On Tue, Feb 1, 2011 at 1:02 PM, <exarkun@twistedmatrix.com> wrote:
On 22 Jan, 10:14 pm, tom@recursivedream.com wrote:
>In this thread, I hope to find a resolution to the issue of the Finger
>tutorial and efforts to sufficiently improve it or remove it.
>In the course of reviewing documentation-related tickets, I stumbled
>#1148 (http://twistedmatrix.com/trac/ticket/1148). Therein, Glyph
>put down a lot of things we've been discussing and agreeing upon in the
>Refactoring Documentation thread. One of the issues still up for debate
>whether or not the Finger tutorial is sufficiently strong to survive
>documentation overhaul. There are various points against it right now:
>   - It isn't tested or even test*able*
>   - It doesn't cover "best practices" as they relate to writing
>   maintainable code, etc.
>   - It attempts to implement basically every main Twisted concept,
>often in
>   contrived or poorly-executed ways
>   - It has been said it has, "...at best, the potential for
>There are also enough tickets related to refactoring / rewriting it
>that a
>resolution would make a significant dent in the list of stale
>tickets. Among these two year-old tickets are:
>   - http://twistedmatrix.com/trac/ticket/532 - Big jump from
>finger18.py to
>   finger19.py in tutorial
>   - http://twistedmatrix.com/trac/ticket/626 - Split tutorial finger
>   into libraries
>   - http://twistedmatrix.com/trac/ticket/2205 - Documentation
>   need updating and tests
>This shouldn't be a blocker on anything Kevin and I are doing, but it'd
>nice to concurrently have discussions on issues we'll need to address
>I'm also pretty anal about ticket lists and if these aren't going
>I'd love to close them ;)

In an attempt to elicit some feedback on this, let me try casting the
issue in a different light.

Does anyone think the finger tutorial shouldn't be deleted?  Why?


I actually found it very helpful back when I was first learning Twisted.  I admit I haven't really read it critically in several years though.  If it is to be removed, I think it would be a good idea to replace it with something similar (i.e. a step-by-step implementation of a protocol).

Kevin Horn