fact is that we're going to have to mark some
comments as invalid, because we don't have the wherewithal to
deal with random (and often not-GNOME) support requests. But
we will get good, valid user feedback as well.

>  Also if you are interested in deploying a comment
> system purely for online usage, you might consider using Disqus[1] or
> IntenseDebate[2]. These work really well for preventing spam but call
> non-free systems on the backend, which might be a problem. These also
> prevent the need to rewrite the library server code to work with
> threaded comments, etc.

There is absolutely no way we could ever deploy a proprietary
system. I'm open to using an existing open-source solution,
but I don't think there are any that provide the experience
we want.

> I've written a program called Splatter [3] that retrieves comments and
> bugs from bugzilla and caches them locally. What I learned while doing
> that is it's pretty hard to keep a server resource and local resource
> in sync, both for the server and the client (then again I am pretty
> naive and inexperienced), but writing a desktop client to work with a
> cloud service and download _only_ the updates since last time, and to
> cache all the existing updates can lead to all sorts of problems.

For this specific case, I don't think it's a very hard problem.
We don't have to sync, per se. We only have to post and then
pull updates. Syncing only becomes tricky when local clients
are able to modify existing data.

--
Shaun