I'm assuming you realize that the severe squooshing makes the numerals in the draft very uncomfortable - that's at least as bad as the bad node in the "6".

I'm sure people can suggest good fonts* with narrow numerals and the diagonal-style "6" and "9", and I can see that budget is an issue, but to me this is crying out for custom numerals! Something that will make people think "port". Please feel free to contact me if you'd like to discuss pricing for such custom work: hpapazian at gmail dot com

* I might suggest you try Tarzana Narrow.

Also, I have to think that Times-Bold (for the large code in the middle) is sort of pedestrian in this context, especially since you're resorting to the classiness of Greek letters. What would be ideal is to make the font both more expressive and harmonious with the large two-digit background numbers.

Lastly, although I do see the clever circular ring of text all around the placemats, it might be useful to make the central large code also readable from more than one viewpoint; maybe have three (two would be less balanced) copies of the code pointing in 120 degree steps.

hhp

Thank you for the prompt reply. Lots to say.

• “squooshing”! I admit to using slightly different font sizes for different digits, to make them the same height, but for each digit the x and y sizes are the same. Squoosh-free!

• My limited experience of these matters suggests that there is likely to be great conflict between “budget is an issue” and “custom”.

• I took a screen-shot of a web example of Tarzana Narrow Regular, and moved things. A tightly spaced “66” was square. But A4 less my default margins has a ratio about 111:76, far from square. Or are you suggesting that I do some “squooshing”?

• “Times-Bold … is … pedestrian”. Well, yes. But it does have two-and-a-half advantages.
1. It has matching lower-case Greek letters.
2. It is free.
2½. I won’t offend any Port drinkers, who are, by nature, conservative creatures.

• There is no need to “make the central large code also readable from more than one viewpoint”. I sit in front of my own glasses. Everybody else sits in front of their own, which are arranged identically. (E.g., pictures taken at a recent tasting in which everybody had two 8½″×14″ US Legal pages, and another in which everybody had five sheets of A4.)

I just looked at FusiThinNormal, and you certainly did not do
any squooshing... the font itself is squooshed! :-/ In fact it's a
squooshed version of Futura. Which is why it's free. It's also
probably an illegal clone.

> there is likely to be great conflict between “budget is an issue” and “custom”.

:-)
The good news is we're talking about only 10 glyphs.

Tarzana-Narrow: you're right, it's not narrow enough. The thing is, most
ultra-narrow fonts don't feature the diagonal-style "6"/"9". What about this?
http://new.myfonts.com/fonts/parachute/pf-din-text-comp-pro/

Lastly, your rationales for Times and the one-directional code make sense.

hhp

Again, not narrow enough. The lining 6, used to make a super-tightly-spaced 66, is 1.22:1, significantly less tall than 1.46:1.

Also, I hate the 7. Both the corner just below top-right, and the top-left tail.

You're picky. Good for you. Now it's really saying "custom" to me. :-)

hhp

Hrant iz right.

http://new.myfonts.com/fonts/nicksfonts/turista-flaca-nf/
http://new.myfonts.com/fonts/liebe-fonts/liebe-erika/
http://new.myfonts.com/fonts/paratype/futura-book/cond-book/

Maybe not compressed enough:
http://new.myfonts.com/fonts/andrijtype/oksana-sans-compressed/

But Frode is right none the less.

Hmmm, Oksana Sans Fat Compressed Italic makes me think that -especially since the numerals will be outlined- omitting the counters might be a useful trick here...

In fact, that's a free font, and since it's missing the counters it might actually survive squooshing (and skewing upright). However it's extremely likely to need some tweaks after such brutish treatment.

hhp

I love the 0 and 7 in Turista Flaca NF. Shame that the bowl of the 6, and to a lesser extent of the 3, is so dis-proportioned.

Me, fussy? You don’t say.

Ten custom digits: ¿£$¥€? (Lining; sitting on baseline; matching heights; sans-serif; non-italic and non-oblique; the bounding box of a two-digit string in height:width of about ( (2^-1.75)*72/0.0254 - 2*30 ) / ( (2^-2.25)*72/0.0254 - 2*30 ) ≈ 1.46; parsimonious strokes so no excess corners nor 180° turns; bowl height about half total height.)

Maybe you can get a good deal asking Nick to customize Turista Flaca:
http://www.typophile.com/user/6470

Message sent to Nick — thank you for the link.

