@HG Muller

@Aurelian: Just out of curiosity: what happens when you use the Play-Test Applet to convert the Interactive Diagram you made to GAME code? Does the castling work as intended then?

I'm not sure I know how to do this. I have read this:

https://www.chessvariants.com/invention/game-code-generation

but I did not understood how to do the weird castling, or something that needs doing in game code like initial pawn pushes. (the * in Xbetza).

I think I may have found an issue in Grand Cavalier Chess. In Self-play mode - 2 people on same PC. The Cavalier (Knight Pawn) is not able to get to the last rank for capture or promotion.

===

Question: according to the rules If you do not have any captured pieces the Cavalier can not promote. Can it go to the last rank for a capture and not promote if you have no captured pieces? ===

Can you check this?

Thank you.

Jim

Unfortunately Erik has not logged in since May and I don't know how to contact him. Any ideas?

@Bn Em I meant a piece that takes the first 2 steps like a rook and then towards outside like a bishop. The osprey jumps as a dabbabah and then moves towards outside like a bishop. So they are not the same piece.

In wikipedia Fairy chess pieces they are the Osprey for DthenB and the Ostrich for AthenR

Charles Gilman called these Zephyr and Lama respectively, though by his description they cannot make the one‐step move that I'm not sure whether you're including. The latter also turns up as an Osprey in Expanded Chess

Thank you, Fergus!

I still need to understand how to delete a value from an array in order to remove the necessary value from wcastle or bcastle.

You can unset an array element if you know its key. For example:

set ra array a b c d e f g h i j k l m n o p q r;
unset ra.4; // removes e from array

But what if you just know the element you want to delete and not its key? I have now added a diff operator to return an array difference. When an argument isn't already an array, it gets turned into a single element array. So, this code removes j1 from wcastle.

set wcastle d1 c1 j1 k1;
set wcastle diff #wcastle j1;

@Fergus,

I still need to understand how to delete a value from an array in order to remove the necessary value from wcastle or bcastle.

No ideea I'll Try it!

@Aurelian: Just out of curiosity: what happens when you use the Play-Test Applet to convert the Interactive Diagram you made to GAME code? Does the castling work as intended then?

I do not understand why after I move the rook I can castle with the cannon on the field that was supposed to be used only for the rook.

So far, you have not made it clear that the King moves to a different space when castling with each piece. The rules you have underneath the game do not even mention castling. The castle subroutine was not written with such a rule in mind. You might be able to make it work by adjusting the values of bcastle and wcastle after a Rook moves.

I think I haven't explained that properly but only one castle is aloud.

Castling is allowed only once, which I understand, but what you say next doesn't fit with you trying to say that. Are you trying to say that the King has only once space it can go to when castling with a certain piece?

Regarding your preset, I don't understand why you are using four colors for checkering the board. Also, the dark blue squares do not contrast well with the color used for highlighting legal moves, which makes it hard to see legal moves highlighted on those squares. Additionally, your rules could use images of the pieces, so that someone looking at a piece on the board can more easily tell how it moves. Remember to use the shortcodes for displaying pieces, so that the piece images below match those on the board no matter what piece set is used.

@Fergus Then I do not understand why after I move the rook I can castle with the cannon on the field that was supposed to be used only for the rook. I think I haven't explained that properly but only one castle is aloud. This is not free castling. So this is why I asked about wcastle and bcastle.

Adding response with Chrome on my Likebook Mars without being signed in first.

Replying from Chrome on my Likebook Mars.

Adding response from Chrome on my Likebook Mars.

Posting from Chrome in Windows 10.

@Fergus,

I am almost done with the castling. It was not that hard. I have left the normal stalemated subroutine which took care properly of the moves display. Also I am handling the flag sets and unsets properly as far as I can see. What I still need to do is delete a variable from wcastle or bcastle as the rook gets moved. How do I delete a value from an array?

