Software that makes placemats

[jdaw1]

Placemats updated here.

Updated here and emailed to you as .ps file.

Observe that PhilW has used different URLs for the two versions, ‘a’ and ‘b’.

I have tended to have one name for my placemat files, repeatedly overwritten as updated. You have done versioning. Hmm, a question for elsewhere, but do people need versioning?

Thoughts?

Edit:

Habitual due to software releases etc, where source is under version control, but releases (and release candidates) get revision numbers. Not expected to be part of the postscript.

[PhilW]

Two suggestions for additional features:

1. Water box formatting
Currently the designer can specify the desired number of water tick boxes, but cannot control the number of bozes per row. In some layouts it would be more aesthetically pleasing to have a 'clump' of boxes rather than a long single line, even if they would fit. Two suggested 'simple' options:
(a) Add a parameter to control the maximum boxes per line, e.g.
/WaterCountTarget 10 def % already provided
/WaterCountRowMax 6 def
(b) Add an optional array parameter to specify the number of bozes per line, e.g.
/WaterCountTarget 10 def % already provided
/WaterCountRows [4 3 2 1] def

2. Packing style selection
Currently the code allows the designer to specify a preferred ordered list of packing styles; User-specified layout can also be used. An additional feature which could be useful in some situations would be the ability to specify on a per-page basis, which of
the list should be used (rather than the program automatically determining this). An example would be a two-page user-specified layout where the style on each page is to be different, e.g.
/PermittedPackingStyles [
[ [2 5] [0 4] [4 4] [2 3] ] % desired layout for glasses page 1
[ [1 1] [0 2] [4 2] [2 1] ] % desired layout for glasses page 2
] def
/EnforcePackingStyles [ 1 2 ] def

n.b. for consistency with other naming, the above variables might perhaps be better renamed as /PackingStylesPermitted and /PackingStylesEnforce

[jdaw1]

Water box formatting
Currently the designer can specify the desired number of water tick boxes, but cannot control the number of bozes per row. In some layouts it would be more aesthetically pleasing to have a 'clump' of boxes rather than a long single line, even if they would fit. Two suggested 'simple' options:
(a) Add a parameter to control the maximum boxes per line, e.g.
/WaterCountTarget 10 def % already provided
/WaterCountRowMax 6 def
(b) Add an optional array parameter to specify the number of bozes per line, e.g.
/WaterCountTarget 10 def % already provided
/WaterCountRows [4 3 2 1] def

Form (b) has two awkwardnesses. First, what if the requested length overlaps a circle. Second, the array needs to be of huge length to cope with all cases. I could avoid this by renaming to /WaterCountMaxRowLengths, with the last item applying to all subsequent rows. Would that be OK? Would this second form be materially more useful than a single integer WaterCountMaxRowLength?

[jdaw1]

Packing style selection
Currently the code allows the designer to specify a preferred ordered list of packing styles; User-specified layout can also be used. An additional feature which could be useful in some situations would be the ability to specify on a per-page basis, which of
the list should be used (rather than the program automatically determining this). An example would be a two-page user-specified layout where the style on each page is to be different, e.g.
/PermittedPackingStyles [
[ [2 5] [0 4] [4 4] [2 3] ] % desired layout for glasses page 1
[ [1 1] [0 2] [4 2] [2 1] ] % desired layout for glasses page 2
] def
/EnforcePackingStyles [ 1 2 ] def

n.b. for consistency with other naming, the above variables might perhaps be better renamed as /PackingStylesPermitted and /PackingStylesEnforce

What about something like

Code: Select all

/PermittedPackingStyles [ {SheetNum 0 eq {/This} {/None} ifelse} ! ]

or even

Code: Select all

/PermittedPackingStyles [ {[/This /That /None] SheetNum get} ! ] def

(where This and That are to be replaced with something else)? (Neither of which need any change to the code.)

