Media type

[m.mohr]

Dear all,

the GeoTiff SWG decided on a media type for GeoTiff: image/tiff;
application=geotiff
See
https://github.com/opengeospatial/geotiff/issues/34#issuecomment-514078289

Based on the assumption that IANA accepts it, I'm wondering whether COGs
could also get a recommended media type now.

Two proposals from my side:
- image/tiff; application=geotiff; cloud-optimized=true
- image/tiff; application=geotiff; cog=true

What do you think?

Having a recommendation would be very helpful for STAC and openEO, for
example.

Best,
Matthias
_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Vincent Sarago]

Thanks Matthias,

I’m +1 on `image/tiff; application=geotiff; cloud-optimized=true`

> Le 24 juill. 2019 à 13:35, Matthias Mohr <[hidden email]> a écrit :
>
> Dear all,
>
> the GeoTiff SWG decided on a media type for GeoTiff: image/tiff; application=geotiff
> See https://github.com/opengeospatial/geotiff/issues/34#issuecomment-514078289
>
> Based on the assumption that IANA accepts it, I'm wondering whether COGs could also get a recommended media type now.
>
> Two proposals from my side:
> - image/tiff; application=geotiff; cloud-optimized=true
> - image/tiff; application=geotiff; cog=true
>
> What do you think?
>
> Having a recommendation would be very helpful for STAC and openEO, for example.
>
> Best,
> Matthias
> _______________________________________________
> COG mailing list
> [hidden email]
> https://lists.osgeo.org/mailman/listinfo/cog

_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Sean Gillies-3]

Matthias, Vincent:

A boolean mime type parameters is not the way to do this. Defining whether "cloud-optimized=false" has any meaning is just the first weird detail that would have to be handled. Something like "profile=cog" would be better and would be open to other profiles.

Debates about the name of the new parameter may be moot because TIFF has registered only the "application" parameter: https://www.iana.org/assignments/media-types/image/tiff and according to Mark Nottingham [2] third-party parameters are seen to have the potential to break the internet [2]. But maybe IANA will be fine with it in 2019.

If a new parameter were not an option, a "geotiff-cog" value for the application parameter would be consistent with the TIFF spec and IANA registration and is worth a consideration. "geotiff" is also a completely new parameter value and practically no software understands what it means right now. There's a small window perhaps to mint a second parameter value and get all the software that would implement "application=geotiff" on board with "geotiff-cog" (or whatever).

Sidebar: I feel like "cloud-optimized geotiff" is pretty vague about the implementation and "cog" even more so. If this sub type were to be formalized, I'd hope we could come up with something more technically clear and less buzzy like "http optimized" or "direct access".

[1] https://tools.ietf.org/html/rfc6906

[2] https://www.mnot.net/blog/2012/04/17/profiles, comment on Thursday, April 19 2012 at 5:35 AM

On Wed, Jul 24, 2019 at 12:46 PM Vincent Sarago <[hidden email]> wrote:

Thanks Matthias,

I’m +1 on `image/tiff; application=geotiff; cloud-optimized=true`


> Le 24 juill. 2019 à 13:35, Matthias Mohr <[hidden email]> a écrit :
>
> Dear all,
>
> the GeoTiff SWG decided on a media type for GeoTiff: image/tiff; application=geotiff
> See https://github.com/opengeospatial/geotiff/issues/34#issuecomment-514078289
>
> Based on the assumption that IANA accepts it, I'm wondering whether COGs could also get a recommended media type now.
>
> Two proposals from my side:
> - image/tiff; application=geotiff; cloud-optimized=true
> - image/tiff; application=geotiff; cog=true
>
> What do you think?
>
> Having a recommendation would be very helpful for STAC and openEO, for example.
>
> Best,
> Matthias

--

_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Pfister, Marc-2]

> Le 24 juill. 2019 à 13:35, Matthias Mohr <[hidden email]> a écrit :
>
> Based on the assumption that IANA accepts it, I'm wondering whether COGs
> could also get a recommended media type now.

+1 on `image/tiff; application=geotiff; cloud-optimized=true`

- Marc Pfister
_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Brad Hards]

As I posted over at https://github.com/opengeospatial/geotiff/issues/34 I think it would be better to have tiling and overviews brought out as characteristics of the geotiff MIME type. So modify the existing registration.

