lee/hpr-knowledge-base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit: HPR Knowledge Base MCP Server

Browse Source 
- MCP server with stdio transport for local use
- Search episodes, transcripts, hosts, and series
- 4,511 episodes with metadata and transcripts
- Data loader with in-memory JSON storage

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Lee Hanken
2025-10-26 10:54:13 +00:00
commit 7c8efd2228
4494 changed files with 1705541 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

281
hpr_transcripts/hpr2797.txt Normal file
View File

@@ -0,0 +1,281 @@
Episode: 2797
Title: HPR2797: Writing Web Game in Haskell - Simulation at high level
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr2797/hpr2797.mp3
Transcribed: 2025-10-19 16:58:51
---
This in HP are episode 2797 entitled, Writing Web Game in Hackel, Simulation at High Level.
It is hosted by Tuku Toroto and in about 26 minutes long, and Karina Clean Flag.
The summary is, Tuku Toroto gives over new on simulation in their 4x game.
This episode of HP R is brought to you by archive.org.
Support universal access to all knowledge, by heading over to archive.org, forward slash, donate.
Hello, this is Hackel Topic Radio, and I'm Tuku Toroto.
Today I'm going to talk about the Writing Web Game in Hackel, and especially the Simulation
at High Level.
So far, we have been concentrating on separate pieces, so now it's time to pull those pieces
together and have something more, more concrete going on.
So when I was thinking how to build the game, and think in all kinds of different simulations,
I, in the end, I often do the system, the simulation is done in steps, and they are taking
every step is one month, more or less, one or more or less.
The time is in decimal system, so for example, current time now will be 2019 or more three,
and that's stored in the data phase as a 2020, 193, because interquests are much more
nicer than floats when you are adding continuously to them.
It floats you at some point start to get rounding errors and all kinds of surprises.
So in the end, I want that the turn resolution happens every 24 hours, a real time, automatically,
but that part I haven't been, I haven't yet figured out how to do, so currently I'm just
triggering manually.
So there's an interface where I press a button at that point, this process that I'm going
to describe next, start, and it simulates what happens during one decimal month in the game world.
So main processing is, well, this has couple parts that I haven't run through those yet,
and I'll probably explain those later, and I'm not going to read out the time signature
of the function, it's rather long, but it's in the show notes if you're interested on seeing it.
Any idea is that it's a function that has access to the data phase, it can both read and write
there, and it returns a time in the end, it tells what's the time in the universe
after the processing has been done.
And I found it kind of interesting that in bed systems, the libraries that I'm using
for the data phase access, you have very granular control over what kind of things your functions can do.
So if you have some process that you know that it only reads from the database,
you can tell that in the database that this function only does reading, and if somebody tries
to add at some point in the future, in that function or in any function that it calls,
it could be a very complicated process that takes really long to run, and it has many parts
and multiple reads. If somebody were to add something there that tries to write into the data phase,
you would have a compilation error, so you have to tell in bed systems if the function or any
function that it calls has read or write access or both into the database.
Okay, so the main idea is that there's a process-turn function that takes no parameters,
and it's a, well, I'm reading it out a little bit.
So it starts a process-turn equals two, so it's a monotic code.
And the new time arrow to the left at the end of the time.
Here we are calling advanced time function that updates the time in the database for one month forward.
Then Android turns it, so now that's in the new time variable.
Andersko arrow left, remove, export status is now time.
So here we are, if you remember previous episode, we are going to talk about planetary
statuses that planet can have a good harvest, bad harvest, they might be some monster
attacking the, or things like that, this part goes through all the, all the
statuses in the database and removes the those that are marked to be removed
when the time is what we just passed in, or later.
And the Andersko arrow left means that whatever removed expect statuses returns via
discarding. We could call this without that Andersko arrow to the left,
but then we would be getting a compilation warning.
And I like to keep the code so clean that there's no warnings at all.
Because while the code might work in some cases with those warnings, there might be some
hard cases where it doesn't work. And if you have lots of warnings, you might,
lots of warnings that aren't important to you. And then there's one important,
you might miss that if there's lots of warnings and you have all, and you have gotten
used to the fact that there's a 500 warnings, so one more, you don't spot it.
So I keep the code as clean as possible.
So then statuses have been removed. Then these factions are all to the left,
select list, open-fucked, close-fucked, open-fucked, ask, as a faction ID, close-fucked.
Here, select list is a persistent method. Here we are loading from the faction table,
we are loading all the factions that are there, and storing them by the IDs,
and storing them into the factions variable.
And now the factions variable is a list of faction entities, and entity is a data type
that has a primary key and the actual data of what you found from the database.
So now we have a list of those. And onwards, underscore,
are all to the left, map M, open-parent-handle-faction-events, new-time-close-currents-factions.
Again, discarding, let's see, receiving from this function, we are not interested in doing that.