Why in the castle subroutine in fairychess.txt include file the line unsetflag #RPOS; is commented?

I suppose it could be uncommented if you wanted to be thorough, but it isn't necessary to prevent illegal castling. Once the King has moved, the space it moves from and the space it moves to both get unflagged, and every space the King moves to gets unflagged. So, no matter what, the King is still unable to castle. As long as the King can't castle, nothing can castle with the King. So, leaving a flag on the Rook's original space doesn't change anything critical to the game.

@Fergus,

Why in the castle subroutine in fairychess.txt include file the line unsetflag #RPOS; is commented?

I think I accidentally deleted a comment while deleting some test comments.

1.The castling moves are not shown although they work.

The code for this goes in the stalemated subroutine, and here is what you have in that subroutine:

// Castling code removed, since Apothecary Chess does not include castling.

So you need to add some code for it there. You can probably borrow some from the one in the fairychess include file. I'm not sure if it will require any modification. Try it without modification first and see how it works.

2. The long cannon castle is not done entirely although the king moves correctly.

You don't seem to be unflagging spaces properly. I confirmed that you left the Rook's space unflagged by moving the Rook back after moving it and then castling with it. When the Rook's space remains flagged, you will not be able to castle with the Cannon. You should do as I do in Chess, which is to unflag every space moved to or from. The code used in Chess is

unsetflag $origin $dest;

Test. I keep getting the error message, "There is no comment to post."

I have repeated all my tests. The same two problems appear.

set wcastle d1 c1 j1 k1; set bcastle d12 c12 j12 k12;

From these, I gather that the King lies somewhere between d and j.

setflag a1 b1 k1 l1; setflag a12 b12 k12 l12;

But here, nothing is flagged between d and j. Make sure to flag every single piece that can castle, including the King.

I have forgotten about that:

https://www.chessvariants.com/play/pbm/play.php?game=Grand+Apothecary+Chess+2&settings=Default&submit=Edit

Can you give me a link to the preset you're working on?

@Fergus, Now I am trying to make the castling work.

The king is supposed to be able to castle with a rook by moving 3 spaces towards it or with a cannon by moving 4 spaces towards it, Usual conditions apply. I have used the following flags and variables:

set wcastle d1 c1 j1 k1; set bcastle d12 c12 j12 k12; setflag a1 b1 k1 l1; setflag a12 b12 k12 l12;

I have 2 problems. 1.The castling moves are not shown although they work. 2. The long cannon castle is not done entirely although the king moves correctly.

@Fergus,

Brilliant, it works and I understood them well enough to be able to make even other pieces.

Since Markdown altered your code by removing the multiplication operators, and HTML interprets the less than sign as the beginning of a tag, I am responding in Text. Here is your code for the Lyon: def Lyon fn (checkride #0 #1 1 0 and empty #0) where #0 * 2 sign - file #1 file #0 * 2 sign - rank #1 rank #0 #1 and not fn Knight #0 #1 or and checkride #0 #1 1 1 <= distance #0 #1 2 or fn Dabbabah #0 #1; To rule out regular Rook moves, this should fix it. This produced the correct pattern on an empty 8x8 board with e4 as the origin and also when I blocked the diagonal move at f5. def Lyon checkaride #ts #1 0 sign #rd or checkaride #ts #1 sign #fd 0 and checkride #0 #ts 1 1 and empty #ts =ts where #0 * 2 sign #fd * 2 sign #rd and match 2 abs #fd abs #rd =fd - file #1 file #0 =rd - rank #1 rank #0 or and checkride #0 #1 1 1 <= distance #0 #1 2 or fn Dabbabah #0 #1; Instead of using a lambda function, this assigns values to variables that get reused. These are rd for rank distance, fd for file distance, and ts for turning space. After calculating fd and rd, it makes sure that one of them has an absolute value of 2. It then uses them to calculate ts. After checking that ts is empty, it checks that a diagonal move from #0 to #ts is legal, and it then uses the value of fd or rd with checkaride for checking a Rook move in a specific direction from #ts to #1. It first checks for a Rook move along the file of ts. If that returns false, it then checks for a Rook move along the rank of ts. When I tried checkride #ts 1 1 0, it gave inaccurate results. The following code gave the right results on an empty board, but it gave false positives when I blocked it on f5. def Lyon fn (checkride #0 #1 1 0 and empty #0) where #0 * 2 sign - file #1 file #0 * 2 sign - rank #1 rank #0 #1 and not fn Knight #0 #1 and not checkride #0 #1 1 0 or and checkride #0 #1 1 1 <= distance #0 #1 2 or fn Dabbabah #0 #1; I will let you figure out how to fix the Tiger, since you should have more of a clue now.

