seanegan at gmail.com
Wed Dec 5 01:26:55 EST 2007
On Dec 4, 2007 8:48 PM, Gary Kramlich <grim at reaperworld.com> wrote:
> As I've expressed before, I read and type all the time, and hey look, my
> lines now move while reading. Just like reading in a car really, and
> not something I should be forced to endure while I am in fact stationary.
That's fair. Of course, you are forced to endure the same exact
phenomenon should your friend send you a message before you finish
reading the previous ones. This happens all the time in chats, and
frequently enough in IMs that surely people are used to this. The
auto-resizing behavior should actually be less annoying since it
behaves in a perfectly predictable way based entirely on user
I'm not saying you have an invalid point; obviously showing new
messages when they arrive is way more important than maximizing the
backscroll. But like showing new messages when they arrive, I think
that the benefits outweigh the cost here. You obviously disagree.
> See elb's comment on this.
I probably agree with him that I take that argument more seriously
than any other, but I still don't find it very convincing. I don't
think running over remote X has been a major goal of ours (or any
modern desktop program, really) for some time, and this feature is,
again, no more expensive than sending or receiving a message (and
occurs less frequently). I think we get more benefit from this than we
hurt the three people running pidgin over remote X (who should
probably be using Finch anyway ;)).
> Do you _really_ want to play the white space card here... Look at my
> screen shot and tell me we're not already wasting a ton of space...
I'm guessing you're referring to the horizontal space being unused by
the infopane and toolbar. I've told people that they only "waste
space" if you give them more space than they require. You have a good
point that when I say the input area is wasting space, it's only
because we give it more space than it requires.
(as I mentioned earlier, I dislike Stu's change because it gives the
input more space than it requires by itself)
And important distinction, I think, is that the amount of space
required by the input area is very dynamic, we can calculate it, and
are able to give it exactly the required height. The infopane and
toolbar are wasting *horizontal* height, which cannot be given to the
backlog anyway. Also, the backscroll, toolbar, and tabs are all
There are some UI features that are potentially worth taking room from
backscroll; presenting status information is one thing I think is
worth it. Saving bandwidth on remote X or saving your eyes from
adjusting themselves 12 points don't meet that bar, I don't think.
> You know what I've realized over the years... Blanket cases are never
> the correct approach. Attempting to provide more context by showing
> more backlog that according to your quote isn't even used, seems like a
> waste of time and effort in even making it available. By your own
> admissions you've pointed out that it's pointless, and I'm in awe at why
> we're still discussing this.
Basically (and this is the case with all these UI wars), we have a
subjective difference of appearance. I can go on talking about
academic research that supports my case, and point out that basically
every IM client since 2003 works this way. You can talk about saving
bandwidth and reading in cars, but neither of us are ever going to
change the other person's mind.
I do try to be reasonable about this stuff, and I'm sorry when I'm
not. My crazy tab ideas proved quite unpopular; so we reached a
compromise people seem generally happy with. The infopane was
unpopular at first, but I think the feedback has made it a lot better,
and a lot of people have come around on it. Right now, there are
certainly mixed opinions: some people like it, some don't. Most have
no strong opinion. I'm not sure which side I'm leaning towards now,
but as I've mentioned a bunch of times, I may wind up bringing manual
sizing if too many people thinks this sucks.
More information about the Devel