[PhilW]

I could avoid this by renaming to /WaterCountMaxRowLengths, with the last item applying to all subsequent rows. Would that be OK? Would this second form be materially more useful than a single integer WaterCountMaxRowLength?

Yes, that sounds good and second form more useful imo.

What about something like

Code: Select all

/PermittedPackingStyles [ {SheetNum 0 eq {/This} {/None} ifelse} ! ]

Would be fine, so can ignore this one; I tried something similar which was rejected by the parser so thought that might not be allowed for some semantic reason I was missing, but the semantic error was user-based

[jdaw1]

I could avoid this by renaming to /WaterCountMaxRowLengths, with the last item applying to all subsequent rows. Would that be OK? Would this second form be materially more useful than a single integer WaterCountMaxRowLength?

Yes, that sounds good and second form more useful imo.

+WaterCountMaxRowLengths done.

Easiest way to update your code would be to search for ‟WaterCountMaxRowLengths”, and splice in that line. Then search in the new code for ‟%%%”, copy from there down, and paste into same place in yours. If being pedantic also substitute line 6.

[jdaw1]

I do agree that it would be neater if the two A1s touched. But how should that be expressed that in the parameter?

So [ [0 2] [2 2 3 2] [4 2 3 2] [6 2] [0 1] [3 1] [6 1] [0 0] [3 0] [6 0] ] would make the two A1s touch.

Done.

[DRT]

On my last trip to Porto and the Douro I purchased a notebook that I intend to use for future trips. I know of others who take the same notebook time after time to use as an immediate source of reference at tastings.

Perhaps it would be good to have the option of a Tasting Summary Sticker that included:

Tasting Name:
Date:
Location:
Attendees:
Wines:

These stickers could then be placed into notebooks in which tasting notes are to be written.

Yes, I know. It's just laziness not to write all of those things directly into the notebook. But not everything produced by the placemats programme is essential, so these would simply be for convenience of those who need them.

Default should be /TastingSummaryStickers = False

[jdaw1]

My immediate reaction is not enthusiastic.

Tasting Name:
Date:
Location:
Attendees:
Wines:

Much of this information isn’t available. Yes, there are headers defined, into which we typically put some of this, but not in a structured way. ‟Tasting Name”, ‟Date”, and ‟Location” will need extra parameters unless reproducing headers is sure to be satisfactory. Attendees are known to the code; as are the ‟Wines” (with the information spread around Circlearrays, Titles, Belowtitles, Abovetitles, and Overtitles). So it seems that this request doesn’t fit the mojo of the rest of the code.

But if others are enthusiastic I’ll reconsider.

[DRT]

it seems that this request doesn’t fit the mojo of the rest of the code.

I am unlikely to lose sleep as a result of this suggestion being rejected

[jdaw1]

In implementing WaterCountMaxRowLengths I forgot to re-include the parameters that control which pages hold water boxes. PhilW remarked on this bug at the Warre versus Fonseca vertical, and in the PM quoted here. It has been remedied, and is regretted.

[jdaw1]

• Layout was enforced since with minor margin change the default layout was sometimes varying from that intended (rows of 1-2-1-2-1 wanted, default varied between this and 2-2-1-2).

Code: Select all

% /PseudoHexagonal trivia: fix the number of glasses at eight, and vary the page's usable Height/Width ratio. At which ratios does the number of rows change?
% 2 -> 3    0.44444444      4 9 div                                                = 4/9
% 3 -> 4    0.68572971      7 sqrt 1 sub 5 mul 12 div                              = (Sqrt(7)-1) * 5/12
% 4 -> 3    0.70562062      35 sqrt 15 sqrt add 2 mul 21 sqrt sub 3 sub 17 div     = ( (Sqrt(35)+Sqrt(15))*2 - Sqrt(21) - 3 ) / 17
% 3 -> 5    1               1                                                      = 1
% 5 -> 4    1.41719214      35 sqrt 15 sqrt sub 2 mul 3 sub 21 sqrt add 4 div      = ( (Sqrt(35)-Sqrt(15))*2 - 3 + Sqrt(21) ) / 4
% 4 -> 5    1.45830052      7 sqrt 1 add 2 mul 5 div                               = (Sqrt(7)+1) * 2/5
% 5 -> 8    2.25            9 4 div                                                = 9/4

Excepting unity these are in pairs of rationalised reciprocals.