@Fergus

I am trying to write the game code for 2 new pieces I want for my newest chess variants. The first, a bent rider, called the lion should move like a bishop for 2 squares and then towards outside like a rook. It may also move like a dabbabah. I have written the code bellow that has the problem that allows the final rook moves without checking for the first bishop part of the move. Please take a look:

def Lyon

fn (checkride #0 #1 1 0 and empty #0)

where #0 2 sign - file #1 file #0 2 sign - rank #1 rank #0 #1

and not fn Knight #0 #1

or and checkride #0 #1 1 1 <= distance #0 #1 2

or fn Dabbabah #0 #1;

The second, it is also a bent rider, called the tiger should move like a rook for 3 squares and then towards outside like a bishop. It may also move like an alfil. In the same way as before the bishop moves are allowed by my code whitout checking the rook ride. Code so far:

def Tiger

fn (checkride #0 #1 1 1 and empty #0)

where #0 0 * 3 sign - rank #1 rank #0 #1

or fn (checkride #0 #1 1 1 and empty #0)

where #0 * 3 sign - file #1 file #0 0 #1

and not fn Knight #0 #1

or and checkride #0 #1 1 0 <= distance #0 #1 3

or fn Elephant #0 #1;

I could not find code that block the ride in the first part.

That's probably fine, yes.

It looks like the Game Courier play-counts, but I'm not sure how it got there.

I build such a table in the ad hoc reporting page, but I don't write it back to the database. I could have created it at some point in developing the script? (The script's last change was in Dec 2018.)

The database has a table called ct. Its two columns are Game and Count. What is this for?

I cannot find the matches in the Apothecary Chess Tournament 1st Edition anymore.

What is the tournament id?

@Fergus: I cannot find the matches in the Apothecary Chess Tournament 1st Edition anymore.

I'm starting this subject thread for broad submissions of mathematical or scientific subjects that involve chess.

To kick it off, 3D chess is Turing Complete.

I don't understand your question, Fergus. I think I have explained it above.

A third thing if I may ask in these game the king may castle in 2 ways with the rook that starts near the corner of the board or with the corner cannon. Omega chess implements a castle move with a non corner piece. Can I combine this with the regular castling subroutine in order to easily make my castle subroutine work?

What castling rule do you want it to work for?

Could you find the time to also solve for berolina pawns the problem that regular pawns used to have in fairychess.txt

Okay, I've done that.

@Fergus,

Could you find the time to also solve for berolina pawns the problem that regular pawns used to have in fairychess.txt

I can't figure it why but I have a green background and a blue highlight.

At some point, I started using highlight color instead. I included this in the form players could use to customize the appearance of a game, but I neglected to add it to the Edit form used by game designers. I have now added it in.

I can't figure it why but I have a green background and a blue highlight. Look here, please:

https://www.chessvariants.com/play/pbm/play.php?game=Grand+Apothecary+Chess+1&settings=Default

Fergus, What about the ability of changing the blue color of the highlighting when showing possible moves.

It uses the border color to highlight possible moves.

@Fergus, Hello,

I had asked 3 questions on this thread 2 weeks ago. It seems that you have missed them. Can you find time to take a look?

