Drop Gwibber from default install
ken.vandine at canonical.com
Mon Mar 12 00:04:52 UTC 2012
On Sun, 2012-03-11 at 17:21 -0400, Evan Huus wrote:
> On Sat, Mar 10, 2012 at 3:01 PM, Kai Mast <kai.mast at freakybytes.org>
> Hey Guys,
> I would like to open a discussion about removing Gwibber from
> default install. Gwibber has several defects that make it way
> worse to
> use than the Twitter-Website.
> - No support for Twitters Streaming API makes it a lot slower
> than the
> NOTE: This feature is nearly 2 years old!!
> - No autocompletion of usernames makes it much more difficult
> communicate with Gwibber
> Also some other stuff that is really annoying:
> - No smooth scrolling
> - No multi-column view
> Is there anybody actually using Gwibber on a daily basic? I
> have it
> uninstalled on all my machines and use Twitter through the
> I would love to use a native desktop-application but Gwibber
> is just
> missing soo many features.
> What do you think?
> Kai Mast
> On Sun, Mar 11, 2012 at 4:37 PM, Bedwell, Jordon
> <jordon at envygeeks.com> wrote:
> On Sun, Mar 11, 2012 at 2:35 PM, Rodney Dawes
> <rodney.dawes at canonical.com> wrote:
> >> > Even if I use the Twitter web site to interact with it,
> and gwibber is
> >> > "slow" to update, it is useful to see the notifications
> which pop up
> >> > due to gwibber running on my system (though I have a bug
> with where
> >> > the notifications are popping up, in notify-osd).
> >> Because you can't use Chrome and a small little addon to
> send updates
> >> from twitter.com to notify, or hell, better, just use
> built-in desktop
> >> notifications in Chrome. I'm sure you could even build an
> addon that
> >> does that for Firefox if one doesn't exist. So your point
> is moot.
> > Actually, no I can't use an add-on for Chrome, as I don't
> use chrome.
> > And even if there was one for Firefox, I wouldn't want
> Firefox running
> > all the time to do that. If I did, I could also just leave a
> tab open
> > to twitter.com, now couldn't I. But while Firefox is sitting
> > using constant CPU and 500M of resident (real) memory,
> > is only using CPU when it needs to, and only 31M resident.
> Some people
> > do value their battery life. My point isn't moot. You just
> disagree with
> > it. There is a difference.
> Because leaving a tab open to twitter.com clearly coordinates
> your reason not to remove Gwibber right? Your point is moot,
> I came up with a valid solution, I suggest you look up what a
> point is and you made it more moot with your latest statement
> just leaving a tab open without popups/notifications of any
> Your memory usage statistics are pretty bad too, looking at a
> tab of
> twitter that's been open the last couple of hours it's using,
> oh wow,
> a whole 86MB... oh wait it fluctuated to 102mb and then
> dropped down a
> bit. And memory usage really drains the battery more then CPU
> time, or
> that memory is the only factor (if any major factor in modern
> in batteries. Remember users have to open up Gwibber too, to
> read the
> messages back again if they weren't there, so tack on another
> 86MB of
> memory each time you open it, the CPU time it takes to open
> it and
> the battery killing memory it uses too. What makes it so much
> than an app tab in Firefox which can be used for other
> purposes too?
> >> > If you haven't got a reasonable replacement, then your
> only goal is
> >> > to punish others by removing something they might use,
> because you
> >> > have some frustrations with it yourself.
> >> Or is their goal to save them the agony?
> > Oh, despair. Rage rage, against the dying of the light. Do
> you have
> > a valid point to make, actually related to the thread, or
> are you
> > just trying to troll Canonical? If you want to reply again,
> please try
> > to get back on topic with it. Do you have a bug or
> scientific research
> > which suggests gwibber causes actual physical pain (agony)
> to users of
> > Ubuntu, simply by existing on the default install?
> I don't need to troll Canonical, you lot do just fine at that
> yourselves. But I'll leave it to a Canonical employee to
> assume that
> everybody who disagrees must be a troll and therefore is one
> default because they are blunt and don't sprinkle love and
> care all
> over their statements as if it's a new born baby. By
> standards though, the pain and agony it causes me every time I
> try to
> use it is scientific enough. Add the other 2 users in this
> thread who
> do not like it and we have ourselves a bona fide study.
> Remember, the
> Canonical usability test only consisted of 15 people (that
> sir, is a
> huge and I mean HUGE sample group) and that was good enough
> for ya'll
> this right here is a perfect study :).
> I think it would benefit everyone to take a step back and calm down
> please :)
> What we have now is a problem with two possible solutions. The
> problem, as nicely articulated in the original parent's message, is
> that Gwibber is slower and less feature-full than the Twitter web
> site. The original parent provided several verifiable examples, so I
> don't think anyone can agree with the problem statement, even if you
> might disagree with its severity.
> Solution One: Remove Gwibber from the default install.
> Solution Two: Improve Gwibber.
> If the Twitter website is sufficiently fast and feature-full that
> Gwibber can never provide sufficient advantage to outweigh it's cost
> in maintenance and CD space, then the correct solution is to remove
> it. If it is believed that it can, in future, become more useful, then
> the correct solution is to improve it. Obviously neither of these are
> going to happen for Precise, as we are well past FeatureFreeze at this
> I have CCed Ken on this, since he is one of the main Gwibber devs and
> is probably best equipped to answer questions on what sort of effort
> would be required to improve it.
Thanks Evan for forwarding this, I apologize I haven't had time to read
the full thread yet. Let me try to add a few thoughts here and
hopefully get the ball rolling towards a solution.
First thing is to remember the reasoning behind including gwibber in
Ubuntu wasn't to provide a twitter client, it was for social networking
integration in the desktop. Provide an easy way to share information
with your friends and see what your friends are up to. This doesn't
mean the gwibber client user interface providing you a stream of your
friends activity, but real application integration. Like a simple way
to post to all your accounts from the former MeMenu or to share user
reviews in software center. Software Center is a great example of why
Gwibber was included, after you write a review you can choose to share
your review from right there inside the software center interface. Sure
applications can implement the functionality themselves, but that would
require quite a bit of effort. To use libgwibber to post, it really
only takes a few lines of code. And this type of desktop integration
really can't be done with the browser.
I would really like to see Gwibber get "fixed" as opposed to dropping
it. Not because I am now the maintainer, but to continue with the
original spirit of why it was included. If nobody cared to fix Gwibber
I would have no problem spending my time working on other projects, I
took on Gwibber because the previous maintainer stepped down and I had a
vested interest because of Ubuntu.
The real problem we have with Gwibber is lack of contributors, we only
have a few regular contributors all with other responsibilities.
Fortunately I get to do some work on Gwibber as part of my day job
working for Canonical, which is awesome. However, most of the work I
put into gwibber is in my spare time. None of the complaints I've seen
from people are unsolvable, but we need to have a plan and people to
work on it.
My proposed solution: Put together a plan of what needs to be done and
do a call for volunteers to help work on it. I do think we can find
some people with enough interest to do their part. This is much more
constructive than just saying we need to drop it.
For those that haven't tried the latest version, I really suggest
getting 3.3.91 and taking that for a spin. Most of the effort we've put
into it recently have been quality, so no stunning changes but more
reliability. Duplicate detection and handling of the content.
Scrolling and keyboard navigation has improved quite a bit, but we
really need smooth scrolling. I would say that would be at the top of
the todo list.
More information about the Ubuntu-devel-discuss