So PermittedPackingStyles being [ /PseudoHexagonal ] causes the less elegant four-row arrangement to happen with a height÷width ratio in the narrow range between ½âˆš35 + ¼√21 ”“ ½âˆš15 ”“ ¾ ≈ 1.4172, and â…–(1+√7) ≈ 1.4583.

[RAYC]

http://vintuswines.com/blog/2012/09/ An intersting approach to placemants by Noval - i like!

[jdaw1]

http://vintuswines.com/blog/2012/09/ An intersting approach to placemants by Noval - i like!

I like too.

[Alex Bridgeman]

http://vintuswines.com/blog/2012/09/ An intersting approach to placemants by Noval - i like!

Round glass bases on a rectangularly divided tasting mat - where the divisions are too small to hold the base of the glass? Hmmm. Not as much to my taste as the current tasting mats we use.

[jdaw1]

People liked the circle watermarks on the TN sheets. The code has just been changed such that these default to being present.

[jdaw1]

+CircletextsSameFontSizeIfRadiiShrunkToBeSame. The need for a feature like this had been on the placemat software’s to-do list since 2007.

[jdaw1]

Is The Bung Hole a confirmed venue or simply the default parameter now in the placemat software?!

It is the default of the placemat software’s main user but not of the software itself.

Though it’s not a bad idea. Why not write ‟The Bung Hole” rather than ‟London”?

Appropriate values also added to the default of ExternalLinks.

[jdaw1]

Very rarely one might want the centre heading non-centred (e.g.). This can be controlled with HeaderFooterCenterRelativeX, which is the centre of the HeaderCenterText, 0 being at the left margin, 1 being at the right, and 0.5 in the centre.

Also see the organisation thread for the tasting of a smattering of Graham.

[jdaw1]

The following quotations come from the parameters for the Smattering of Graham and Malvedos tasting.

Code: Select all

/HeaderFooterCenterRelativeX {TypeOfPagesBeingRendered /Glasses eq {GlassPositions SheetNum get 6 get 0 get MgnL sub PageWidth MgnL MgnR add sub div} {0.5} ifelse} def

Also the position of centre header needed to be changed, for which the code had been altered to add the parameter HeaderFooterCenterRelativeX. This was then assigned to some 132 characters of code, the complication of which might suggest that the form of this parameter was not optimally chosen.

The default value for HeaderFooterCenterRelativeX is plain ‟0.5”, as simple as can be. But the first time this parameter was used the value had to be changed to something computed from something deep in an array. Is that acceptable? Is this a question for the Taylor 1977 committee?

[DRT]

Would it be possible to print the left have portion of GM76 and G91 on the left hand sheet so that when the sheets are overlapped there is continuity in the spacing of the circles?

[jdaw1]

Would it be possible to print the left have portion of GM76 and G91 on the left hand sheet so that when the sheets are overlapped there is continuity in the spacing of the circles?

I thought about that, and it seems to me that would require a major reworking of the software.

What might perhaps be possible would be to print it to a custom paper size (about twice A3), and have a post-process to cut the print job into pieces.

But even if that could be done, in practice would the paper be positioned sufficiently accurately? And stay so positioned? Methinks not. And a similar question could be asked again with the answer presumed to be negative about the accuracy of the trimming of the margin area on which a printer won’t print.

[PhilW]

Would it be possible to print the left have portion of GM76 and G91 on the left hand sheet so that when the sheets are overlapped there is continuity in the spacing of the circles?

I thought about that, and it seems to me that would require a major reworking of the software.