Instead of setting the body text at 18px, I have erased code for setting the size of the body text, so that it will now use the system default. On the Windows desktop, at least, this is typically 16px. I checked multiple devices, and on each one, it seemed to make the text smaller, though Android and iOS devices did not let me inspect the elements and determine the precise size.

One reason for making this change is that Literata, which is now the body font instead of Lora, has a clearer look at smaller sizes. Another is that it should make the display more responsive to particular devices. Instead of using a fixed size across all devices, it lets the device determine the size. Also, I had previously set the font-size for comments to 16px, and I haven't had trouble reading comments. However, comments will now display text at the same size as the main content instead of at a smaller size. So, comments should generally display at the same size as they used to. Finally, the font size will now better match whatever zoom level someone prefers to use his browser at.

Since the heading sizes are defined with the em unit, which is determined by the body font size, they adjust according to the body font size. So, with the body font a bit smaller on most devices, the headings will also be a bit smaller.

But can you also do that with headers?

Someone might want to add paragraphs in a smaller-than-standard font.

If you want small text, you can enclose your text in <small>small tags</small>.

I looked into the default sizes of H1-H6, and I learned that by default, H4 is the same size as the body text, and H5 and H6 are even smaller. This doesn't make sense to me. Headings should always be larger than the body text.

Someone might want to add paragraphs in a smaller-than-standard font. And these might need headers too.

Four header levels already seems quite generous to me.

I think the H2 text size is very, very large. I'd like to see that scaled down a bit, but things are looking better.

Is that because you are used to it being smaller, or because there are specific pages on which the H2 heading is too long?

I looked into the default sizes of H1-H6, and I learned that by default, H4 is the same size as the body text, and H5 and H6 are even smaller. This doesn't make sense to me. Headings should always be larger than the body text. Additionally, higher level headings should be larger than lower level headings, and size differences between headings of the same style should be large enough to tell them apart. Bearing these considerations in mind results in the headings all being larger than they are by default. This doesn't leave a lot of room for adjusting the size of H2, which is already closer to the size of H3 than to the size of H1.

Additionally, text is appearing larger on Chrome than on Firefox. I normally develop on Firefox, but if you're using Chrome, you will see larger fonts than I normally do. When I inspect the page h1-h6-test.html, which has no CSS or styling, it gives me the same values for the font sizes even though it appears larger on Chrome. In inspecting, Xiangqi, I get the same values for the style elements, though I cannot get Firefox to give me the size in pixels. I need to look into why text is larger on Chrome and see if there is some way to standardize the size in different browsers on the same computer. You may inspect these pages yourself with Developer Tools on Chrome or Web Developer Tools on Firefox.

There are pages on which I plan to use a different style or a lower heading level. These would include link pages, Zillions pages, and Game Courier preset pages, which normally have fewer sections and less text than other pages. I'll work on that later.

You can press ctrl+f5 to clear cache for current page only

You might want to read the later comments before replying to earlier comments that are no longer pertinent. If you still see the double underlines, you need to clear your cache.

You are correct - my appologies.  I'm travelling and on my father's computer, so I didn't want to clear cache, but I've opened a different browser so I can see how it looks now.  This is a definite improvement.  I think the H2 text size is very, very large.  I'd like to see that scaled down a bit, but things are looking better.  

The h5 style makes the impression of being 'incomplete', because it doesn't have left and right vertical borders.

It also doesn't have horizontal borders. If you are seeing it with any borders, you need to clear your cache.

The typeface of h1-h4 'blends in' very well with the typeface of the main text, but h5 and h6 somehow don't.

Also, those no longer use small-caps. Clear your cache to see how it currently looks.

After looking at some pages, I decided that the pseudo horizontal rule works better for H2. It more clearly sends the message that this is a major section. Since Wikipedia does it, it should be familiar to people, and I don't think it would be considered an unprofessional look.

