From Dyna
Jump to: navigation, search


The Request Tracker is our request tracker. This is where we keep track of issues we need to solve -- everything from long-term design issues to bug reports and help requests from users. It includes all the subprojects of Dyna.

Most conversation among developers should take place within the request tracker. That way, we can see where we stand, search for old discussion, link bits of discussion together, prioritize, etc.

Priority levels

These are taken from the RT wiki. (When we change them, they'll become "inspired by" the RT wiki.)

  • Priorities 11-20 : A minor inconvenience. There is a good workaround
  • Priorities 21-30 : An inconvenience. There is a workaround
  • Priorities 31-40 : A major inconvenience. There is a troubling workaround
  • Priorities 41-50 : A real problem. Someone can not do part of his work until this is resolved.
  • Priorities 51-60 : A disaster. Someone can not work at all until this is resolved.

Not sure how to prioritize tickets for the design of new features. They shouldn't block urgent fixes, but they shouldn't be forgotten either. It would be nice to have a way to distinguish a bug's urgency from its long-term importance. We could introduce a custom field (some bug trackers distinguish priority from severity), or use RT's "due date" field.

Email interaction

You can email RT to start a new ticket or add discussion to an old ticket. Every queue has a dedicated email address for ticket submission at dyna dot org (details here). Also, discussion on any ticket will be emailed to the developers who are "watching" that ticket's queue.

Thus, RT will maintain conversation archives, grouped by topic.

Any random fool can email these addresses. So we have SpamAssassin set up to prevent spam from getting into RT (and through RT, into developers' mailboxes).

Web interaction

Most RT controls are available only over the web. For example, you can create new tickets and add to old tickets by email, but you must log in via the web interface for some other functions, e.g.:

  • close tickets that have been resolved
  • view lists of tickets (e.g., search)
  • set ticket priorities and due dates
  • link tickets to one another with relationships like "parent," "depends on," "refers to"
  • merge tickets that focus on the same issue

If you have an account on the NLP network, your account at has the same username. Your username does have a password set for it; please change the default as soon as you can.


The search box in the upper-right corner of the web interface appears to allow three kinds of queries:

  • a ticket number brings up that ticket
  • a queue name brings up all new or open tickets on that queue
  • anything else searches the subjects of all new and open tickets

If you want to do something fancier, click on Tickets / New Query in the left margin. This allows complex searches. By default, these searches are still confined to "new" and "open" tickets. You can override this by changing "Status is -" to "Status isn't rejected" on the search form.

Full documentation

MIT has written some documentation for RT users. Consider skimming the "Basics of RT" document.

The official introduction to RT is a decent quick overview. It is part of a documentation wiki.

Personal tools