Perhaps others could suggest possible requirements that I might have forgotten.

• Digits should have level bottoms (on baseline) and tops. But a pointy top, ^, has a lot less visual weight at the top than a horizontal line (e.g., ‘7’) or a smooth curve (e.g., ‘0’ or ‘8’). So no pointy tops please.

Sorry I can't oblige, but my schedule doesn't permit and, even if it did, my experiences with working with people who know EXACTLY what they want have been less than ideal. Perhaps it has something to do with the definition of a Perfectionist: someone who takes great pains...and then gives them to others.

However, I can suggest that you might look at this...

http://new.myfonts.com/fonts/jonahfonts/credititle/caps/

Nonetheless, thank you for replying, and for the suggestion of Credititle-Caps. Again, the ‘7’ is not to my liking.

However, this thread has prompted one idea. Some ‘squooshing’ (is that a technical term?) might be acceptable. The code could gain additional parameters specifying a min and max y/x scaling ratio. This would widen the range of appropriate typefaces.

Is there a standard term for squooshiness, and for direction expressed? Perhaps one of y/x, x/y, y/x–1, x/y-1, Ln(y/x), Ln(x/y), where y and x refer to the scaling.

Universalis ADF Cond: Ascender/Descender = 750/250, proportional lining numbers, widest box ('eight') is 516.

Paratype's PT Sans Narrow: Ascender/Descender = 800/200, monospaced lining numbers, box width = 475

The font with the PostScript name UniversalisADFStd-BoldCond is generally good, but insufficiently tall. So I have added to my PostScript program two extra parameters:
% y/x scaling, bigger being taller, smaller being fatter
/BackgroundTextsSquooshMin 0.5 def
/BackgroundTextsSquooshMax 2.0 def
and added code to use them.

The program then chooses a squoosh of 1.72775, that allowing the 0 and 3 to touch top and bottom, and the ‘66’ to fill the width. The result is, again, at www.jdawiseman.com/port/20111011.pdf.

Is it perfect? No. Is it good enough? I think so.

Thank you all for help and suggestions.

Any typographer with a head attached to their body will tell you stretching or squooshing anything* beyond a couple of percent will destroy the original intent of the type designer. At this point you're slowly moving towards that grey zone between actual design and just tossing stuff onto your canvas. By now, Hrant -- or any other decent type designer for that matter -- could've easily drawn a set of numerals to fit your needs. Judging by the amount of time you've spent poking around here without satisfying result, hiring a professional seems a much better investment.

On a side note: Where does this fear of hiring a professional actually come from?

* Disregarding those extremely few typefaces designed with squooshing in mind.

My view is that usually 3% is pushing it.
Interestingly AFAIK this is the percentage used on supermodel photos.

> Judging by the amount of time you've spent
> poking around here without satisfying result

On the other hand poking around can be fun, and educational. Also Julian is flexing his math skills, and just like a muscle the brain does need to keep fit.

> Where does this fear of hiring a professional actually come from?

Usually the wallet, where else? Not a trifling thing Frode. But often there's also the admirable desire to do it yourself.

Frankly I've noticed you're sometimes overzealous in promoting professional work.

--

> www.jdawiseman.com/port/20111011.pdf

This is less bad than the first one, but the squooshing remains evident, so anybody who notices/minds it in the first one will do so almost as much in the new one. :-/ I'm sure you're mindful of the qualities of the ports you taste; you're the one to decide how much of that mindfulness to port over :-) to the activities peripheral to that particular finer thing in life.

hhp

> Any typographer with a head attached to their body will tell you stretching or squooshing anything* beyond a couple of percent will destroy the original intent of the type designer.

Well, yes. True. If this were more text, five words rather than than two characters, and set at a smaller size, 24pt say, then this squoosh would look foul. But it is only two characters, at several hundred point on A4, outlined faint grey.

Indeed, the original intent was probably not to be clip … stroked in faint grey with lots of differently-directioned seriffed black text on top, and glasses of port on that. So put aside the historical (US constitutional) issue of intent, and instead ask whether it looks satisfactory. And I think it does†. Not perfect—absolutely not perfect—but satisfactory for this task, and indeed for this class of tasks.

Indeed, there are improvements more important to me than the squoosh. I would prefer all the characters to have exactly the same heights; and the ‘6’ to have a flat or rounded top; and other changes too. But it is satisfactory as is.