OK, this is better. Not necessarily better than last week's styles, but at least not far worse. The typeface no longer sticks out as typewriter-like, which I think is a huge improvement. (It is true it was only used in headers, but visually the headers dominate the impression the page makes, because they have such large font sizes.)

Presented in the comment the full-width underlining struck me as a bit weird, but I must admit that in the typical article with an Introduction / Setup / Pieces / Rules / Notes section it works.

Having headers in italics still strikes me as odd, though. But having 6 header levels seems overly generous; I was nottt even aware that HTML supported h5 and h6, and I cannot imagine I would ever need more than 3. So I suppose there isn't any real downside to having some of the lower-level headers in an uncommon style; most people won't need them, and the few that do need more than 3 levels but less than 6 can skip the styles they don't like.

The h5 style makes the impression of being 'incomplete', because it doesn't have left and right vertical borders. Even then, it is still a very exotic header style. The typeface of h1-h4 'blends in' very well with the typeface of the main text, but h5 and h6 somehow don't.

There still is a lare size gap between h4 and the main text. I would expect there to exist headers that had nearly the same size as the main text, but distinguish themselves by being boldface.

The font that is now on HTML buttons still seems fixed space. I don' think that is a good idea; it makes the buttons unnecessarily wide.

You might want to read the later comments before replying to earlier comments that are no longer pertinent. If you still see the double underlines, you need to clear your cache.

It's important to distinguish each heading level by a different visual style.

This is a hobbyist site, weird is a personal opinion, and there are only so many different ways of distinguishing six levels of headings by style.

Wavy underlines, which I already tried out, look much worse.

Fergus, I'm pretty surprised at what I'm reading.  You have suddenly decided, unilaterally, that six levels MUST be displayed so differently that it slaps you in the face.  And, if the double underline looks terrible, hey, at least they are not wavy?  Is that really your answer?

I think you would agree that I have complained about very, very little all these years, despite several large issues worthy of complaint.  I'd rather not go into enumeration, but I will if you doubt it.

The fact is 99% of the content on this site was not created by you.  You cannot just do whatever you want with it.  The huge, bold, double-underline type is categorically unacceptable to me.

I switched back to using Noto Sans for headings on the main site, though I will use Courier Prime for headings in Game Courier for thematic reasons. I could not find any good substitute for Noto Sans among sans serif fonts. I wanted one that put bars on the capital I, and there were few of those. I tried out IBM Plex Sans, but it didn't look as good. I figured out how to keep letters in Noto Sans from touching the horizontal rule by using some bottom padding. Since I'm no longer using small-caps or underlining and overlining, H5 and H6 no longer look as bad as they used to when I was trying out Noto Sans. With these problems taken care of, I figured it was okay to go back to Noto Sans.

After looking at some pages, I decided that the pseudo horizontal rule works better for H2. It more clearly sends the message that this is a major section. Since Wikipedia does it, it should be familiar to people, and I don't think it would be considered an unprofessional look.

I am trying out a set of headings styles that don't use any horizontal lines. Instead of making all the styles completely different, I am using sharp differences in size between headings using the same style. H5 has the same style as H3, and H6 the same as H4, but each is 0.55em smaller.

I think I'm fine with the (pseudo)underline for h2, but the under/overlining of h5 makes it stand out more than h3-4.

It works for me, and the only page I know of with H5 and H6 headings is the Game Courier Developer's Guide, which I wrote anyway. If I'm to do away with the lines, I need an alternative that works as well, and I don't have one. I did try small-caps for H5 and small-caps plus italics for H6, but italicized small-caps differs from small-caps merely by being slanted. So, I chose not to do it that way.

Since you pointed out wikipedia's, note that h4-h6 are all the same, and h3 only differs from those in font-size.

Yes, I'm not a fan of how Wikipedia is styled. I borrowed one idea from Wikipedia, because it seems to work, but I don't like how some of their headings differ only by size. It makes it harder to tell what level something is.

