opengeolabs / shapefilemustdie Goto Github PK

going to https://switchfromshapefile.org/ results in a certificate not sure error, maybe you misconfigured something in GH?

"While the format is proprietary, the specification is open."

I found this website due to this sentence, and while I was hoping to learn more about it, there is no clarification at all to this sentence. What does it mean? I understand that the "specification is open" (and there's a link to it and everything), but what does "the format is proprietary" mean? What is the distinction between format and specification in this context?

Some suite, which will make performance tests for each format

It is fun to poke holes at Shapefile but I don't think it is constructive to highlight limitations that aren't actually fixed by any of the competition.

• topology
• 3D
• others?

The link "No coordiante reference system definition" introduced with a2197f4#diff-eacf331f0ffc35d4b482f1d15a887d3bR85 points to anchor #nocrs which doesn't exist on the page. Maybe it should instead also point to #projection? (But if so, why are they to separate list items? I know that a coordinate reference system isn't the same thing as a projection, but the ways (if any) to define one or the other are usually closely related.)

Suggest you correct index.html#L124. The .prj file is not mandatory, the .shx file is.

- files</a> (*.shp, *.dbf, *.prj). Users cannot share just one file; you
+ files</a> (*.shp, *.dbf, *.shx). Users cannot share just one file; you

refer to

http://georgemauer.net/2017/10/07/csv-injection.html

The biggie for me. I'm surprised it isn't listed! CRS specification is essential for data interchange

It is hard to take this website seriously if it recommends ESRI GeoDatabase. As it says itself, that is a "proprietary, closed format"...

I share the Spanish translation of this interesting document:

Repository fork: https://github.com/GeoTux2/shapefilemustdie
Publication via GeoTux: http://geotux.tuxfamily.org/index.php/es/geo-blogs/item/337-cambia-shapefile

Would be good to cover that alternative

There are no NULL values in numeric field data types. It is not possible to distinguish between no data or "0".

I would like to be described more about the support for 3D vector data storage and representation. I think KML is good for this case, I don't know if it's the same for the other formats, including Shapefile.

It would be great to aggregate the information about the different candidates in a table.

Something like this:

I think you are missing one of the bigger frustrations with Shapefile - not only do you need multiple files to represent a single feature type but you also need one set of files per feature type so you can get an explosion in the number of files depending on your data model.

The text already says that KML has had its heyday. I see no point in recommending a limited, useless format like it?

Hi,
A very serious limitation to the use of shape files is their inability to support arc and circle geometry. This causes any CAD data containing arc and/or circles to be turned into small straight segments that do not reflect the true geometry, are not accurate and make the files much bigger then needed.
When importing data from shape files all arc or circle data (Originally) is imported as straight lines.
Arad

I would also like to ask about formats such as TopoJSON and NoSQL databases, such as MongoDB and others.

The spec is pretty clear that all GeoJSON is WGS-84. Using coordinates in any other CRS isn't going to be very interoperable. See https://tools.ietf.org/html/rfc7946#section-4.

And that's good! one of the issues with shapefiles is that it tried to be all things to all men, so instead we can recommend different formats for different things

data interchange

For making data available and sending it places geojson, csv, kml and gml fill that role, with geojson being the leading one, csv used for super simple point data, and gml used for more complex stuff that geojson can't handle, with geojson being the recommendation for most stuff because 99% of users aren't going to have data that needs gml.

data editing

This is where geopackage, file geodatabase, and spatialite come in, and this is going to honestly depend on what software you are using, file geodatabase is probably a good choice if you're editing in ESRI and geopackage or spatialite would be a great choice for non esri world, just don't post them online when making your data available, post it as geojson.

Within the features of KML is that it is an open standard OGC.

Thanks.