Hello
my tuppence worth on fuzzy space is possibly a fudge, but here goes.
When I worked in Worcestershire SMR (Usual caveats about official policy
apply, though I no longer work for them anyway) we discussed the problem of
data which had dodgy locational info. One suggestion (I forget whose, so
wont take any credit/blame) was the idea of a 'parish polygon', which linked
to records which could fit into that parish. This answer was suggested in
response to VCH/Domesday data, where you might be able to tell Parish, but
not more precise info, and at least could flag up the possibility of unknown
sites being relevant to Dc work/smr search/whatever. This could be on a
seperate layer etc, and turned on and off as necessary, and records could be
removed from the Parish Polygon if more accurate information was made
available. That would not work for the Anytown barrow (bottom) but might
help with some problems.
I think also the thing to consider is that, in my view, the SMR/GIS is a
guide, not a be all and end all. Only when all available data has been
looked at can you determine what the edge of a polygon means. That means
that if its an excavation report, it will be obvious that the edge is the
edge of the event, not neccessarily the monument. As a monument is a
nebulous entity interpreted from all the event info, doesnt this mean that a
polygon is a snap shot of someone's judgement of a monuments extents, at the
time the polygon was drawn, rather than a definitive description? In which
case, each polygon can be seen more as a freeze frame of an ongoing process,
rather than an end point of that process. If that is so, then the problem of
boundaries is less of an issue, as archaeologists can (and do, often, and
usually loudly) disagree over their interpretations. The key thing, though,
is to make sure that the idea of what a polygon is representing is clear to
whomever gets the data. And this view of what the polygon represent can also
change as the data is cleaned, more data added, and the edges become more co
nfident. Please correct me if I have confused myself.
Having re-read the above paragraph, I think that the suggestion of a
confidence level associated with the polygon edges seems like an
increasingly good idea. This could also be done in levels so eg Level 1 was
a point in the general area, level 2 is a polygon guestimate, etc. An
individual record could progress up the levels as the data was cleaned,
checked and added to, but some records may never be able to progress due to
incomplete/inaccurate data recording/lost data.
Finally, the point about standard symbols potentially has more importance
for exporting SMR/GIS data than importing it. If the data standard is robust
enough, we can import any data and make if fit the standards we want.
Exporting it should perhaps be done to a more universal standard. Eg In
Durham (standard Caveats apply again), we are in the process of exporting
our data for one of the districts so the Planners can look at it in regard
to PLanning policies/applications, and call us if they need advice. Only
limited info is being migrated (grid refs (the data is mainly points, thogh
current SAM's have been polygoned); name and Smr number, to allow easy
communication. Training will also be given to help the planners understand
what they are looking at. But it would be useful if we could provide that
info so that say earthworks showed up differently to AP's, and Find spots
differently to SAM's. The point about exporting Standards is relevant
because I would be happier knowing that any theme providing such info had
some guidelines created by archaeologists, rather than council IT staff or
Planner (no disrespect to them, but they arent archaeologists). Have any
other people out there done this sort of data provision, and if so how did
they deal with this issue (if they considered it an issue!)
Nick Boldrini
Assistant Archaeology officer
Durham County Council
(From Home email)
The views in this email are my own, and not Durham County Council Official
Policy
Peter iles Wrote:
> One such is 'Fuzzy Space'i.e.
> what you do with sites that are insufficiently well spatially referenced,
> e.g. Mr Smith reported a prehistoric burial mound 'near Anytown'. If you
> accept this as a valid site - Mr Smith may have provided sufficient
> information to show that it wasn't made up - then how do you plot it on
the
> SMR? The paper map approach adopted in this SMR back in the 1970's was to
> write the SMR number assigned to it on the margin of the OS 1:10,000
quarter
> sheet that contained 'Anytown' with a pencil line joining the number to
the
> placename on the map. This is not a solution on a GIS where a site needs
a
> definite location. Drawing a polygon on the GIS to surround 'Anytown' is
a
> possible solution, but where do you set the boundaries? Come to that how
> (without full excavation information) do you set the polygon boundaries
for
> the majority of site types? few will have clear, definable edges you can
> draw a line around. You can use your professional
judgement/estimate/guess
> boundaries, but what we really need is a boundary that shades out, rather
> than an abrupt line. Unfortunately this is not yet available in ArcView.
> This is perhaps just a practical reflection of a more theoretical
> difficulty, but is still important.
|