I think I'm fine with the (pseudo)underline for h2, but the under/overlining of h5 makes it stand out more than h3-4.

By the time someone gets to h6, I don't think much distinction really needs to be made; boldface but otherwise p-style would be nearly enough. (Since you pointed out wikipedia's, note that h4-h6 are all the same, and h3 only differs from those in font-size.)

I temporarily tried Lexend for headings. It looked better than Noto Sans, because it had a geometric design like Futura rather than a grotesque design like Helvetica. However, it did not have an italic style, and using weight to distinguish the heading styles didn't distinguish them enough.

The typeface makes the articles look like typewritten manuscripts rather than published articles,

Courier is being used only for headings, not for the body text. I find that it looks better than Noto Sans, and it works better with horizontal lines. When I tried Noto Sans, its descenders touched the horizontal line beneath the H2 heading, and the lines in the H5 heading were not the same distance from the text. I also dislike how barren and uninteresting sans serif fonts look when used for headings. I find that using a non-proportional font for the headings makes them stand out a little bit better and distinguishes them from the body text more.

and having underlining in the headers looks weird and unproffessional.

This is a hobbyist site, weird is a personal opinion, and there are only so many different ways of distinguishing six levels of headings by style.

Since pages were looking too cluttered with horizontal rules at the top, I took some steps to reduce this clutter. First, I removed the H2 heading for the introduction section in member-submitted pages. Second, I changed the notice that sometimes appears at the top of a page to a paragraph with the notice class instead of text with a horizontal rule above and below it.

To prevent the H1-H6 styles from being used in the wrong place, I limited these styles to headings with an article, or a section within an article, as the immediate parent. In other contexts, such as comments, footers, nav bars, or tables, these styles will not be used. They may be used in comments by enclosing them in an appropriate parent element, like so:

these huge, bold, double-underlined headings look absolutely horrible.

Wavy underlines, which I already tried out, look much worse.

Please change it back.

It's important to distinguish each heading level by a different visual style. I am now trying out a style used on Wikipedia for H2. It puts a full-width line underneath the heading. I have done this with the bottom border. So, the line no longer gets broken up by characters with descenders.

I have to say, I think these huge, bold, double-underlined headings look absolutely horrible. Please change it back. Or, at a minimum, get rid of the underlines.

As far as I am concerned this header business is a large step in the direction of uglifying the website. The typeface makes the articles look like typewritten manuscripts rather than published articles, and having underlining in the headers looks weird and unproffessional. On a scale of 1-10 I would judge the old header styles perhaps as 8, and the new situation as a 3. And this new (fixed-spaced) typeface now also appears on HTML buttons...

I adjusted the margins of the headings to be the same, and I put more space above a heading than below it so that each heading is closer to the text it is a heading for. But I also avoided making it look crowded by putting a heading too close to the text following it.

Here is why I am using small-caps for headings five and six instead of for heading two, as I used to do in Game Courier:

Using small-caps reduces the width of the line in Courier Prime even though it is not proportional.

Using small-caps reduces the width of the line in Courier Prime even though it is not proportional.

I've been working on redesigning the headings so that one can intuitively tell one level from another. Besides decreasing in size, they also differ in style.

Heading One: Centered

Heading Two: Double Underlined

I had originally just used underlining, but I changed it to double underlining to avoid confusion with links, which are normally underlined with a single line.

Heading Three: Normal

Heading Four: Italics

Heading Five: Small-Caps with Overline and Underline

Because small-caps are more compact, they are used for lower headings rather than for higher ones. The use of lines parallels the use of lines in the second heading. Because small-caps all have the same height, words have a more straight appearance, which works better with the overline.

Heading Six: Small-Caps

Like the change from heading two to heading three, heading six omits the lines used in the heading just above it.

I learned last night about a new type of font file called a Variable font. This includes all the styles for a typeface within a single file. Previously, each style of a typeface would get its own font file. Since the typeface used for the main body of this site was now available as a Variable font, I updated it. But while I was at Google Fonts today, I decided to check out other fonts.

