Wednesday, November 22, 2017

My Halloween Homage to Shapefile and Johnny Rotten

It is part of punk rock lore that in 1975 a young John Lydon turned some heads with an "I Hate Pink Floyd" shirt. He took a normal band t-shirt, scratched out the members' eyes, and scrawled his message on top with a marker. Soon he was invited to join a punk band (which then named itself the Sex Pistols) and took the name Johnny Rotten. And the rest is history.


Meanwhile back in the geo world, there has been a lot of back and forth about Shapefiles. The industry definitely has a love-hate relationship with them. On one hand they are indispensable but on the other hand there are a lot of people who want to banish them. Then there is the occasionally funny, occasionally annoying @shapefile on Twitter. When I saw two opposing items on Redbubble, my idea for a Halloween costume was born. All I needed was some markers, safety pins, hair spray, and blue hair dye.

The dirty little secret is that Lydon was actually a Pink Floyd fan—his shirt was more of a publicity stunt than a statement of opinion. Likewise, I can't help but have respect for the humble data format. Shapefiles are an important part of geo history and are for all intents and purposes the first interoperability the industry has achieved. While they have their limitations and probably are not the ideal format for many modern applications, they have enabled a lot of successful activities over the years and will continue to do so for many more.

Three November Updates

I have three GeoPackage-related updates for you. First of all, the GeoPackage SWG has wrapped up its work on the Tiled, Gridded Coverage Extension (formerly the elevation extension[1]). Look for OGC to make a press release soon. There will be an open comment period followed by an adoption vote (probably electronic) by the OGC Technical Committee. Of course if you have any comments, you do not have to wait until the press release. You can go directly to the GitHub repository and open an issue from there.

Second, you may have noticed a slight change to the geopackage.org page. We now have separate links to the adopted version of the standard and the working version on the top right. We have made a few administrative changes[2] that we want to get into the public eye even though they are not yet part of an approved document. These are now incorporated to the working version which is tentatively titled 1.3_SNAPSHOT.

There is a third development that will affect the standard, though it should not directly affect implementers. OGC is developing a Tile Matrix Set Encoding Standard[3]. This document extracts the TileMatrixSet definition from Web Map Tile Service (WMTS) and makes it a independent document designed to be referenced by other standards such as GeoPackage. If OGC adopts this standard, we will update the GeoPackage standard to bring them into alignment. I don't have a timeline for any of this, but whenever this work is complete we will probably release it as GeoPackage 1.3.

Happy Thanksgiving!

[1] For a recap of this topic, see paragraphs 2 and 3 of http://geopackage.blogspot.com/2017/05/good-news-bad-news.html
[2] See https://github.com/opengeospatial/geopackage/pull/387, https://github.com/opengeospatial/geopackage/pull/392, and https://github.com/opengeospatial/geopackage/pull/394 
[3] The working version of this draft standard, available to OGC members, is here: https://portal.opengeospatial.org/files/?artifact_id=76617.