Philosophy of minimal disturbance of existing users

Joshua Cranmer pidgeot18 at
Fri Sep 14 21:20:55 UTC 2012

On 9/14/2012 12:49 PM, Kent James wrote:
> As another user interface issue, I just tried some new Daily runs, and 
> immediately notice that my menus are gone.
> That forces a dramatic change in the way that I am used to working. 
> While I understand the reasons behind it, and probably agree with 
> them, for the typical user who had gotten used to doing things a 
> certain way, this is just an annoyance. Today may not be the day that 
> I really wanted to stop what I was doing, and figure out how to turn 
> the menus back on. The reaction of most existing users, IMHO, is going 
> to be a mild round of cussing at the arrogance of developers who are 
> constantly pushing change on users who really don't want it. (Now I've 
> been the guy getting cussed at in the past, and there are times that 
> change is needed, I understand that).

Our userbase is by no means homogeneous: some people don't notice or 
don't care about the change, while others are extremely vociferous about 
the change. The problem with suggesting that UI change should be 
minimized is by defining the threshold. Look at the removal of the 
folder columns in the tree pane; when it was done in 3.0, there was a 
vociferous reaction by users which prompted Joey to write the Extra 
Folder Columns addon. Judging from ADU and usage stats, this group of 
people would comprise about 5-6% of the population. Is that so large 
that we shouldn't have changed the UI? 90% or more of our users probably 
never noticed the feature in the first place.

Nor are users consistent over time--the introduction of tabbed browsing 
into Thunderbird was received negatively by a large subset of the 
population. Now, two tab-related bugs are in the top 51 bugs, with one 
of them making number 10 and comprising the third 6-digit bug in the top 
10 (the other two being CardDAV support and Outlook-style rich tree 
view). Actually, implementing the 9th most wanted bug would likely be 
one of these contentious you-changed-it-now-it-sucks UI issues.

There also needs to be a recognition that some of the most contentious 
UI changes are to conform to the (often vociferously negative) UI 
platform conventions. The example I have here is Firefox's use of the 
single menu button, which conforms more closely to MS's Windows 7 
guidelines. How can you tell if a user wishes to receive changes to 
improve look-and-feel integration with the OS or if the user just wants 
to be UI-conservative?

> Why not instead adopt a philosophy of minimal disturbance of existing 
> users? Couldn't we instead enable the menubar by default for existing 
> users, and only disable it by default for new users?

There are several problems with this approach:
1. If we had been following this approach for the entire history of 
Thunderbird, we would have to be maintaining the UI consistency of 
Netscape 6. Here's a link to a page with pictures: 
<>. As well as every intermediate 
release in between. I've heard complaints from UI people about having to 
support just two or three versions of UI; maintaining a dozen different 
versions of it just ramps up the complexity.
2. Testing and QA. One of the things we would have to do is start trying 
to figure out which version of the UI people were using to figure out 
what broke (doubtless, there will always be one of the UIs broken), and 
it's not clear to me that this is a simple task (getting version 
information can be surprisingly hard).
3. Migration. I migrate computers by copying my profile from computer to 
computer. But others may decide just to download Thunderbird and create 
a new profile--to your approach, this is a "new user" instead of an 
"existing user." Thus, the UI they see could be drastically different 
(especially when you consider the long-term implications: people seem to 
be working with a ~5yr upgrade cycle).
4. The press may talk about cool new UI features (positively, I'd hope). 
Then the user would look forward to the change, download the new 
version... and see no change. Oops?

The problem of determining who wants new UI and who doesn't isn't easy; 
the best way to guarantee it would be to add something like the 
Thunderbird 3.1 migration assistant, which proactively informed users of 
the changes and offered to do the changes to keep the old UI, but it's 
not clear that this is a process that would be well-received if it 
happened on every upgrade.

Joshua Cranmer
News submodule owner
DXR coauthor

More information about the tb-planning mailing list