What might perhaps be possible would be to print it to a custom paper size (about twice A3), and have a post-process to cut the print job into pieces.

But even if that could be done, in practice would the paper be positioned sufficiently accurately? And stay so positioned? Methinks not. And a similar question could be asked again with the answer presumed to be negative about the accuracy of the trimming of the margin area on which a printer won’t print.

Alternatively, just using A3, you could print the left-hand sheet as is, and print a right-hand sheet which included the right-hand 5 columns (i.e. from 75/90 inclusive). Then, any vertical slice on the right-hand sheet through 63/83 would create a page which would overlap directly onto the left page to provide the continuous image, without onerous precision on the cutting.

[jdaw1]

Alternatively, just using A3, you could print the left-hand sheet as is, and print a right-hand sheet which included the right-hand 5 columns (i.e. from 75/90 inclusive). Then, any vertical slice on the right-hand sheet through 63/83 would create a page which would overlap directly onto the left page to provide the continuous image, without onerous precision on the cutting.

That was completely obvious, but not quite so completely obvious that I thought of it. An excellent solution to the cutting problem. (Even if not quite to the precision-positioning problem.)

And the only change in the code needed, though not even needed in this case, would be in the internal variable TitleAboveBelowOverRandomSeeds, which would need to change such that the same item of Circlearrays/Titles/Belowtitles/etc in multiple locations would have the same ‘random’ positions of stars of flowers. Easily done. (Edit: done.)

[jdaw1]

There is the side-effect that it causes there to be two pre-pour sheets for the overlapping circles. The obvious and easy solution would be to have an extra boolean parameter, PrePourRemoveDuplicatesAndSortByWithinTitles (having the obvious interaction with PrePourReverseOrder), and StickyLabelsRemoveDuplicatesAndSortByWithinTitles.

[RAYC]

What proportion of people own magnum decanters?

Not I, and therefore i vote for a fancy bit of code that produces 2 decanter labels whenever the word "magnum" appears in titles, subtitles or overtitles etc.!

But i might look out for a magnum decanter also (now that i own 7 double magnums of port and quite a few more of wine, perhaps also a double magnum decanter...)

[jdaw1]

What proportion of people own magnum decanters?

Everybody who is anybody, surely?

Not I

Oh.

But i might look out for a magnum decanter also (now that i own 7 double magnums of port and quite a few more of wine, !)

Splendid.

perhaps also a double magnum decanter...

No: too heavy to use. A lot of wrist strength is needed to have the required careful control. Indeed, some have said the same of even a magnum decanter.

therefore i vote for a fancy bit of code that produces 2 decanter labels whenever the word "magnum" appears in titles, subtitles or overtitles etc.!

‟/DecanterLabelsNumCopies 0 def” admits the possibility of ‟2”. It doesn’t auto-detect the word ‟Magnum”, but can be set easily enough by hand.

[RAYC]

therefore i vote for a fancy bit of code that produces 2 decanter labels whenever the word "magnum" appears in titles, subtitles or overtitles etc.!

‟/DecanterLabelsNumCopies 0 def” admits the possibility of ‟2”. It doesn’t auto-detect the word ‟Magnum”, but can be set easily enough by hand.

thank you - the auto-detect feature was slightly tongue-in-cheek!

[jdaw1]

Alternatively, just using A3, you could print the left-hand sheet as is, and print a right-hand sheet which included the right-hand 5 columns (i.e. from 75/90 inclusive). Then, any vertical slice on the right-hand sheet through 63/83 would create a page which would overlap directly onto the left page to provide the continuous image, without onerous precision on the cutting.

That was completely obvious, but not quite so completely obvious that I thought of it. An excellent solution to the cutting problem. (Even if not quite to the precision-positioning problem.)

