[Tagging] remove “rendering” from proposal template

stevea steveaOSM at softworkers.com
Sat Apr 24 18:33:35 UTC 2021

On Apr 24, 2021, at 11:18 AM, Brian M. Sperlongano <zelonewolf at gmail.com> wrote:
> Suppose that someone proposes that tag "B" (which currently does not render) should be a replacement for tag "A" (which currently DOES render).  If tag "B" were approved, and suddenly people started replacing tag "A" with tag "B", we would suddenly see that feature start disappearing from maps (until some future time that the renderer adds support for tag "B").  This is the kind of thing I'm talking about and it requires no sample rendering, just an impact statement and/or transition plan.  I would consider a proposal incomplete (bordering on dishonest) if it would remove a currently rendering feature without clearly describing and assessing that impact!

OK, this (tag replacement) is a critical topic to address, and differs from the specifics of whether we keep "Rendering" as a section, modifying it as "Rendering Suggestions" (and saying "this is all optional, don't let this affect your vote").  With tag replacement, we DO flip "optional" to "required."  So, yes, this is an exception to all I say and must be accommodated in how we proceed.

> Further if someone proposes to expand the definition of tag "A" to include additional features not currently tagged with tag "A", it could similarly cause undesired behavior to downstream consumers/renderers that are depending on the narrower definition of tag "A".

Again, this is both true and critical, but it is more related to the topic of one syntactic representation replacing another, rather than the instant case of making the topic of rendering "suggested / optional."  Though, as I (and you, and others, especially renderer authors) think about this, once again we get into the realm of these decisions (of "hm, there is some replacement going on...") being squarely in the domain of renderer authors, rather than those specifying syntax in a tag proposal.

The "someone propos(ing) to expand A to include additional features" must do so in the realm of syntax / tagging.  The "how does a renderer cope with those additional features and choose to render them" must happen in the realm of renderer authors.  I know I'm hammering this point hard, but there is tagging (syntax;  grammatical utterances in OSM's language of tags on data structures) and there is rendering (the semantic interpretation here expressed not as "an IDEA conveyed by a grammatical sentence," but rather "IMAGES and GRAPHICS conveyed by a grammatical sentence."  The entire "how" that happens, the actual rendering, happens as the "grammar" is handed off to the "parser."  Syntax and semantics are related, they are coupled, but they are also separate.

More information about the Tagging mailing list