Closed omad closed 6 years ago
Was never fond of EO as the name of the extension, "satellite" by itself is certainly way too general. Are you recommending "satellite-eo" to distinguish this from drone imagery? It is a bit more descriptive than just EO, but doesn't remove all the ambiguity. If SAR is an extension that is also a satellite. Mosaic imagery is also made up of satellite images.
Earth Observation generally refers to the set of satellites that are used for environmental monitoring and for mapping. With the EO extension we really want to capture the set of satellites that include visible bands, but also may include near-infrared, short-wave, but also possibly mid-wave and long-wave IR as well. I'm not sure there's a better choice than EO, but fine with the additional qualifier that "Satellite-EO" gives so that we might eventually have a "Drone-EO" extension.
EO as defined at the moment fits very well to 'Satellite-Optical-EO' domain, with some fields that also apply to other domains such as SAR (see #155) but others that do not. The nice thing about being more specific about the extension domain is that the domain works as a guide to people implementing the SPEC, defining the attributes that are considered to be relevant by the users.
I would go to with extensions for each domain such as Satellite-Optical, Satellite-SAR, etc., and later when this is more consolidated try to work on structuring the common attributes between domains.
+1 to renaming what's currently EO to Satellite-Optical
Not really fond of satellite-optical. Optical remote sensing refers to the part of the spectrum of visible through short-wave IR. However, mid-wave and long-wave IR bands (which are not considered optical because of the thermal sensing component) are often included and packaged together (ie Landsat) and can easily be included in this extension, especially MWIR which has the same units of reflectance as the optical bands.
So having optical is a bit of a misnomer if data using this extension includes MWIR and LWIR which is certainly possible. What is wrong with satellite-eo?
"satellite-eo" would imply it covers radar data, but it does not. Conversely, it can apply to aerial imagery, but the name precludes it. Maybe "passive" for optical/IR/MWIR/microwave and "active" for radar?
+1 on the change - happy for whatever name you all agree on.
And yes, it's cool if other extensions like SAR and Drones share some of the fields here - indeed it's encouraged. Once we get those fleshed out we can elevate some to a common extension that they can share.
@simonff Yes, that is true, I suppose that having a name that suggests SAR when it does not is worse than a name that suggests it only includes Visible-SWIR when it does not.
So something including "optical" would be the most descriptive, but we should include in the description that many "optical" sensors often include midwave and even longwave IR bands even though they technically aren't optical.
Maybe eventually we might see a separate but similar extension for thermal remote sensing. But for now I think that the occasional sensor that publishes thermal bands alongside optical bands is fine.
so, with that in mind I propose we change the name to EO extension, standing for for Electro-optical, which refers to the Vis-SWIR range, rather than Earth Observation. :-)
It could be sat-eo, but as pointed out this extension as is can be used for drone imagery.
@matthewhanson you've got me thinking on that but my view is that, at least from the payload side, TIRS for example is an optical sensor - it has an optical subsystem. I agree however that from the applications side the term optical and its connection to visible may be confusing. I'm ok with Electro-optical.
I support EarthOptical for its pun value.
Reiterating Matt Hancher's concern that I mentioned in giitter: "Maybe it's better to have fewer larger extensions (even if some of their fields go unused in some cases) rather than a lot of hyper-specialized extensions."
+1 for EarthOptical. But again, I'm cool with whatever, and Electro-optical sounds good too. And keeping eo prefix sounds good.
I think we should aim to get to some sort of 'master extension' for all the different types of sensors, so aspire to what Hancher proposes. But I think we should start narrower with things we know, and let the parts (drone, sar, mosaics, etc) develop a bit independently, and then group them. But perhaps we could put them all in the same folder under extensions? So group them by a folder, but stop short of saying 'this is one larger extension'...
PR welcome if someone is up for doing the rename.
Every extension should live in its own folder anyway and not on the top level of extensions like EO and collection. You usually already should have three files at least (schema, spec, example).
Once there is another +1 I'll prepare a PR.
+1 for Electro-Optical (did I already vote?)
EarthOptical: @simonff and @cholmes Electro-Optical: @matthewhanson, @fredliporace and @cholmes It was a draw before, that's why I asked for another +1. So seems like Electro-Optical is the winner.
I actually meant to support ElectroOptical as well, just to get things going, though this Freudian slip indicates that this name is fragile as well.
resolved with PR #276
Several people have pointed out that the current
eo
extension is very specific to satellite earth observation data, including many fields which are specific to satellite data, but not eo in general.To avoid confusion we should rename the extension and prefix to better indicate their purpose.