What are bugs? Are bugs really work items?
chicks at chicks.net
Mon Oct 11 00:19:38 UTC 2004
On Sun, 10 Oct 2004, Gervase Markham wrote:
> Christopher Hicks wrote:
>> But would we have had firefox if people hadn't piled features onto mozilla
>> to the point that it needed to split up?
> In this analogy, Bugzilla is Mozilla. I think most of the developers would
> prefer not to see Bugzilla get into such a bloated state that people wanted
> to do a slimmer rewrite...
I don't view Firefox as just a "slimmer rewrite", but I can understand why
you and others do. Firefox, Thunderbird, Calendar, etc. were all part of
the "suite" and the suite was always somewhat bloated. Users viewed the
various chunks of Communicator as seperate apps, but until recently it
wasn't implemented that way. The fact that it took a while for that to
become obvious to the developers is totally understandable, but viewing
that process as tragic or avoidable is making a value judgement with which
I don't concur. The refactoring was going to happen eventually.
> Turn it around. If people hadn't piled features into Mozilla in that
> way, we wouldn't necessarily have needed Firefox.
Firefox was always needed. Occassional refactoring is just a fact of life
in good software.
>> I think Mozilla is one of the better examples of shoving tons of different
>> pieces into the pot, not particularly fearing feature creep, and letting
>> folks take as big (or small) a bowl of soup as you want.
> I cannot believe you are using the way Mozilla developed as an example to
I'm not trying to follow the Mozilla example religiously, but I do feel it
exemplifies the life cycle of good software. Folks fighting feature creep
are only slowing the inevitable growth and refactoring cycle of software.
The process doesn't have to be pretty or always right for the result to be
what the result needs to be. I'm totally open for taking the methodology
experience gained in mozilla development to heart and I think there's at
least as many things "done right" as "done wrong". Repeating the
successess and avoiding repeating the mistakes seems like a great
philosophy until you get to getting people to agree on what's a mistake
and what's a success, but I haven't ever heard of any better philosophy.
(I read a quote today from quotationspage.com that "the only thing we
learn from history is that we don't learn from history." I'm not that
cynical yet, but I see his point.)
> :-) And the browser in the Mozilla Suite is a single-size bowl of soup
Precisely. It was always a single bowl of soup. Discouraging folks from
"bloating" mozilla would only have delayed this realization
>>> Linux follows the Unix Philosophy most definitely, where all the
>>> different parts of the kernel can be modularized and are written mostly
>>> separately. (Though it is a monolithic kernel, so it does give some
>>> credit to your theory.)
>> When I said Linux I was referring to the kernel.
> The fact that it's a monolithic kernel isn't really relevant.
> The kernel is still very carefully modularised.
Which bugzilla is doing better at through time.
> The fact that the modules are linked into one big binary and communicate
> via shared memory rather than being a lot of services communicating via
> message passing is orthogonal to the level of code modularity.
That's all fine. Linux still exemplifies the "throw it all in and deal
with the consequences" philosophy. There have been numerous refactorings
through time. If folks started saying Linux is only a kernel for single
machines or kernel features were thrown out because people were fearing
feature creep it would only slow down the process.
I'm sure people can find plenty of examples where folks in the mozilla or
lkm communities stomped on something for feature creep reasons, but I
don't think that exemplifies the overall attitude in effect. YMMV. :)
Also - I'm not saying that totally opening things up and not thinking at
all about the direction or value of things is the way to go. But the
bugzilla attitude has gone so far in the other direction that it has
stifled healthy progress.
"A coupla months in the laboratory can save a coupla hours in the library."
More information about the developers