Tilemill Modernization

[xycarto]

Perhaps this is naive question, but I am curious to know what it would take to keep Tilemill alive for the next several years?

It seems no matter how hard the geospatial world tries to kill it, raster tiling still plays an important role to web mapping and geospatial in general. The Tilemill project still receives issues on a regular basis, indicating there is still a world out there relying on this tool for editing and basic rendering. For those who work with styling, blending, and serving up raster data in base map formats, this tool is still the go-to for many professionals. Further, it seems there is no lack of raster tile servers out there; however, there are very few tools to develop or style the raster tiles to put on these servers.

So my questions are this:

1. Is there truly a solid interest in keeping Tilemill alive in the geospatial community?
2. What would it take to revive this project again into an actively maintained tool? Something that could be relied upon in the future.
3. Are there any developers with the time and resources left able to maintain this tool?

Open source does not mean free labour, so if there was an interest in modernizing Tilemill what would it take?

1. Is it merely a Kickstarter campaign rising funds so solve the most pressing issues?
2. Does the project need legitimate sponsorship with a dedicated developer?
3. If the project could get funding, is there anyone capable of doing the work?

Please know, this is not to discredit all the hard work from contributors that has gone into keeping this project alive. I am forever grateful to the community still out there doing their best to keep Tilemill up and running through this project. These questions are more-so trying to determine the possibility/feasibility/reliability of this projects future and how the geospatial community can contribute to maintaining a tool they still find useful.

[csytsma]

You've asked some great questions, which Tim and myself have also asked, and spurred us to take over Tilemill and try to keep it alive several years ago. While we've been able to keep it working, and expanding its features for our own use, we've been disappointed that no one else has come forward to also help maintain it. That has led to the assumption that there aren't many people that are still using it. We have learned enough along the way to be effective, but the project could greatly benefit from experienced software developers to take it to the next level.

The main challenge for most people is the complexity of setting up a working instance. We've created installation scripts to aid in that, but I think it's still beyond the skills of most casual mappers. The previous native apps of Tilemill greatly simplified this, but still required a separate Postgres instance if you wanted to use its full features.

We considered setting up a hosted multi-tenant instance of Tilemill that people could subscribe to and have access to, in which the full environment is also setup for them, including Postgres. But this does remove the "opensource" and free component, and gets you right back to Mapbox.

I like your idea of finding alternate ways of getting paid labor to maintain it, through Kickstarter or a sponsor. But I have no idea if there's enough interest out there for that. The Github project appears to still get about 20-30 unique views a day, and about 10 clones a day. Something happened a couple days ago (April 15-16), that caused about 50 clones per day for those 2 days. So there is some interest. It's hard to know whether the individuals who cloned the project ended up successfully getting it setup though.

I'm very interested in continuing this conversation, and seeing if we can find a way to keep Tilemill relevant.

Is there anyone else out there who has an interest in helping to contribute to Tilemill to maintain and grow it?

[stevage]

Curious what is actually involved in maintaining it? Is it a question of keeping up to date with new MacOS/Windows/whatever versions, just stopping the thing breaking?

I don't think I'm putting my hand up here, but I'm an interested spectator. I really don't actually use raster tiles any more, but I have very fond memories of them, and there are things that were possible with Tilemill that still really aren't possible with Mapbox GL for instance.

[csytsma]

The core aspect of maintaining I'd say is keeping the libraries and code up to date (including addressing security issues), fixing bugs, updating documentation, etc. This ensures that it continues to function and is usable.

The second part of maintaining I'd say is adding new features that are identified as needed, so that the product stays relevant. This helps keep existing users, and attract new users. If the users disappear, the product will also die.

[xycarto]

Thank you for the response and insights. It is good to know there is some one on the other end.

I am not a developer, but I can offer my capacities in other facets, like research, how to get funding, and where that funding would go to. I put this issue up in the hopes the Tilemill project might know who those developers are.

In my opinion I think there might be four initial steps:

1. Getting Tilemill into a Docker in its most stable form. At least moving forward, there would be a version usable for the short term. Although I am not a developer, I could probably develop this for Ubuntu and get it in the Docker universe.
2. Determine if there is a need to update the project. I think Github is a good place to get started in determining who is still cloning the project, but there likely a need to reach out to the wider community to see who is still actually using the project. I have limited reach in the geospatial community, so something like this might the effort of more than one person.
3. Determining what those core aspects are according to your suggestions.
4. Determining if there is even a developer out there capable of taking on the project and, if so, how much funding is needed.

I am only one test case. I am a Tilemill user. I work a lot with raster data and blending that data with other bits into raster tile sets. I often use a combination of raster and vector tiles. For me personally, the use of Tilemill is just an editor and producer of the MapnikXML document. Beyond its raster data handling, Tilemill is the only raster tiling tool I can use in custom projections. My needs are selfish. I want it to run smoothly in Ubuntu, have an easier time setting projections, and to work relatively quickly in refresh when I make changes in the editor.

That said, I am enough of a user that I need to start giving back to the project where I am capable. In things like documentation, that would be a labor of love for me, but the hook seems to be determining if the project is still relevant and finding the developer.

[csytsma]

@xycarto In response to your suggestion 4 initial steps, which I think is a good place to start:

1. Getting Docker setup will definitely make it easier for more technical people. I don't think this makes it any easier for the casual mapper. I'm not able to help with the Docker creation, but there seems to already be a few forks that have setup Docker. Can we get someone to take this on? Create Docker for Tilemill releases #2742
2. There's a continual need to update the project, just to update libraries, fix security vulnerabilities, etc. I may take another whack at this in the coming months, to get it updated again.
3. Perhaps we setup a survey, and try to get as many active users to provide feedback, and see what the community looks like. I can help create the questions, but would love help on this. Survey Active Users on Usage & Features #2743
4. I think our best bet is to find an active user who is also a developer. Not sure we could get the funding to pay full rates. Can We Get Funding for Tilemill? #2744

My primary use is to create rendered, static maps for Adventure Races that I put on as a hobby, and to create MBTiles that I use when going out on my own outdoor adventures, and want off-line maps. I've created a fair amount of static map features to aid with creating paper-sized maps, in a separate branch, that I haven't merged back into Main.

I'd suggest we create new Issues for each of the 4 steps above, to aid in discussion and track them. I'll create them now.

[SaidAlSouti]

I love this project. I used it on daily basis.
Thanks guys for you awesome tool.