[CJA]

thaaaanks!

I'm looking for some criticism for these. I've heard that it needs to be more consistent re: chiptunes and non-chiptunes, and I'll probably end up replacing the dungeon and town themes with something more consistent and less chippy, probably recycling the motifs from Catacombs or Steam Zone... and making all of the battle music more chippy.

[CJA]

So, a source of inspiration came from an unexpected place: The ambiguity of my battle engine was frustrating me, so I decided to do something that directly opposes my usual strategy. I took my battle engine out of MZX and and gave it a home in a text file, and then I commented every line.

Every.

Line.

(Well, not every line, but there IS about one comment per command now.)

I've found that not only did this fix a few wriggling bugs, but it made item drops work again and abstracted units so that I can add a party switcher later instead of one static party for the whole game. Woo!

This was brought on by me absolutely slamming down some story lines into the story planner. Creating a character is such a nice way to move things forward. Also, story interpolation is really good. If you know where your story starts and ends, put those as two lines in a document and just keep inserting new events that happen between any two adjacent events until you get your groove on.

[Dr Lancer-X]

I find battle engines get big quickly. I have more trouble with the 64k per-robot limit with battle engines than with anything else, and that's with designing the engine to use multiple robots from the beginning. Code that gets big can be annoying to work on outside MZX because you get no warning when your code gets too big and fails to load properly.

[Graham]

CJA, on 04 January 2017 - 04:36 PM, said:

So, a source of inspiration came from an unexpected place: The ambiguity of my battle engine was frustrating me, so I decided to do something that directly opposes my usual strategy. I took my battle engine out of MZX and and gave it a home in a text file, and then I commented every line.

Every.

Line.

(Well, not every line, but there IS about one comment per command now.)

I've found that not only did this fix a few wriggling bugs, but it made item drops work again and abstracted units so that I can add a party switcher later instead of one static party for the whole game. Woo!

This was brought on by me absolutely slamming down some story lines into the story planner. Creating a character is such a nice way to move things forward. Also, story interpolation is really good. If you know where your story starts and ends, put those as two lines in a document and just keep inserting new events that happen between any two adjacent events until you get your groove on.

Story interpolation sounds like a neat suggestion to try. Thanks for sharing that.

What is the purpose of commenting every line, exactly? As a debugging tool?

[Dr Lancer-X]

I'd say the comments aren't even necessary, it's just the act of mentally stepping through all your code and reasoning about what it does that causes you to catch a lot of bugs you previously missed.

[Graham]

Dr Lancer-X, on 04 January 2017 - 08:23 PM, said:

I'd say the comments aren't even necessary, it's just the act of mentally stepping through all your code and reasoning about what it does that causes you to catch a lot of bugs you previously missed.

Oh gotcha. Yeah I could see that.

[CJA]

Yeah, essentially, that's what it did. But commenting will make things easier in the future, especially fixing bugs.

And... Thanks Lancer, I just noticed that I'm at 62K right now. Guess I'll separate the logic for moves and damage. Or honestly I'd just find it easier to run through the file and strip out all the comments before loading. fok.

[Dr Lancer-X]

CJA, on 06 January 2017 - 11:16 AM, said:

And... Thanks Lancer, I just noticed that I'm at 62K right now. Guess I'll separate the logic for moves and damage. Or honestly I'd just find it easier to run through the file and strip out all the comments before loading. fok.

Is this engine something that needs to grow as you add more stuff to the game? Because if so, the latter approach is only kicking the can down the road for a bit.

[KKairos]

My current project uses about 80 robots at a time that could be doing nothing, but they're each far, far away from the 64k limit all the time so far, and they're all prepared to load one of the so far 12 or so different sprite-driven object types via hub.txt. (I will probably never need all 80 objects at once on a single board, but I'd better be sure of that before I plan otherwise.) I've found piecemeal code is my friend for sanity and reset-able things like the "rooms" of the game and which enemies are in them. Maybe since your game doesn't (seem to?) need real-time you can have a separate bot calling .txt files as subroutines with a "main" routine doing the linking between the two? Or even one robot loading multiple .txt files in sequence, with some careful variable manipulation to get back to the right "spot" in the main routine? (Thinking about it for five seconds, the right string and "goto" could do it?)

