In doing the Bugzilla Survey for the Mozilla Foundation, one of the organizations I surveyed was ALMWorks, the makers of Deskzilla. I customized the questions for them, in terms of what was difficult when integrating with Bugzilla. I've created this bug to track all the major issues that you would have if you were writing a new Bugzilla client against the latest trunk code. The most important issues to track are the ones that either: * You can't do at all or * You can't do without screen-scraping. Clients should never have to screen-scrape.

Max: Great to see interest in this. Rob: Please list all of our problems that are not already subtasks.

Igor (the Deskzilla developer) had some really good thoughts on things we should think about when developing a new feature, in terms of client integration: - How can a remote client download a bug's current state [pertaining to that feature]? (remember bulk download!) - How can a remote client search bugs that have that specific state? (Keep Boolean charts full with new fields and operators) - How can a remote client get all related metadata about your feature (field names, numeric ids, options, is the feature enabled, what options are valid at what times, etc.)? - How can a remote client upload changes?