[rfk-dev] Getting this train moving

[Peter A. H. Peterson]

Quoting Eric S. Raymond:
> It's been quiet on the list; perhaps Hurricane Sandy drove everyone
> into Internetless bunkers.

I was waiting for Alexey's forgotten updates; it took him a while to
get his SF.net login recovered. (He now has commit access.)

> 1. I have removed color messages, they're now black-and-white again. The
> ergonomics of some of those color combinations just didn't work, and I
> saw no point in adding an option to enable a misfeature.

That's fine with me.

> 2. X and x exit keys are no more, following list discussion.

I would be fine with them being undocumented quit keys, although I am
a little concerned that undocumented quit keys could be seen as a bug.
(I think that Q/q and ^c are probably sufficient.)

> 3. Code now complains and exits gracefully when no NKIs are found. There
> was some talk of wiring in defaults, but I think the minimalist approach
> is effective and more in conformance with rfk style.

Great. I think I do prefer a wired-in approach and might submit a patch
at some point but this is not a bug per se.

> 4. New major version number is now 'e' (2.7182818), making the sequence
> monotonic-increasing. Policy will be for the minor number to be a count
> of shipped APIs. 

Akittenic Positioned Items?

I like this. Especially with the new codebase, and other changes like
external NKIs, e seems like a good choice that actually works with the
long minor number tradition. 

In a desire to keep the minor number "sensible", I would be OK with
shortening this to 5 significant digits (at a minimum) but am also
completely happy with up to 7. 8 is right out.

> 5. The playfield frame has been retained, as unclepedro said he was
> "warming up to it" and other reaction seemed favorable.

I think we should keep the playfield frame unless there are other
objections. 

> Issues *not* resolved:
> 
> A. Behavior on resize is somewhat wonky.

I think this should be fixed before we ship. Here are the two simple
solutions I would be in favor of (OTOMH):

1. Resize the frame but don't move the NKIs. If the frame is resized,
robot should be able to explore the space. 

2. Don't resize the frame, ever.

In both cases, if the terminal is resized *below* the original size, I
think the message "Stop, you're crushing me!" (or whatever it is in
the code) should be displayed. I think that's a fun (and frankly
surprising) behavior and I loved it the first time I saw it.

I find it somewhat confusing when the NKIs move or are respaced minutely
as a result of resizing the frame. I'd be willing to consider resizing
triggering a "shuffling" of the current NKIs (perhaps with a
Boggle-like animation) but that seems really elaborate.

> B. The "You found kitten! Way to go, robot!" message only shows if you
> set term=VT100, and does so after curses mode has been exited.

This should definitely be fixed before shipping.

> C. The two very oldest versions have not yet been merged into the repo.

We should try to do that, but it's not an impediment to shipping.

As for other tasks, I think it might be worth making a pass over the
NKI list and/or adding some new ones in honor of the apocalypse. I'll
address this in another post.

Thanks again for your work on this stuff, Eric.

-pedro

-- 
Peter A. H. Peterson
Graduate Student Researcher
Laboratory for Advanced Systems Research
University of California, Los Angeles