[Dr Lancer-X]

KKairos, on 06 January 2017 - 06:07 PM, said:

Or even one robot loading multiple .txt files in sequence, with some careful variable manipulation to get back to the right "spot" in the main routine? (Thinking about it for five seconds, the right string and "goto" could do it?)

Interesting idea. I used this technique in Taoyarin and it worked surprisingly well there - all the enemies would load_bc multiple times per frame to load the different parts of their logic. I used this as a method of making it easier to reuse code between enemies, but it would be a fantastic way of reducing robot memory usage as well - I've never thought about using it for that. The only problem is that you really need to design the robot from the ground up to use this approach - I think converting over might be a little bit troublesome. Depends how compartmentalised your code is.

There's a performance concern, but given how much Taoyarin abused it and it didn't really cause any problems I wouldn't worry about that.

[KKairos]

Yeah, everything past the player engine was pretty modular from the start. Literally there's hub.txt and then *monster/objectname*.txt for all the others. Since I rewrote the structure of the file that tells the game what to put in each room, it's actually streamlined hub.txt a fair amount. And again in my case robot memory wasn't the main concern, just a side bonus. There is the benefit that the biped enemies will probably AI/move basically the same way, just with slightly different attacks, which could be a good general principle if CJA has a bunch of slightly different effects/displays that need to happen.

In CJA's case I suspect this would be most useful for an auxiliary robot loading different actions with a hub structure, and maybe the main robot(s) using a subroutine and then sending stuff off to aux-bot and waiting to be snapped out of their 'pause' routines.

Anyway before I turn this into a robotic talk thread, good luck CJA and I hope some of this is marginally helpful!

[Bramble]

CJA, on 04 January 2017 - 07:36 PM, said:

This was brought on by me absolutely slamming down some story lines into the story planner. Creating a character is such a nice way to move things forward. Also, story interpolation is really good. If you know where your story starts and ends, put those as two lines in a document and just keep inserting new events that happen between any two adjacent events until you get your groove on.

I love this, CJA! I agree with your assessment about characters. In a way, if the story is driven by a character, it's easy to see how events should progress. Further, I really like the idea of story interpolation. I feel like that's an evolution in gamemaking that I have never been able to accomplish. With your technical skills and creative ability, I bet you're churning some pretty worthwhile stuff.

[Graham]

KKairos, on 06 January 2017 - 12:07 AM, said:

Maybe since your game doesn't (seem to?) need real-time you can have a separate bot calling .txt files as subroutines with a "main" routine doing the linking between the two? Or even one robot loading multiple .txt files in sequence, with some careful variable manipulation to get back to the right "spot" in the main routine? (Thinking about it for five seconds, the right string and "goto" could do it?)

Funny enough, CJA wrote a tutorial on this some time back
http://www.digitalmz...=0

[Dr Lancer-X]

Graham, on 07 January 2017 - 12:31 PM, said:

Funny enough, CJA wrote a tutorial on this some time back
http://www.digitalmz...=0

Bit different. CJA was talking about programmatically pasting text files from multiple sources together, saving the result to a file, then loading it as a robot.

KKairos is talking about having the robot actually load different text files' worth of stuff to do.

e.g.

code.txt

. "... do a bunch of stuff ..."
set "code2.txt" "load_robot"

code2.txt

. "... do a bunch more stuff ..."
wait 1
. "Loop back to the original code"
set "code.txt" "load_robot"

The thing about the second approach is that you can use this to essentially have one robot with an unlimited robot memory capacity, if it's all laid out carefully.

[Graham]

Dr Lancer-X, on 06 January 2017 - 08:57 PM, said:

Bit different. CJA was talking about programmatically pasting text files from multiple sources together, saving the result to a file, then loading it as a robot.

KKairos is talking about having the robot actually load different text files' worth of stuff to do.

e.g.

code.txt

. "... do a bunch of stuff ..."
set "code2.txt" "load_robot"

code2.txt

. "... do a bunch more stuff ..."
wait 1
. "Loop back to the original code"
set "code.txt" "load_robot"

The thing about the second approach is that you can use this to essentially have one robot with an unlimited robot memory capacity, if it's all laid out carefully.

Ah I see. I've actually used both of those approaches in my game with splendid results.

[CJA]

I'm probably going to statically split the main battle engine into pieces. Move execution and damage code is 23K and is the only part that will really grow substantially from here, and the main logic is 36K (which won't move much -- maybe 20 lines for a new status condition).