It worked well:

[jdaw1]

There can, rarely, be an additional complication. If a placemat is made from several A3 placemats, guillotined and overlapped !, then the same glass will appear on two sheets. But it should not have twice as many pre-pour sheets. This technicality is fixed with PrePourRemoveDuplicatesByWithinTitles, which, if true, removes duplicates, equivalence being determined by GlassesOnSheets having the same value of WithinTitles. The guillotine-and-overlap trick also messes with the order of the pre-pour sheets, sometimes fixable with PrePourSortByWithinTitles.

[DRT]

There can, rarely, be an additional complication. If a placemat is made from several A3 placemats, guillotined and overlapped !, then the same glass will appear on two sheets. But it should not have twice as many pre-pour sheets. This technicality is fixed with PrePourRemoveDuplicatesByWithinTitles, which, if true, removes duplicates, equivalence being determined by GlassesOnSheets having the same value of WithinTitles. The guillotine-and-overlap trick also messes with the order of the pre-pour sheets, sometimes fixable with PrePourSortByWithinTitles.

In summary: ask JDAW to produce the placemats when things get complicated.

[jdaw1]

In summary: ask JDAW to produce the placemats when things get complicated.

The new better behaviour is on by default, so you can do it.

[jdaw1]

At the Smattering of Graham and Malvedos, the consensus was that for the glasses that had to be lifted out, so those not at the edge, the radius should not be any smaller.

Radii = 120.2, all of them

Radius = 120.2 pt   ⇒   diameter ≈ 3.34″ ≈ 84.8 mm.

For comparison, five or six glasses on A4 (with 30pt margins all round) have a radius of 133.978, so just over 11% more.

[PhilW]

There can, rarely, be an additional complication. If a placemat is made from several A3 placemats, guillotined and overlapped !, then the same glass will appear on two sheets. But it should not have twice as many pre-pour sheets. This technicality is fixed with PrePourRemoveDuplicatesByWithinTitles, which, if true, removes duplicates, equivalence being determined by GlassesOnSheets having the same value of WithinTitles. The guillotine-and-overlap trick also messes with the order of the pre-pour sheets, sometimes fixable with PrePourSortByWithinTitles.

In summary: ask JDAW to produce the placemats when things get complicated.

Perhaps a simplification (in documentation) might be helpful; e.g. "For some layouts, it may be desirable for a specific bottle to appear on more than one glasses sheet (e.g. when using overlapping sheets). To avoid additional consequent duplication on subsequent sheets, define
PrePourRemoveDuplicatesByWithinTitles as true." Provided the duplication does not also occur on the tasting notes sheets, I don't think anyone would worry about the different pre-pour sheet order.

[jdaw1]

Documentation will be simplified thank you.

Which glasses are on which TN sheets is controleld by GlassesOnTastingNotePages. If the glasses are on A3, so sixteen-ish per page, then GlassesOnTastingNotePages shouldn’t be at its default value of equalling GlassesOnSheets. So the need to change the TNs is unaffected by the new functionality.

[jdaw1]

By request of DRT, yesterday evening, TastingNotesCirclesBehindTopX and TastingNotesCirclesBehindBottomX added. Documentation not yet updated, but see the eight TN sheets in the draft of the placemats for the March 2013 Taylor vertical. Comments welcomed.

[jdaw1]

Documentation updated with PhilW’s suggestion, and with new feature suggested by DRT.

see the eight TN sheets in the draft of the placemats for the March 2013 Taylor vertical. Comments welcomed.

Comments still welcomed.

[DRT]

Documentation updated with PhilW’s suggestion, and with new feature suggested by DRT.

see the eight TN sheets in the draft of the placemats for the March 2013 Taylor vertical. Comments welcomed.

Comments still welcomed.

My preference is for the layout of circles shown on the page starting at 1908, albeit that the positioning of the circles within the horizontal separator a is not quite perfect.

[RAYC]

I like the page starting 1942 but don't quite understand why it can't be spaced horizontally across the entire page (or at least centred), notwithstanding that the circles would not "touch".

