[OSM-talk] Good practice, and should we rely on defaults?

[stevea]

This is a bit like shouting into an echo chamber or looking into a hall of mirrors, or some blend:  "guessing at what downstream uses 'thinks' are defaults" turns something not-quite-crisply-defined (the data we have today, which are good, but not perfect and not perfectly predictable) into "less crisp."  Let's not do that.  I don't think we can properly predict what all the downstream use cases will be as these already-fuzzy concepts have few places to go except for round-and-round-and-round discussion.

What looks like a good idea now turns into a bank of fog in no time.  Like pouring murkiness into mostly clear water.

The data we have today are what they are.  If we wish to improve them (we do), a cleave along a boundary of "assumed defaults" or "defaults except with this/these exception(s)" and that smeary edge of syntax extraction becomes its own distraction, it isn't a good idea.  There are many other ways we can improve our data, "guessing more" about them (which is what this would quickly become, and with squabbling almost endless about what and how...) doesn't seem like a very productive one of them.

I'd start with an "assume defaults of today in the real world" and apply those assumptions to the data we have.  Without what would or will no doubt become stale 'default' data in no time flat.  This endeavor seems an endless sinkhole of lose.  I believe not everybody will not come to this conclusion.  BTW, this is what we now do.

Personally, I do not "guess at defaults" as I enter data into OSM.  I don't plan on starting anytime soon.  I enter what I know and stop there.