I do like knowing what's going on with the back ends of your projects, though. Interesting Stuff!

EDIT: No, seriously, the comments are massive. I have 20K of comments in here. Holy crap.

This post has been edited by CJA: 08 January 2017 - 01:16 AM

[CJA]

Oh yeah, here, have a crappy CSV reader. It was designed for multiple files and minimal babysitting, just plug in the file list and don't worry about anything else.

Spoiler

. "csv_reader"
. "=========="
. "Loads a CSV file and parses it into a two-dimensonal array of
. "counters, in this format:"
. "  &file name&&row name&&column name& = &value&"

. "Starting with $csv0, list your .csv file paths and count."
set "$csv0" "csv/mycsv.csv"
set "csv_file_count" 1

. "If this string is not empty, it be used as the name of a 'virtual'
. "column containing the line number.  This gives you a unique ID."
set "$line_number_column" ""

. "This represents the symbol that will be used as the lookup operator."
. "  String cells that begin with this symbol will be treated as a"
. "  reference to a MZX counter."
. "i.e., if a cell contains *HEALTH, that cell will contain the value"
. "  of the HEALTH counter when the CSV is processed."
set "cast_to_counter" 0

. "Set seperator between cells.  Use 44 for comma, 9 for tabs."
set "seperator" 44


. "============================ INITIALIZE ============================"

. "Init"
set "this_csv-file" 0

. "Letsa go!"
: "readNextCSV"
set "&$csv('this_csv-file')&" "FREAD_OPEN"

. "Check that the file exists."
set "fread_pos" -1
set "$catch" "File does not exist: &$csv_file('this_csv-file')&"
if "fread_pos" = -1 "throwError"

. "File's open.  File cursor (a.k.a. fread_pos) is now at the end"
. "  of the file; note this value for later and return cursor to start."
set "file_length" "fread_pos"
set "fread_pos" 0


. "============== DETERMINE LINE (ROW) ENDING STRUCTURE ==============="

. "Detect the line ending character, start with \n"
set "fread_delimiter" 10
set "$readline" "FREAD"
. "The string is as long as the file.  No \n are present"
if "$readline.length" >= "file_length" "lineEndingR"
. "The string ends with \r and stopped before the \r\n"
if "$readline.('$readline.length'-1)" = 13 "lineEndingRN"
set "x" "$readline.('$readline.length'-1)"
& "&x&"
. "Set $line_end to whatever line ending made sense..."
: "lineEndingN"
set "$line_end" "\n"
& "n"
goto "lineEndingDone"
: "lineEndingR"
set "$line_end" "\r"
& "r"
goto "lineEndingDone"
: "lineEndingRN"
set "$line_end" "\r\n"
goto "lineEndingDone"

. "lineEndingDone - We've detected the line ending."
: "lineEndingDone"
. "Our cursor's dirty so reset it."
set "FREAD_POS" 0
. "Read columns"
set "FREAD_DELIMITER" "$line_end.0"
set "$readline" "FREAD"
. "Is the last character a seperator?  If so, we'll use that to 
. "terminate instead."
if "$readline.('$readline.length'-1)" != "seperator" "noSeperatorAtEnd"
. "So there is a seperator!  Smash it in to $line_end !"
set "$line_end" "x&$line_end&"
set "$line_end.0" "seperator"

: "noSeperatorAtEnd"
. "We've figured out if the last seperator is just padding."
. "Note the byte where the header row ends."
set "end_pos" "FREAD_POS"
. "Our cursor's dirty so reset it."
set "FREAD_POS" 0


. "====================== DETERMINE COLUMN HEADERS ===================="

. "Scan the rows real quick to determine how many columns exist"
. "and what their names are."
set "column_count" 0
set "FREAD_DELIMITER" "seperator"

: "readHeaderCells"
. "Append this header's name with this char."
set "$_header('column_count')" "FREAD"
. "If we're past our first row position, we need to break out."
if "fread_pos" >= "end_pos" "doneReadHeaderCells"
. "Otherwise, yes, this is valid column."
inc "column_count" 1
goto "readHeaderCells"

: "doneReadHeaderCells"
. "We're done - but our last column certainly contains line endings"
. "and other junk!  Let's read it again.  Rewind the cursor:"
dec "FREAD_POS" "('$_header('column_count').length')"
. "Take another step back (remember the seperator!)"
dec "FREAD_POS" 1
. "We'll terminate at line endings now:"
set "FREAD_DELIMITER" "$line_end.0"
. "Read our counter in once more so it's properly terminated."
. "We'll use this technique later."
set "$_header('column_count')" "FREAD"