[jdaw1]

I ! don't quite understand why it can't be spaced horizontally across the entire page (or at least centred), notwithstanding that the circles would not "touch".

Fitting circles tightly is part of the mojo of the whole software. I could add a boolean TastingNotesCirclesBehindFitAndCentreInRow, but don’t like it.

[jdaw1]

Fitting circles tightly is part of the mojo of the whole software. I could add a boolean TastingNotesCirclesBehindFitAndCentreInRow, but don’t like it.

Added. On the usual five-glass A4 example, this being true reduces the radius of the circles on the TN sheets from 87.7071 to 77.7381. Equivalently, setting TastingNotesCirclesBehindFitAndCentreInRow to false would increase the area of these circles by +27.3%.

The following come from the draft of the placemats for the March 2013 Taylor vertical (thread). The first image is old-style, with /TastingNotesCirclesBehindFitAndCentreInRow false def.

The next eight images are new-style, with /TastingNotesCirclesBehindFitAndCentreInRow true def.

(The exact selection of vintages has not yet been finalised. Some others might be added; perhaps a few of these will prove to be unavailable.)

[jdaw1]

The effect of the boolean TastingNotesCirclesBehindFitAndCentreInRow can be seen in the following two-frame GIF.

I prefer the bigger circles (radius 87.7 rather than 77.7); others don’t.

[jdaw1]

At the Bell AHB and RAYC did not like the layout of the circles on the TN sheets.

(Circles are usually fainter, at 0.2× the ink density of the glasses sheets, but on a small bitmap that was too faint, so here have been darkened to 0.5×.)

IIRC, the original complainants liked the first two sheets, diagonals \ and /, but not next two verticals |, and I don’t recall a comment on the fifth. Perhaps they could say more, and others could concur or otherwise.

[Alex Bridgeman]

My comment was that it caused my OCD to react badly when I was met with lines of circles which did not continue from the same point when moving to the next sheet in the tasting notes.

[jdaw1]

continue from the same point when moving to the next sheet in the tasting notes.

Does that mean that you would have been happy with an ordering like the following? (With obvious changing of which glasses on which sheets.)






Note to self: 0→1→0→½→½→1→1→½→0 gives the eight page types seen in the draft of the Taylor placemats for 22 March 2013, but in an order satisfying the ‘Bridgeman condition’.

[Alex Bridgeman]

Does that mean that you would have been happy with an ordering like the following?
Note to self: 0→1→0→½→½→1→1→½→0 gives the eight page types seen in the draft of the Taylor placemats for 22 March 2013, but in an order satisfying the ‘Bridgeman condition’.

Yes! This would have been much more comfortable for my aesthetics!

[jdaw1]

0→1→0→½→½→1→1→½→0

Or 0→1→1→0→½→½→1→½→0.
Or 0→1→1→0→½→1→½→½→0.
Do express a preference or I’ll have to choose. (Obviously, 0=Left, 1=Right, and ½=Centre.)

[Alex Bridgeman]

Depends on how many pages of tasting notes there are.

But say I prefer 0->.5->1->0.5->0

[PhilW]

Peronally I much prefer the diagonals over the verticals (and am not one of those previously expressing such view). I think probably because much of the page already contains grid-based elements, and the diagonal layout with circles adds good counterpoint. I also prefer the slightly larger circles which overlap the horizontal lines to the smaller ones which fit within; these are both purely aesthetic judgements.

[jdaw1]

Peronally I much prefer the diagonals over the verticals (and am not one of those previously expressing such view). I think probably because much of the page already contains grid-based elements, and the diagonal layout with circles adds good counterpoint. I also prefer the slightly larger circles which overlap the horizontal lines to the smaller ones which fit within; these are both purely aesthetic judgements.

I agree with your first point (diagonals), and for the reason you give. Which suggests 0→1→0→½→1→1→½→½→0.

I agree with your second point (larger rather than smaller), but the consensus was otherwise.