Hi-aïl,
Le 18/04/2016 14:13, Jean-François Moyen a écrit :
Le 18/04/2016 12:58, Pierre Chevalier Géologue a écrit :
(Ok, shortly, and on non-troll portions :-) )
True of course, but in that case what we have is a sample ("grab" in
your terminology, so a point). We crushed it and extract 100's of
zircons, 50 of them or so are mounted in epoxy, polished and prepared,
and each zircon is analyzed using laser ablation in one, two or many 23
µm spots. For each spot we record a continuous ablation signal of about
60s, at a rate of one reading every 1/10th of second or so, for 12
isotopes. We then average the one minute of signal into a "value" for
each isotope, and convert that into an age, and combine all the ages
from all the grains into something we call the age of the rock.
So -- Ok, I *have* heard about one-to-many relations, and no doubt all
the data can be bound to the sample location. In that case however, this
is not the most important information, the real key information is the
averaged value for each spot -- AND its location with respect to the
zircon grain (core, rim...). So in that sense, this is "non spatial"
data (or vaguely spatial).
Also note the nested data : one sample = 50 grains, one grain = 2-5
spots, 1 spot = 12 time series. Again, I know about one-to-many links,
but we are moving away from a simple "map" paradigm, and that was my
point: there is some geological data that is not direclty map based, or
only in a very tangential way, and for which a map-based treatment (GIS
or CAD or 3D, does not really matter here) is at least overkill, and
probably prevents us from catching some key features.
What I retain, after digesting your arguments a bit, is that we may also
have to limit ourselves to a reasonable space dimension. For instance,
in mapping, one meter seems to be reasonable; although the decimeter can
be useful for detailed face mapping. Centimeter is the reasonable best
accuracy needed along drill holes. Going under the millimetre seems to
be beyond all reason, in my humble opinion. Unless someone disagrees?
I do. The attached image is an electron microprobe map of a pyrite
crystal. Note the scale :-)
And again, the connection with "real" geographic space it weak --
granted, the sample comes from somewhere, but (i) it is not the relevant
information [it could come from the mine's dump for all we know, the
story wouldn't be different]; (ii) even if it's not, we do not know the
absolute coordinates of the pyrite : the sample is probably not located
in space with a precision down to micron level (and it's been hammered
on, grabbed and rotated, cut....).
(NB: from here :
https://www.researchgate.net/publication/276486183_Implications_of_pyrite_geochemistry_for_gold_mineralisation_and_remobilisation_in_the_Jiaodong_gold_district_northeast_China
by the way)
- Déconnectées de la réalité du terrain : un modèle thermo par exemple
va ressortir une grille de propriétés (composition des minéraux) dans
l'espace PT. Ou une donne expérimentale faite sur une roche synthétique.
Well, are synthetic rocks what a geologist deals with??? (dunno, I
never had...)
Ever heard of experiments ? ;-)
Try that, for instance :
https://www.google.fr/search?q=experimental+study+gold+pyrite
Although some of the hits are actually analytical studies on real
material, this gives you an idea of what experiments can mean. In that
case clearly you use synthetic material (or at least material spiked in
Au).
Bref, c'est quoi une donnée géologique pour vous ?
Big Question!
I will *just begin* to try to answer this question, from a very basic
field geologist's point of view.
And this is my very point.
To each of us, a "geological data" is a different animal. In you case,
field based, it is spatial (2 or 3D).
But, being more of a geochemist (academic kind), a large portion of my
data is either not spatial -- or the spatial information is of little
use to me. In this paper for instance
(https://www.researchgate.net/publication/245540556_Forty_years_of_TTG_research),
although we do know the location of each and every point in our figures
(such as fig. 12 for instance), this is an information we do not use at
all. In fact I'm lying -- in Fig. 12 we know the location of maybe 1/3
of the data, another third we know only vaguely ("Stolzburg pluton, NW
of Tjakastad") and the last third we have only very poor information on
("Carajas province, Brazil"), if at all. Yet in this case ... we did not
care.
I'm not saying here that "my" sort of data is better than yours (well
actually I am, but that's beside the point -- oups, sorry, get that
troll back into its cage !! :-) ).
I'm merely saying that we all dealwith different types of data, and
therefore your (or mine, or anybody else's) idea of what is "geological
data", what properties it should have, and by extension what sort of
tool do we need to process it may differ from the next guy's.
And so -- writing a tool that suits you, or me [done that already
to a large degree, it's called GCDkit, thanks very much] may, in fact
will result in a tool that is totally useless for him. Or she. Or that
other guy there.
That's not necessarily a crisis, and at any rate should not stop us from
doing something. But, well, it is just impossible to develop a tool that
will work for all geologists.
I would agree, though, that a spatial component is something that is
important to most of us; coming from an academic background, it is
perhaps the one thing that separates us more clearly from our fellow
natural scientists (physics, chemistry, biology, etc.).
But then, are we just talking about glorified GIS here ?
3D GIS, in a way ?
