[OH-Dev] Mission Assumptions

[Asheesh Laroia]

Excerpts from Grant Bowman's message of Sun May 13 14:43:51 -0400 2012:
> I have been having a problem with a mission that Asheesh asked me to look at.
> 
> https://openhatch.org/bugs/issue429
> 
> This kind of bug should become obvious as more and more people try it
> and fail. The error messages are cryptic and I think the test need to
> be loosened. Nomatter how 'right' the system is if it fails for new
> users in a very typical use cases it needs to be changed somehow. The
> point of the missions is to help new users learn, not encounter
> problems with the learning system no matter how many tests it might
> pass. If this is for newbies as I suspect it is aimed at, better error
> messages may be the right way forward with hints for typical types of
> failures or relaxing the pattern matching a bit.
> 
> Am I understanding the intent of this mission and the nature of this
> particular bug correctly?

Hey Grant (and all),

Yeah, you ran into a pretty bad bug. We definitely don't consider that
buggy behavior to be 'right'!

What error messages did you get specifically? We're committed to making
the training missions easy for newcomers, and that includes providing
helpful error messages for all sorts of likely mistakes.

It's pretty clear there was a bug here; again, sorry about that.

With the fix to issue429 deployed, would you give the svn mission
a fresh run-through, and make sure it's working properly for you as well?

You seem to be understand the intent of the mission right, and I think
the bug as well. When I said to you, "But it passes the tests", I meant
that I was disappointed that there was (clearly) a problem with the
automated testing we're doing that it fails to catch this problem.

It's probably best to add more end-to-end testing for the missions; I've
filed issue732 for that.

I hope that explains things! Thanks for the testing, and your patience.

-- Asheesh.