This page presently holds a letter to Poikilos.

Letter to Poikilos

On 01/07/2019 2:01pm, Poikilos wrote:

Is there anything I should know about why 3d_armor doesn't work, or does it work, but without correct visuals?

1. Yes.

To respond completely, I need to mention some non-technical issues.

It's apparently easy to understand the 3D Armor problem, if one has time to focus on it. However, I gather that only one person can explain it presently: Jordach.

Jordach could fix the problem, but he's declined to do so. The issue seems to be stress related the events of Spring 2017. He threatened suicide later in the year and told me at the time that he was en route to the hospital.

However, he's theoretically still involved in Old MT. This assertion is based on an announcement in Fall 2018 to the effect that he was going to appear in person at a U.K. MT event.

So, perhaps you'd be able to get him to chat with you in IRC.

On 01/07/2019 2:01pm, Poikilos wrote:

Since you have included a mobs mod, I must mention: I am also interested in testing my fork of mobs_sky with the new release of Minetest and adding more birds. I'm also really interested in adding better flight paths and a landing animation.

2. You went on to mention bugs related to your mod.

The mod could be added to CoderMobs. But it probably would have the same issues that it does now. They'd be in Mobs Redo, in your code, or both. There is no reason, however, not to try it. If you're able to identify changes needed in Mobs Redo, I'll make them in my copy.

On 01/07/2019 2:01pm, Poikilos wrote:

The new "Features" page should probably offer specifics to clarify the following:

* can Pet Ores be spawned naturally (probably should not for multiplayer)

* how armor class is used to calculate damage, or at least say name of source file that calculates it

* how will lesser-used homedecor items be removed/split (will they be in a separate modpack/mod? suggested directory structure: morehomedecor/* such as morehomedecor and morehomedecor, morehomedecor and morebuilding_blocks)

3. The Features page is just an overview. We might condense it further to make it more accessible to the casual reader. Or perhaps it's at the right level of detail presently.

What you're interested in is the manual, which is separate. It should be ready for the release in 3 weeks, but may not be complete.

The manual is an expanded copy of the Features page that goes further into settings and the types of things that you've listed. We'll add anything that's missing over time.

* Pet Ore spawning is, or will be, controllable by config settings. This is on the TODO list.
* I'll put armor class questions on the TODO list.
* Input on what to do with Home Decor (or our Decor Pack derivative) will be welcome.

I should add that our copy of the Lua API document will need work. We had to hack our way through 100s of commits to make Final MT. I'll need to check everything in the Lua API document over time and start to get it into sync with the code.

On 01/07/2019 2:01pm, Poikilos wrote:

I like all of your changes, and consider them critical — including:

* picking a "standard" mobs framework and expanding it to include more behaviors (I am interested in seeing a log sometime of all changes; I always wanted to base my kc_modpack on mobs redo, but I was concerned about limitations and was considering jordan4ibanez' openai)

* including a mod with sea mobs

* including a mod with generated structures (shipwrecks from CoderSea mod in this case)

5. We'll have generated structures in multiple contexts. I've been doing this in my worlds for years.

On 01/07/2019 2:01pm, Poikilos wrote:

Since Microsoft owns GitHub, some people are concerned about current or future Terms of Service — I'm not sure if you are, but if so, you might consider where some minetest mods have moved: GitLab (homedecor) or NotABug (mobs redo & other TenPlus1 mods).

6. Comment in more detail on possible issues. Soon, if possible.

The main reason that we'd go to GitHub is that it's expected. git-based is a must, but I could host a git repo myself using Gitea. Discuss your thoughts on the pros and cons of GitHub, NotABug, and/or Gitea.

On 01/07/2019 2:01pm, Poikilos wrote:

I highly suggest adding an API function that gets the player model version AND player model type (mod name? some way to identify beyond numeric — may not be needed if version is a string that includes type and version, such as "minetest.org-4" or "minetest.org-20170604". For why I mentioned this date, see README.md for playeranim at link below.

You might not be planning on changing things like the `y` of the origin of the model like minetest.net has, but in the case of their dev version, not having any clear way to know has been a headache.

Modders had to go into the commit history to find out, and make users add a variable to tell the mod what version was being used.

7. Certainly. Not this month, but I'm adding it to the list.

On 01/07/2019 2:01pm, Poikilos wrote:

By this many years into development, with a knowledgeable community of modders, I would have hoped that a previous release of Minetest would have interwoven and optimized major mods such as mobs redo, sea, and moretrees — I have always tried to ignore that and some other shortcomings, and I'm very glad you've corrected this and other shortcomings mentioned.

I really look forward to trying the new release. Though I am usually mostly exciting with modding Minetest, I am actually really excited to play this one!

8. If it crashes for you initially, give us a chance to fix it.

Additionally, understand that some major features are still on the TODO list. The main example that comes to mind is natural spawning of More Trees.

Presently, in that group of trees, only Rubber Trees spawn naturally. They came first because they're related to Technic, which is now part of the standard game.

Or, rather, a light version of Technic is standard now. Slopsbucket can provide details.

Slopsbucket has been working with snapshots extensively. But both he and I know what to do with it. Others will need to rely on documentation. Which will still have gaps in 3 weeks.

I could provide you with a Windows package perhaps 1 or 2 weeks sooner if you like. I forget; can you run Windows and, if so, 32-bit or 64-bit? The package won't support your maps, due to the need for new "games", but it should be fine for review and debug purposes.

On 01/07/2019 2:01pm, Poikilos wrote:

I also look forward to focusing on modding to add more features instead of spending half the time fixing stuff and uploading bug reports like I was (mostly regarding issues related to changes in the backend in the [upstream] git version).

Excellent work.

9. We think that our engine and standard "game" are close to stable and that they provide a solid foundation for world host and modder purposes. Our CoderEdit (a drop-in replacement for WorldEdit) is especially useful for people who need to do serious work.

Bugs still show up periodically. We fix them to the extent possible. Slopsbucket is skilled at identifying problem commits and that helps.

We feel that it's past time to focus on stabilizing the core and to give use of the framework to develop mods and worlds the respect that it has been due from the start.

10. People consider "issues" between the core devs and myself to be not relevant. But they go to the heart of what I've just said.

A Junior Troll visited me recently in IRC. I think that this one was only 12 years old. He or she said:

"You're not listed on the Contributors list. Look, your name isn't there, so you didn't do anything and you're not important."

Then he or she ran away without allowing me to respond.

I've been told that this is the retcon of history that "Shara" is pushing now in the back alleys of the old project. Didn't you yourself quote a similar statement to me?

So all of the world hosts, modders, IRC admins, forum admins, and people who helped others for years are disposable garbage?

The MT core group has always thought like this. So, it was doomed from the start. Middle-school "kool kids klub". Not a real project.

It's time to declare that the mess which was the old project has ended.

It's true, you know, for reasons that have been explained previously. Nobody with an IRL identity can be associated long-term with the old project without risks related to the prosecutable crimes that were committed.

Regardless, a mismanaged and hostile "kids klub" project run by anonymous Ninja Kittens, Sexy Fox Girls, and trolls and DDoSers in Halloween costumes of other types isn't to be taken seriously.