Less bugs & more features by delaying the code freeze

I would always have liked to be an optimist, but sometimes I can't. So my feeling is that Dries' call for testing will produce some results (written by a test case creating minority) but won't be nearly enough to defer the code freeze to October, as Dries has promised for a fully covered Drupal core.

So, why not have a proposal for writing lots and lots of simpletests and getting Drupal core fully covered? I can see various benefits with this:

1. The student gets to see all of Drupal and will be masterly skilled both in Drupal (and testing) after understanding all those different parts - which comes automatically because of writing simpletests.
2. Drupal 7 can become the desired killer release as all the important pieces can go into Drupal before the code freeze - which is deferred until October because there's so much coverage.
3. Drupal quality and bugfree-ness goes through the roof.
4. And last but not least, modularity: no huge module that needs to be maintained by the student after the SoC has ended (or otherwise goes unmaintained and obsolete). No long-time commitments for the student. No "need to fix it to work at all", simpletests are small (and many) enough to be valuable each on its own. Important point, really.
5. Large direct contribution to Drupal core (fame!) without the same level of scrutiny that run-time code gets reviewed with. Doesn't need to be perfect from the start.

Possible drawback: The community doesn't take on its responsibility to also do lots of simpletest, and leaves its work to the student saviour.