I ended up replacing Lora with Literata. This typeface was designed for use with Google Play Books, and its basically the Google counterpart to Amazon's Bookerly. I liked it better than Lora for a few reasons.

1. It looks cleaner at small sizes. At the size of text used on the website, some characters in Lora have a bit of a blur to them, because they are being anti-aliased to look good at a smaller size than they were actually designed for. At the same size, characters in Literata lack this blur.
2. Its lowercase letters have a lower X-height, which creates more contrast between upper and lowercase letters.
3. Although its Q does reach underneath the next letter, it's only by a little bit, and I otherwise prefer the look of its Q, whose curved tail is similar to the one used in Century Schoolbook.

I also replaced FreeMono with Courier Prime, because with Game Courier on the site, I wanted a Courier monospace font. I have begun using it instead of the sans serif font for headings.

While I did try some other sans serif fonts, I stuck with Noto Sans. Presently, Literata is the only Variable font on the site. The others use individual font files for each style.

@Fergus, I think the berolina pawns have the same problem that the regular pawns had!

Fergus, What about the ability of changing the blue color of the highlighting when showing possible moves.

You were correct in identifying the problem. The range functions for the Pawns did not take into account the value of fps. However, replacing 2 with fps would not allow for first Pawn moves that are shorter than the maximum possible distance. For example, if fps were set to 3, this would include 3-space moves, but it would not include 2-space moves. This is what I ended up with:

def White_Pawn-Range mergeall where #ori -1 1 where #ori 1 1 filter lambda (onboard #1) aggregate lambda (where #ori 0 #0) range 1 var fps =ori;
def Black_Pawn-Range mergeall where #ori -1 -1 where #ori 1 -1 filter lambda (onboard #1) aggregate lambda (where #ori 0 neg #0) range 1 var fps =ori;

These use aggregate to create an array of spaces a Pawn may advance forward to, and they use filter to weed out non-spaces that are out-of-range. They use a named parameter in order to use its value in the lambda function used by aggregate.

I have since modified the functions above to use the filter on all results:

def White_Pawn-Range filter lambda (onboard #1) mergeall where #ori -1 1 where #ori 1 1 aggregate lambda (where #ori 0 #0) range 1 var fps =ori;
def Black_Pawn-Range filter lambda (onboard #1) mergeall where #ori -1 -1 where #ori 1 -1 aggregate lambda (where #ori 0 neg #0) range 1 var fps =ori;

Finally, I replaced aggregate with an expanded version of the values operator. With a lambda function and an array, it will work just like filter except that it returns a sequential array of each value calculated by the lambda function. Unlike aggregate, it will include values of zero, because it does no filtering on the results it calculates. As it did before, it will return the values of an array if that is what's passed to it. If it does not get a lambda function or an array, it will return the rest of the arguments following it as a single array.

def White_Pawn-Range filter lambda (onboard #1) mergeall where #ori -1 1 where #ori 1 1 every lambda (where #ori 0 #1) range 1 var fps =ori;
def Black_Pawn-Range filter lambda (onboard #1) mergeall where #ori -1 -1 where #ori 1 -1 every lambda (where #ori 0 neg #1) range 1 var fps =ori;

Fergus, Is it not that in the fairy chess include file the functions :

def White_Pawn-Range array where #0 0 2 where #0 0 1 where #0 -1 1 where #0 1 1; def Black_Pawn-Range array where #0 0 -2 where #0 0 -1 where #0 -1 -1 where #0 1 -1;

for generality they should be:

def White_Pawn-Range array where #0 0 var fps where #0 0 1 where #0 -1 1 where #0 1 1; def Black_Pawn-Range array where #0 0 - var fps where #0 0 -1 where #0 -1 -1 where #0 1 -1;

?

Fergus, is there a way to customize the blue color that highlights possible moves?