Author: tom[at]compton.nu
[Added to the original trac issue at 2.21pm, Sunday, 9th March 2008]

I'm not going to change the semantics of the API merely based on one person's opinion - if you think empty values should be prohibited then I suggest you start a discussion on the mailing lists and try and get a consensus from the community as to the correct approach.

Author: Robx
[Added to the original trac issue at 9.19am, Wednesday, 12th March 2008]

Sorry, I must have forgotten to set the component. I meant to file this against merkaator. I was just hoping for consistent behaviour between the different editors, not for a change to the API.

Author: koying
[Added to the original trac issue at 9.37am, Thursday, 13th March 2008]

If Tom says the API actually allows this, there is no reason for Merkaartor to disallow it.
Furthermore:

• An empty KEY is rejected
• I doubt the other editors consider a tag as void if the VALUE is empty
• For boolean tags (e.g. "oneway"), I consider an empty value could be as valid than the other possibilities (Yes/true/1/...), i.e. la "#define"

Author: Robx
[Added to the original trac issue at 2.54pm, Thursday, 13th March 2008]

I don't particularly mind if Merkaartor continues to allow empty values. I do believe consistent behaviour between the editors to be desirable.

Both Potlatch and JOSM definitely consider an empty tag value as meaning the tag is to be deleted.

JOSM displays a downloaded empty tag fine, but doesn't allow creating it.

Potlatch doesn't even display it. Editing a way with such a tag (I added an extra tag and later deleted it) removes the empty tag also. Potlatch's revert (using 'H') doesn't recreate the empty tag either.

I'm not reporting this as a Potlatch bug, but go ahead if you wish.