Role of addons

Kent James kent at caspia.com
Fri Sep 14 17:06:26 UTC 2012


There is a continual debate that we have in bugs about whether to 
include certain features in the core code or not. I'd like to see if we 
can develop a consensus about a philosophy here.

One of the risks that we face, and have succumbed to in the past, is to 
add features and their complexity to the core code primarily because a 
developer was motivated to do so. I confess that I have been guilty of 
that in the past.

One of the results of that is well stated in a review of Thunderbird by 
Linux Magazine <http://www.linux-mag.com/id/7788/> :

"/One of the problems with Thunderbird is that it doesn't seem to fit 
with most users' needs for email. That is, it doesn't work well for 
business users who need features like calendaring and groupware 
connectivity, and it doesn't work well for casual mail users who have 
mostly adopted Webmail or whatever ships on their computer. ... It's too 
complex for casual users, and not full-featured enough for business users."

/In terms of positioning of the product, I think that we should push at 
both of these boundaries, at the same time realizing that we will 
probably never be able to displace webmail for the most casual user, or 
Outlook for the most complex enterprise.

To push at the simplicity boundary, we must be willing to reduce the 
complexity of the user interface. One of the main ways that we have to 
do that is through addons. The user interface for features that are only 
going to be used by a tiny fraction of our users should be pushed to 
addons, and not included in the core code.

In the long run I would like to see us do this more explicitly by adding 
a category of addon that is maintained along with the core product, and 
shipped with the core product. So these addons would have the same 
commitment to support as any core feature, but are included as addons to 
reduce the overall complexity of the product. Good candidates for that 
in the long run would be chat, calendaring, RSS feeds, bayesian junk 
processing, advanced security models, and advanced search and filter 
functionality.

In the short run, I would encourage us to be selective about adding new 
features that complicate an already overwhelming user interface. Just 
because a developer is motivated should not be a good reason to add new 
user interface items for rarely needed features.

Comments?

:rkent
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/tb-planning/attachments/20120914/87437dac/attachment.html>


More information about the tb-planning mailing list