. "Increase FREAD_POS by any extra line end characters..."
inc "FREAD_POS" "('$line_end.length'-1)"


. "====================== DETERMINE COUNTER PREFIX ===================="

. "Let's figure out the prefix to use for the counters."
set "$prefix" "$csv&this_csv-file&"

. "Clip the file path out by seeking for '.'"
loop start
if "$prefix.('$prefix.length'-'loopcount')" = 47 "clipPathPos"
loop "('$prefix.length'-1)"
goto "doneClipPath"
: "clipPathPos"
set "$prefix" "$prefix+('$prefix.length'-'loopcount'+1)"
: "doneClipPath"

. "Clip the file extension out by seeking for '/'"
loop start
if "$prefix.('$prefix.length'-'loopcount')" = 46 "clipExtPos"
loop "('$prefix.length'-1)"
goto "doneClipExt"
: "clipExtPos"
set "$prefix" "$prefix#('$prefix.length'-'loopcount')"
: "doneClipExt"

. "We trimmed out our prefix."


. "========================= READ THE CSV FILE ======================="

set "row_number" 0
. "readRows - Loop for reading through rows.
: "readRows"
. "Extend this row's prefix with the row number.
set "$row_prefix" "&$prefix&&row_number&"
. "If they didn't set a name for the auto line number column, skip it."
if "$line_number_column.length" = 0 "noAutomaticID"
. "If they did, set the value of that column with the line number."
set "&$row_prefix&&$line_number_column&" "row_number"
: "noAutomaticID"

. "Get ready to read cells..."
set "FREAD_DELIMITER" "seperator"
set "this_cell" 0
. "readCell - Loop for reading through cells."
: "readCells"
. "Read the cell.  See if it's empty."
set "$readcell" "FREAD"
if "$readcell.length" <= 0 "doneReadEmptyCell"

. "See if casting it to int works.  If not, read as a string."
set "readcell" "$readcell"
if "$readcell" != "&readcell&" "readStringCell"

. "We're reading as int - no dollar sign."
: "readIntegerCell"
set "&$row_prefix&&$_header('this_cell')&" "readcell"
inc "this_cell" 1
goto "doneReadCell"

. "We're reading as string - dollar sign and interpolate."
: "readStringCell"
. "If the first char of this string is the lookup char, next block."
if "$readcell.0" = "cast_to_counter" "readCastCell"
set "$&$row_prefix&&$_header('this_cell')&" "&$readcell&"
inc "this_cell" 1
goto "doneReadCell"

. "We're reading a lookup cell!  Use &$readcell& to point to a counter."
: "readCastCell"
. "If it's a string, next block."
if "$readcell.1" = 36 "readCastStringCell"
set "&$row_prefix&&$_header('this_cell')&" "&$readcell&"
inc "this_cell" 1
goto "doneReadCell"

. "The lookup cell points to a string.  Add a $."
: "readCastStringCell"
set "$&$row_prefix&&$_header('this_cell')&" "&$readcell&"
inc "this_cell" 1
goto "doneReadCell"

. "doneReadEmptyCell - The cell was empty, so back up 1 step."
: "doneReadEmptyCell"
dec "FREAD_POS" 1

. "doneReadCell - After reading a cell, see what to do next."
: "doneReadCell"
. "If this is before last cell, keep going."
if "this_cell" <= "('column_count'-1)" "readCells"
. "If this is the last cell, use LINE END as seperator and keep going."
set "FREAD_DELIMITER" "$line_end.0"
if "this_cell" <= "('column_count')" "readCells"
. "We're done this row."
goto "doneRowCells"

. "doneReadCell - After reading a row, see what to do next."
: "doneRowCells"
inc "row_number" 1
. "Skip extraneous line ending characters."
inc "FREAD_POS" "('$line_end.length'-1)"
. "If we're not at the end, keep going."
if "FREAD_POS" < "('file_length')" "readRows"

: "doneThisFile"
. "Clean up - we're done this file!"
set "" "FREAD_OPEN"
inc "this_csv-file" 1
. "If we are past the csv count, we are done!"
if "this_csv-file" >= "csv_file_count" "doneCSV"
. "We aren't done - go back and open the next CSV!"
goto "readNextCSV"


: "doneCSV"
. "Done"
end
. "throwError - Uh-oh, we had an error!  Display what it was."
: "throwError"
write overlay c0c "Error: &$catch&" "scrolledx" "scrolledy"
set "" "FREAD_OPEN"
die

EDIT: Crap, this breaks under some circumstances. I'll fix it and make a thread or something.

