[Tagging] Deprecating phone=* in favor of more ambiguity in the database

[Robin Burek]

> no it doesn't. We should only add semantic helper words if it is required to prevent misconceptions. In this case, there are no competing other uses or even imaginable uses of "phone". It is more elegant and nicer to type if the tags are short, and do not start with a common prefix. These prefixes break widespread editing habits, in particular they tremendously reduce the benefit of autocompletion.
> On the other hand, if you are using a form (presets etc.) to enter the data, you are not bound to alphabetic sorting anyway.

I would rather be on the second point. The key itself is in most cases a minor matter, as such content is generated by presets. Thus, a semantic compilation makes even more sense. If a new communication path is now discovered, one can even let go of a preset only on (in this case) contact:* and the preset would even be able to display contact:new_key. The same applies to all other data users in general. When I am in a namespace, I immediately know what this object belongs to. Even if the exact key may be unknown to me.
Especially when setting up a tagging, one must not only consider the side of the person who enters the data (which is negligible to a large extent today due to presets) and on the other side those who process the tags in some way. The initiator of this thread belongs primarily to the second group.
Translated with www.DeepL.com/Translator (free version)