[Talk-us] Special issues in LA remap
steveall at softworkers.com
Thu Jun 14 00:16:31 BST 2012
>At 2012-06-08 13:49, stevea wrote:
>>I discern a vague remap plan ordering. If you can, please sharpen
>>this up or correct it if it is outright wrong:
>>1) "The redaction bot" (still being written) will do much (to
>>"ease in" the license change),
>Not sure what you mean by "ease in" here. It will return much of the
>area that was touched by non-agreeing users to TIGER05 state, which
>is probably pretty bad.
I am performing very large hand-waving gestures when I totally guess
at what the components of the algorithm of the redaction bot are or
might eventually be. Perhaps that is the point: wider discussion of
the particulars of the redaction bot algorithm is precisely what I'm
trying to see happen, so I can better discern a guide to my own
prioritization of how/whether to remap as smartly as I can, again, I
am not alone here.
>>2) "Nudge nodes to the location where they belong" is smartly
>>next, somewhat unspecified as to how/where,
>"Nudge" is not the right word. Re-align thousands of ways is more
>accurate, and man-years of work.
And once again, we have hit upon how huge is the work required, with
the merest whisper of words to describe the many tasks. I don't want
to (simply) fall into a world of "wishing" that license issues were
simply to magically evaporate, rather, better understand the present
and future of the evaporation process(es). So, people who are in
positions to make redaction bots happen, speak up, say what might be
done, be wide and inclusive in plans you may be / are making (to
better clean the data to meet CT licensing). I realize "this" is
many big tasks, I'm encouraging us to roll up our sleeves and perhaps
specify some goals, which are the in-between of where wishes turn
into results. I think it helpful to state explicitly here/now that
some of us are in earlier stages of understanding what might happen,
and other are more in the middle, and better able to explain. (I'm
in the former camp, as are the vast unwashed masses, you and or
others are now being encouraged and challenged to explain how we're
going to pull us all forward).
>>3) Perhaps update with TIGER 2011 data in select areas, also
>>unspecified as to exactly where.
>I only mentioned re-import for areas that were only touched by
>non-agreeing contributors, and nobody else, and the idea was to do
>it before the redaction, to clean those areas. Thinking about it,
>though, I guess it doesn't matter whether it happens before or after
>redaction for such an area.
>I don't find too much value in trying to apply TIGER11 to areas that
>have already been surveyed/corrected by people, since, while it's
>geometry is better, human alignment to Bing is still better. It also
>seems to have newly-introduced errors in it. I recently used it as a
>reference for Fort Irwin, CA, and I found there were naming and
>other errors in TIGER11 that were not in TIGER05. When trying to use
>it as a naming reference in other areas, I've found it not
>particularly authoritative (i.e. as likely to be wrong as right when
>there is a naming conflict among sources).
So "well-targeted" TIGER11 data to replace TIGER05 data, the large
effort of which takes into account existing edits (large areas where
this is true, actually) are most certainly a part of this parallel
effort. That is a fair bit of effort and likely coordination as well.
Ahhhhh. Talk digests are such a stilted register in which to have
these conversations. I'm trying to make the best use of this
channel. Speak up, one and all: remap, licensing issues have taken
many twists and turns since April 1st, and it is time to freshen up
how we focus our efforts.
More information about the Talk-us