This post has been edited by CJA: 08 January 2017 - 02:26 AM

[Dr Lancer-X]

Pretty cool stuff, especially how robust you've made it. A potential enhancement might be to support "-surrounded text fields, which can contain commas. I can see doing this if you wanted to include dialogue or potentially item descriptions or something.

[CJA]

So I've started trying to complete sub-tasks in this "capsule" way. Y'know how a traditional gel capsule looks? Sort of like a progress bar at 50%, but it's totally enclosed by a hard protective shell. Even if something isn't totally finished, I like to 'close it off' and program the ending code or design the exit or whatever before getting halfway. That way, if I don't really want to work on that part of the project, I can just put that 'capsule' aside and work on something else.

[Dr Lancer-X]

CJA, on 13 January 2017 - 07:02 AM, said:

So I've started trying to complete sub-tasks in this "capsule" way. Y'know how a traditional gel capsule looks? Sort of like a progress bar at 50%, but it's totally enclosed by a hard protective shell. Even if something isn't totally finished, I like to 'close it off' and program the ending code or design the exit or whatever before getting halfway. That way, if I don't really want to work on that part of the project, I can just put that 'capsule' aside and work on something else.

I liken it to architecture. You need to have scaffolding in place in order for construction to go ahead - you can't just make the whole house at once. So it makes sense to get a particular module into a very basic, completed state as quickly as possible so that you can design around it. A battle engine that shows the enemies you're fighting but closes when you press space, for example. Then you can work on the game and actually "finish" it - in other words, you can "play" the entire game from the beginning to the end. Then you start filling out the incomplete systems and you get a really strong sense of how much work you have to do. It's very motivating.

EDIT: By the way, this is in complete antithesis to the 'vertical slice' model of development, but I don't think much of vertical slice and it's not that important unless you need to demo your game at E3.

[CJA]

I googled vertical slice development. Vertical slice sounds like s***.

The player now has the ability to change cores by using key items now (mechanically, it's like changing class in the middle of a battle). And, of course, you can do this out of battle as well. I made sure to program the item system so that the actual item handler (Global robot) does the leg work, and then either it handles the healing outside of the battle with the message system, or it tells the battle engine to do it.

[CJA]

well, it's high noon somewhere...

forgive the horrendous letterboxing, I just wanted to show off this new attack.

[Graham]

CJA, on 15 January 2017 - 10:35 AM, said:

well, it's high noon somewhere...

forgive the horrendous letterboxing, I just wanted to show off this new attack.

:-D
That looks awesome CJA! Games with battle systems have never been by forte but I will give this a try. Keep up the good work.

This post has been edited by Graham: 15 January 2017 - 09:09 PM

[KKairos]

CJA, on 03 July 2016 - 08:29 PM, said:

Late to the party on this image, but: Love the design on those guys, but they kind of look like they're afraid of the player--was that on purpose?

PS Less late to the party, that video looks awesome.

[CJA]

So, except for the main character, the party totally changes in chapter 2, and you end up with a much more... lawless, chaotic party. Writing a different type of dialogue is welcome, because I felt like things were getting too technical and boring. I might actually go back and make some of the dialogue a bit more spicy and add more flair to the characters I have. They're not very well defined, just sort of written as "Shit people would say when faced with X" sort of dialogue.

[CJA]

KKairos, on 15 January 2017 - 08:05 PM, said:

Late to the party on this image, but: Love the design on those guys, but they kind of look like they're afraid of the player--was that on purpose?

PS Less late to the party, that video looks awesome.

I dunno what you mean. Are you seeing the black part as their eyes?

[KKairos]

I was mainly seeing the brown eyes as almost being "scared", but very possible I'm misreading the intent. If others are seeing other things I'd defer to them.

[CJA]

Well either way THEY SHOULD BE SCARED.

Also: tfw you realize that your game project is named "After <name of the company you work for now>"

[CJA]

Aasdfgaksb

Waugh.

I have done nothing