Add RFC 108 text: driver removals for GDAL 3.11 #11862
Conversation
rouault
added
the
funded through GSP
| The decision of which driver to removal comes from a mix of "intuition" and | ||
| community feedback from | ||
| https://docs.google.com/spreadsheets/d/12yk0p8rRK4rAYO_VwXboatWkiILAjE2Cepu32lsHhM8/edit?gid=0#gid=0 | ||
|
|
There was a problem hiding this comment.
Why not consider available data such as traffic to documentation pages? This would show e.g. ~90 hits/month to the documentation for the TIGER driver, which is more than retained drivers such as JSONFG.
There was a problem hiding this comment.
TIGER may be a special case and hit statistics contain mishits because most TIGER data are not in that format.
JSONFG is a format of the future, not yet accepted as an OGC standard, but it will be soon. TIGER is a format from the past. The choise which one to keep is clear.
There was a problem hiding this comment.
I'm not suggesting that we remove JSONFG because it has less web traffic than TIGER. But I don't think we should disregard that traffic either. Given relatively recent discussion or usage of data in this format, I would not write off the site traffic as accidental.
There was a problem hiding this comment.
In the "usage" link they have used Data: County subdivision lines from US Census Bureau 2018 TIGER/line files, thus shapefiles, but it does not mean that some other researcher wants and needs to use historical TIGER data for a good reason, like in the "discussion" link.
| whole codebase enhancements. | ||
| The decision of which driver to removal comes from a mix of "intuition" and | ||
| community feedback from | ||
| https://docs.google.com/spreadsheets/d/12yk0p8rRK4rAYO_VwXboatWkiILAjE2Cepu32lsHhM8/edit?gid=0#gid=0 |
There was a problem hiding this comment.
The community was solicited for feedback about removing write support for certain drivers, with a link to this spreadsheet. The spreadsheet is more broad in its scope and indicates complete removal of certain drivers. The spreadsheet shows Interlis as the only vector driver being considered for removal, but this RFC proposes to remove 5 drivers (of which Interlis is not one?). It's not easy to understand the connection between the community feedback and the list in this RFC.
There was a problem hiding this comment.
Interlis write support removal has already been removed prior to this RFC, given blessing from his maintainer.
It's not easy to understand the connection between the community feedback and the list in this RFC.
Don't try too hard to find logic ;-)
There was a problem hiding this comment.
This is pretty-much DOGE based. We wipe off everything and finds out later what is actually needed.
No description provided.