Bugzilla Helper Replacement

Wednesday June 19th, 2002

In the words of its maintainer Gervase Markham, the Bugzilla Helper "has been creaking for some time now. It's buggy, uses JavaScript, uses popups, has dodgy formatting and is generally broken."

The good news is that a replacement is in development. Read Gerv's newsgroup posting for full details and send feedback to netscape.public.mozilla.webtools.

#5 An idea for a design change in bug report

by jmd

Friday June 21st, 2002 8:26 AM

You are replying to this message

For some time, I've been using the Bugzilla Helper, been reporting bugs, and came to the conclusion the interface is not optimal.

I always really took care of avoiding duplicate, but sometimes even fastidious researches were not enough to avoid that. With the current interface, as Mozilla becomes more popular, number of duplicates will grow exponentially.

The most important is give someone who arrives on Bugzilla a way to known easily whether his bug is alreay known and lead him to discover what similar bugs there is in Bugzilla, and only then add his own.

The current bug report is designed in a way only a Mozilla developer can correctly report bugs. The casual user know what problem he has, but how should he know what part of Mozilla is is related to ? Currently this is the very first thing he has to choose !

After some thinking, I came to the conclusion that the interface that would help most to reduce the number of duplicates would be a tree-like interface.

The top of the tree would be a list of simple question about the type of bug, then about where it happens, then more and more detailed questions about what happened exactly. When the user arrives at the leaf level, he gets a list of very similar bugs. Whenever needed, a new sublevel can be created to keep the number of leaf low. The leaf level should show be default currently living bug (all those that are not closed/solved since more than 1 month).

At any level, the user has a predetermined list of choices to continue, and is told he should try to find one that correspond to what problem he has. Several different path could lead to the same destination to easier things.

If he is unable to classify his bug inside the tree, then he should just be given a way to enter it without classification, and then a dedicated team will try to find the proper classification, and undertand why the bug couldn't be classified, maybe create new entries in the tree, so that's this kind of bug becomes easier to classify.

I don't know exactly how this could be built on top of the current Bugzilla, but this could be in a quite independant database, with the only direct links that the classification of the bug would give a list of key word in the bug description and the leafs would be a bug number inside Mozilla.