Brad

-----Original Message-----
From: COG <[hidden email]> On Behalf Of Pfister, Marc
Sent: Thursday, 25 July 2019 6:39 AM
To: [hidden email]
Subject: Re: [COG] Media type

> Le 24 juill. 2019 à 13:35, Matthias Mohr <[hidden email]> a écrit :
>
> Based on the assumption that IANA accepts it, I'm wondering whether
> COGs could also get a recommended media type now.

+1 on `image/tiff; application=geotiff; cloud-optimized=true`

- Marc Pfister
_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Even Rouault-2]

> Sidebar: I feel like "cloud-optimized geotiff" is pretty vague about the
> implementation and "cog" even more so. If this sub type were to be
> formalized, I'd hope we could come up with something more technically clear
> and less buzzy like "http optimized" or "direct access".

https://gdal.org/drivers/raster/cog.html#file-format-details
could serve as a base for this.

--
Spatialys - Geospatial professional services
http://www.spatialys.com
_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Chris Holmes-5]

I'm +1 on cloud-optimized=true over cog

And I do think we should formalize it into a spec, starting with the gdal details as the base.

I started a repo ages ago https://github.com/cogeotiff/cog-spec but never got anywhere past porting the GDAL page over, as I got overtaken by other tasks :(

But doing it in its own repo had mostly been since OGC hadn't established an actual geotiff spec / repo, so I'd say it probably makes sense to make an extension for COG of some sort that builds on https://github.com/opengeospatial/geotiff/ and hopefully is even in the same repo.

On Wed, Jul 24, 2019 at 1:56 PM Even Rouault <[hidden email]> wrote:

> Sidebar: I feel like "cloud-optimized geotiff" is pretty vague about the
> implementation and "cog" even more so. If this sub type were to be
> formalized, I'd hope we could come up with something more technically clear
> and less buzzy like "http optimized" or "direct access".

https://gdal.org/drivers/raster/cog.html#file-format-details
could serve as a base for this.

--
Spatialys - Geospatial professional services
http://www.spatialys.com
_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog

[Pedro Gonçalves]

I would suggest the usage of the profile link relation
image/tiff; profile=<URI-reference>

the value to be defined (for example it could be https://www.cogeo.org/ )

Further reading at
https://tools.ietf.org/html/rfc6906#section-3.1
and check also
https://ruben.verborgh.org/articles/fine-grained-content-negotiation/

FYI this is already being used by GML
https://www.iana.org/assignments/media-types/application/gml+xml

> On 25 Jul 2019, at 04:34, Chris Holmes <[hidden email]> wrote:
>
> I'm +1 on cloud-optimized=true over cog
>
> And I do think we should formalize it into a spec, starting with the gdal details as the base.
>
> I started a repo ages ago https://github.com/cogeotiff/cog-spec but never got anywhere past porting the GDAL page over, as I got overtaken by other tasks :(
>
> But doing it in its own repo had mostly been since OGC hadn't established an actual geotiff spec / repo, so I'd say it probably makes sense to make an extension for COG of some sort that builds on https://github.com/opengeospatial/geotiff/ and hopefully is even in the same repo.
>
> On Wed, Jul 24, 2019 at 1:56 PM Even Rouault <[hidden email]> wrote:
> > Sidebar: I feel like "cloud-optimized geotiff" is pretty vague about the
> > implementation and "cog" even more so. If this sub type were to be
> > formalized, I'd hope we could come up with something more technically clear
> > and less buzzy like "http optimized" or "direct access".
>
> https://gdal.org/drivers/raster/cog.html#file-format-details
> could serve as a base for this.
>
> --
> Spatialys - Geospatial professional services
> http://www.spatialys.com
> _______________________________________________
> COG mailing list
> [hidden email]
> https://lists.osgeo.org/mailman/listinfo/cog
> _______________________________________________
> COG mailing list
> [hidden email]
> https://lists.osgeo.org/mailman/listinfo/cog

_______________________________________________
COG mailing list
[hidden email]
https://lists.osgeo.org/mailman/listinfo/cog