This is the second time a question of mine has been placed On Hold for supposedly containing multiple questions. This is a link to the current version of How to generate metatiles for use with tapalcatl but here's what it looked like at the time it was placed On Hold (including the bold text at the bottom):
I'm investigating using tapalcatl-py to serve vector tiles on AWS Lambda infrastructure. Unlike other tile servers, it does not seem to use .mbtiles files directly, instead using "metatiles".
The original, Go-based, project describes the metatile process like this:
- The client requests a tile /1/2/3.json
- Tapalcatl makes an upstream request for /1/0/0.zip
- Tapalcatl extracts the file 0/2/3.json from the package.
- The client reads back a tile containing only the format for the specific tile they asked for.
But that raises a lot of questions for me:
Why is the client requesting a tile in .json format (as opposed to .pbf, or even .geojson)?
- Who is Tapalcatl making the "upstream request" to?
- What is the structure of the .zip file?
- How did 1/2/3 get mapped on to 1/0/0? (Is it always /0/0?)
- How did 1/2/3 get mapped to 0/2/3? (Is it always 0/x/y?)
- Does "the client reads back" mean "the server sends back"?
- Does "...containing only the format for the specific tile they asked for" imply that there were other formats within the same zip file, that are discarded after extracting?
In short: how do you turn a regular old .pbf-based .mbtiles file into something that tapalcatl can serve?
How does closing questions like this as "too broad" add any value to the site? I'm clearly asking one very specific question here ("how do you generate the required input directory") and the other questions are just explanations for my confusion, showing evidence of my thought process.
I could have equally worded them as "I don't understand X, I don't understand Y...", which would change absolutely nothing. Or I could have not included them at all - but how does that make the question any better?
The "no multiple question" rule is there to avoid multiple unrelated questions, which are better answered individually. In this case, that makes no sense - none of those questions are better asked separately.
I use a lot of StackExchange sites, and I have never had questions closed for this reason on any other site. It seems pedantic and just plain annoying.
Ok, since I posted this:
- I updated my post here to remove the question marks, and to restate the additional questions I had as information I was interested in, but explicitly not requesting that answersers provide it.
- PolyGeo edited my post to remove it all.
- I rolled back that edit as I didn't agree with it, and PolyGeo has indicated ("you can rollback what I see as improvements to it, if you disagree that they have been made to make GIS SE better") elsewhere that rolling back their edits is ok.
- Now PolyGeo has locked my post.
How is any of this in the interests of GIS.SE, or you know, letting people actually answer questions that are posted? This kind of nitpicking and heavy-handed moderator intervention does not feel good :/
At the time of this update, there are 8 upvotes on this question, and 1 upvote on PolyGeo's answer below. By normal Meta rules, that indicates there is significant agreemement with my view that this "one question per question" rule is being applied in an unhelpful manner.
Please? Can we just stop doing that, and make GIS a nice, friendly place to post?
PolyGeo has rolled back my post again. Why? How could it possibly be so important that it's worth inflaming this situation?
It's so frustrating. In my original question, I was looking for both validation that I what I was doing made sense, and if it did, then how to proceed. Due to the edits, presumably the answerer didn't see the request for validation, so only addresses the tools bit.
End result: the answer is less useful to me, and probably less useful to future users of the site.