And then we are, map M is a, it's a, again, it's kind of like a map or a map for
that we have been talking previously, but little, that capital M in the end is a hint
that it's used in the monadic context. So now that we are doing a database access,
we cannot use the map, we use the map M. I'll try to explain them briefly
on list-plating this episode. So what we are doing is that we are calling
that handle-faction-events, and giving that to our client time,
and we are calling that function to each of the factions that we loaded previously.
So we are doing sort of like a loop here. So this, this will handle special event
as a resolution that I talk, talk previously. So if these are, for example,
a cracky worm attack on a planet and player has chosen what they want to do, or hand chosen.
This part is, this part triggers the handling for that.
Onward, then there's a map M on the score, handle-faction-food-factions.
This handle-faction-food doesn't return anything sensible if returns a unit.
So we don't have to, we don't have anything, anything that we would need to discard.
So we are not using that underscore arrow to the left, but we are using a source hand
map M on the score, that's the same thing. There's just so many of these things that
it's hard to keep track of, track of these sometimes. So handle-faction-food,
here we are, again we are calling this function to each and every of the factions
in turn. So this one will take care of the food production and food consumption.
I haven't gone through through what actually happened inside of this function,
in previous episodes, maybe I will do that at some point.
It's a kind quite simple, the current isn't that much interesting,
but if an event is more interesting stuff happening in there, I will call through that.
But in any case, this will produce food in a planet, and then the population on the planet
will consume that food, and what food is left over or biological resources,
as they are called in the game, are stored in the stores, and if your population
is consuming more food, then the produce and then they are using the food from the stores.
Okay, then map M underscore open-parent handle-faction-construction,
new-time close-parent factions. Again, we are calling the handle-faction-construction
with new times and each of the factions, and this will take care of the various
construction products. This I haven't gone through previously, I will do that.
I have to do some work there still before I can explain the idea and what I'm doing there.
But basically, you have construction projects on various places,
and this is where those construction products are.
Progressing and maybe completing, and then these are generated.
They're usually. Then, underscore arrow to the left map M,
open-parent access-al-events, new-time close-parent factions.
So here we are calling access-al-events, new-time,
to with each and every of the factions, and this is the part that adds a new
special, possible new special events. I'm sure if I have gone through this part in
detail, I might or might not have. But this is the place where new special events
can happen randomly. There's a bit list of special events and then you,
and then they have a randomized, then they have some
representatives of chance occurring and based on, and then they have some
preconditions and based on that, all that information, each faction may or may
not get a special event at any, to some place.
And the final step is map M, underscore open-parent, handle-faction observations,
new-time close-parent factions. Handle-faction observations,
I have gone through, I believe, this is where you, well,
all kinds of reports are generated for you. If you have a population
on a planet, they will write a report to you that this is what we see.
If you have a space ship somewhere, they will write a report to these
what we see. If you have a probe somewhere, they will write a report to these
what we will see. All that information is stored into the database and
whenever information about something is looked upon, like you want to know
what you know, for example, some planet, then all those reports that have been
written for that planet are retrieved and they are combined and that's
your current knowledge of the place. Some of that information might be
mapped to date, some of that information might be obsolete,
some of that information might be downright wrong, and some of that information
might be completely lacking. So you don't have a defect you
of the universe around, only the what you actually have seen and observed,
and there might be some information that is out of date.
Because you haven't been, you haven't been at that place with information.
It's like if you make an expedition to some faraway place and see that,
okay, here's a rocky planet, nothing to see here. Come back.
Then that's the information that you will be shown when you check the
status of the planet. There might be a huge civilization there,
but you don't know because you haven't seen that.
And the final line on the code, that processor function is return new time.
So here we are returning the new time, the time,
current time of the universe after the processing. And why are we using it
return here when we haven't been using it pretty much anywhere else?
And that's the answer is that the processor is a monadic function.
In this case, it deals with input and output,
specifically the database. So there's a bit of extra work that you have to do
to be allowed to do that. In Haskell functions are used here.
So they work in a way that when given same set of parameters,
they always return the same answer and there's no dependencies to outside systems.
So I think another term is referentially transparent.
You could replace function with a huge lookup table that would have all the
combination of the parameters and answers to us.
So there's no, the outside work doesn't affect them,
or some are another part of the system doesn't affect the answer.
The data comes out and the data comes out is always the same
on the given set of inputs. It doesn't change.
But that of course does not work when you start dealing with the database.
For example, here you have a bit of extra instruction.
So if you remember math, that's a way of running a function over a list.
So you have that signature of the math is an open parent,
A, R, O, B, close parent, R, O, list A, R, O, list B.
So when given a function and a list, it will produce you a new list.
And the elements of that new list are what you call when you run that function
to elements of the old list.
FMAT is a similar concept.
It's a type signature is an open parent, A, R, O, B,
close parent, R, O, F, A, R, O, F, B.
Again, same concept, but instead of working on a list,
it works on F and F can be some structure.
It could be a list, it could be a maybe, it could be a either.
So FMAT works for sanctos that I talk about earlier.
