Minders and pipes

[Michael Allan]

Component pipes are now coded.  (Conseo, please see PS at bottom.)
http://zelea.com/w/Category:Component_pipe

Component pipes divide a consensus text into a complex of parts that
may vary more-or-less independently.  See C and D:
http://zelea.com/w/Stuff:Votorola/p/position_space_rationalization

The pentagon component (C) has two variants: the wedge-filled variant
(left) and the star-centered variant (right).  Currently it nominates
the star-centered variant (probably because it has more votes) which
thus becomes the wild variant.  Meanwhile, the star component (D) of
that variant has two variants of its own: the abstract variant (At)
and the bi-sected one that is currently wild.  (It's actually more
complex than shown, because components can cut across each other.)

Terminology: http://zelea.com/w/Stuff:Votorola/glossary


DEMO OF THE BASIC MECHANISM
---------------------------
Here's a component pipe in votespace:
http://zelea.com:8080/v/w/Votespace?p=Sys!p!sandbox&u=Pipe%2B1138-ZeleaCom
Currently it looks no different than any other pipe.

Here's the pipe page in the wiki:
http://zelea.com/w/User:Pipe%2B1138-ZeleaCom

  (maybe choose V on the toolset switch if you're running Chrome,
   Safari or Firefox - but it's not important)

Note that the component pipe has no draft content.  Nevertheless
here's the textual difference between it and the downstream candidate:
http://zelea.com:8080/v/w/D?k=6636.6638-6637
That's actually the difference vs. the upstream wild variant.


   Wild Variant           http://zelea.com/w?oldid=6638
     |
     V
   Component pipe         http://zelea.com/w?oldid=6636
     |
     V
   Downstream candidate   http://zelea.com/w?oldid=6637


Likewise here's the difference between the wild variant and the
component pipe: http://zelea.com:8080/v/w/D?k=6638-6636.6637
By the same token, that's actually the difference vs. the downstream
candidate.


PURPOSE
-------
So not only does a component pipe transparently shunt vote flow (like
all pipes), but it transparently shunts text flow.  If the votes
happen to shift in favour of some alternative content for that
component (a different variant), then the component pipe enables the
new content to flow throughout the tree, while retaining the
possibility of switching back at a later time.  This is the crucial
thing.  There's no loss of information.  In biological terms, the
population retains its genetic diversity, which is the pool of
hard-won solutions/designs on which its further evolution depends.

Or in personal terms, the drafters have unlimited space and time in
which to develop alternative content.


NEXT
----
Next I think we're ready to experiment with the practices in the
social world.  We had to develop rough tools (Galilean telescopes)
just to get a first picture of the practices.  Now we need to develop
the practices in conjunction with the tools and see what comes out of
the combination.

First I'll upgrade my computers and do a full release of Votorola.

-- 
Michael Allan

Toronto, +1 416-699-9528
http://zelea.com/


PS,

Conseo, The harvester code is marked CWFIX where it needs attention.
It should be able to run okay, but it can't harvest the new diff URLs.

$ ff-grep CWFIX
/home/mike/work/votorola/s/wap/HarvestWAP.java:
         //// WAP.doGet already fine-logs request, echoing same info clutters log, CWFIX
                    out.name("key").value(new DiffKeyParse(rs.getInt(9), // CWFIX obsolete form

/home/mike/work/votorola/s/wap/DiffFeedWAP.java:
                    look.setKey(new DiffKeyParse(rs.getInt(9), rs.getInt(10), // CWFIX obsolete form

/home/mike/work/votorola/a/diff/harvest/cache/DiffMessageTable.java:
                // CWFIX obsolete form, e.g. by storing p.key() instead

/home/mike/work/votorola/a/diff/harvest/cache/HarvestCache.java:
                            new DiffKeyParse(pMap.get("a"), pMap.get("aR"), // CWFIX obsolete form
                } catch (IOException e) { // CWFIX review this change