So MAPM is the same thing, but for the monot.
Each type signature is MAPM.
It's a monot M, that arrow, open parent, A, R, O,
M, D, close parent, R, O, D, A, R, O, M,
open parent, D, B, close parent.
So if you have this in a show notes and you look at them
and take couple steps backwards and screen really hard,
then you can see that they are very similar.
All three take a function and they take something
and then they produce something else.
And that works only with the list.
FMAT works with the functions.
It can have different kind of structure.
It's a more generalized.
And MAPM, you have that monot M are hanging around there.
So instead of function being from A to B, it's from A to M B.
So that B value is part of a structure M.
And M is the monot.
There are many kinds of monot.
It's a really fascinating topic.
It's not as difficult as you might have let believe.
Well, when somebody explains that to you in a really well,
it's a really simple concept.
But the problem here is that it's a fairly abstract.
And if you explain the abstract part,
you don't really get why they are important in the programming,
especially the Haskell.
But the way I learned then was that I didn't concentrate on the abstract part,
but I concentrate on what you can do with them,
what they actually do for you.
And here they are learned how to do it.
I'm going to try to explain them at a bit later episode.
But I need to come up with a really good explanation.
But for now, it's enough to understand that the map M is similar to map and F map,
but it also gives, it also deals with the input output in this situation.
Monats are not only for the input and out, they are highly versatile.
You can use them for many things.
And they are sort of the next step in the chain.
Let's start with the function and continue with the applicative that I talked earlier.
But for now, it's enough to know that in this case they are there for dealing with the input and out.
Okay, so, talk to the processor.
Each step is done for all of the factions before moving to the next step.
First, the status removal are done for everything in one go,
but for example, the special element resolution is done for all factions.
And then the food production and consumption is done for all factions.
And so on.
Originally, I handled each faction at a time.
Like, did all these things to one faction and then process another faction and then did the next.
But then I started thinking that it...
Well, it started feeling that these are discontinuation in the universe.
Like, one part of the universe is one month ahead of another part of the universe.
And if there's any interactions between those...
Those universes and those parts of the universe,
the later point in the simulation, it looks funny and...
So, I had it originally working like that, and then I thought that it might be more sensible to run it in a slices.
So, parts of the universe update.
So, like, the whole universe updates in the terms of the special events and then all of the universe updates
in terms of the food products and consumption and so on.
So, then it's still done one place at a time, but it's not so...
It's not so big leap in a long go.
That's the thing with the computer simulation.
You are usually dealing with a discrete time.
Like, you have a step that you are doing it.
You could do continuous simulations, I think, with some fancy mathematics.
But that could be way beyond of my skills.
It would be interesting exercise, but I suspect that that would be way beyond my skills and that might be...
pretty taxing for the computer to do.
Okay, so, after all this, the time the universe has moved one month, month, month, month, month, month,
and players can review reports, special events, tattoos, and whatnot.
They can browse around with the interface and do what they want to do.
And, like I said, this will run when I track on figures times when I get that written.
And that will cost me to figure out how to do that.
So, even if players have not made the choices, then this will run.
It just means that they didn't feel or didn't have time to make those choices.
The world keeps rolling even when somebody doesn't do anything.
And, in the end, I want to add a little bit of automation to the game.
So, you could have instructions.
You could give us some cover on the planet.
Instructions that I want you to cover on this planet in some specific way.
Like, I want you to take the planet to this direction, or that direction, or whatever.
And, here's your resources.
And, just go wild.
And, after that, if the player will not have to micromanage that planet by themselves.
Of course, there might be a situation where some player is actually playing as a governor of a planet.
And, then, the whole point of the game is to take the free sources that they have.
And, to do with the planet what they want to do, or what borders they have been given by somebody else,
there might be a vessel of some star or octopus who governs them.
Who has a domain of multiple planets and has delegated the authority to some, some, some governors.
And, that star lord, could be a dual player, or could be a computer player.
But, in any case, the planetary governor would be a player, and they would be playing and managing that single planet.
And, maybe they would be very ambitious players, and they would secretly start playing a political game,
and maybe a military game, and try to declare independence and start their own empire that way.
Okay, I think that's enough for now.
And, the best thing nowadays is to catch me, is either the e-mails, or a steady verse where I'm to do it and master them, but social.
Okay, Ad Astra.
You've been listening to Hacker Public Radio at Hacker Public Radio.org.
We are a community podcast network that releases shows every weekday, Monday through Friday.
Today's show, like all our shows, was contributed by an HPR listener like yourself.
If you ever thought of recording a podcast, then click on our contribute link to find out how easy it really is.
Hacker Public Radio was founded by the digital dog pound and the Infonomicon Computer Club, and is part of the binary revolution at binrev.com.
If you have comments on today's show, please email the host directly.
Leave a comment on the website, or record a follow-up episode yourself.
Unless otherwise stated, today's show is released on the creative comments,
and the contribution, share a light, three-dot-oh license.