† Obviously it would look a lot lot better with the glasses of port!

How much you can stretch a typeface?
http://www.artlebedev.com/mandership/134/

Frankly I've noticed you're sometimes overzealous in promoting professional work.

On flexing the math: Everyone can not do everything on the same level of professionalism. Very often hiring a professional will get you both a quicker and a better result than you would by yourself. In business, realizing your shortcomings is essential.

Anyway, you're right, H, I certainly can be too pessimistic. After all, I'm notorious for asking stupid questions on this very forum :)

... the original intent was probably not to be clip … stroked in faint grey ...

Exactly, yes. (Seriously, investigate it!) Those other things though, they are things one would expect any typeface to endure, although -- considering original intent -- some more so than others,

It's an interesting assumption that squooshing will be less noticable at a larger size, however your example says otherwise.

Kerning. When there are only two characters, they can each breathe plenty on three sides. So the fourth can be kerned very tightly. But if typesetting a paragraph of Lorem ipsum dolor sit amet (or the linked example’s cyrillic equivalent), then adjacent characters must be clearly separate but not separated. Different quantity ⇒ different kerning rules.

Likewise, perception of the relative widths of vertical and horizontal strokes varies with context. Bulk text must be treated as sacrosanct. Headline text is a little more decorative, and a little less prone to exhaustion through quantity, so there is a little more freedom to play. For example, at the larger size and smaller quantity serifs are less useful. And if setting just two characters on a whole page, things are different again.

Next, read a page of a book. You probably held the book perpendicular to the line of sight.

With two A4 pages on a dining table, portrait, one ‘above’ the other, the further page is at about 45° from my line of sight, and I’m 6′1½″. For me that is a foreshortening of a factor of about 1.4, and for shorter people the factor will be larger. There is enough wiggle-room to tolerate the squooshing, even for a type pedant.

I can't agree that a larger size makes it less of a problem.

hhp

Not larger size by itself.

All of: very few characters; size a very large proportion of the whole page; seen foreshortened.

> very few characters

No difference.

> size a very large proportion of the whole page

Proportion of page: no difference.

Large => Worse.

> seen foreshortened.

Greek-style optical compensation*? Awesome.
Post-rationalization for saving money? Not so awesome.

* Have you field-tested the average angle?
Is the Standard Deviation from that tolerable?

Just make sure any designers in attendance first visit the Bacardi 151 tasting. :-)

hhp

> Have you field-tested the average angle?
Not forgetting that the average angle may well change as the evening progresses!

Just to sum it up: It looks like crap.

If you want it to look great you need to spend a really long time learning how to make it look great or hire someone who already now. Pessimistic enough for ya?

It's port, so: merda.

hhp

All this discussion seems not to have accomplished much, other than to be a classic example of a tempest in a teapot. I sincerely doubt that the typography of the signage will have any effect whatsoever on the wine-tasting experience, and I equally sincerely doubt that anyone except the OPer will care WHAT the signs actually look like, as long as they're legible. However, at least it has provided an amusing diversion for the past four days, so I suppose it's not a total waste.

That doesn't sound like a type designer at all.

hhp

Julian, something just crossed my mind (I don't know why it took so long). Do you know what Metafont is? It's basically a way to define a font parametrically, with variables that can be changed. Especially since you're clearly very handy with code you should be able to take a suitable metafont sans and generate the width you need by changing the right parameter(s); and then you can export a PS font.

hhp

I have a great Port but want it to taste a bit more cokish. Anyone to advise me on wich kind of Coke to choose and how much to drop in?

Metafont: I know of, but have never used, Metafont. Good idea. On quick perusal it doesn’t seem easy. If you can recommend a .mf file specifying a decent sans-serif typeface, please do. Thank you for the idea.

Meanwhile, the typeface in the draft placemats for a tasting on 11th October has been changed to Santana-RegularCondensed, which is not pre-squooshed, and needs less squooshing by me. It will not satisfy all readers of this forum, but is an improvement.

Coke: if you really have a “great Port”, contact me privately to discuss a price. If you just want to drink coke, that’s fine (well, we can pretend it’s fine), but in that case just drink coke.

Andreas, I'm afraid your subtle message got lost ;-)

And not "That's sherry, not port!" lost, but "Where on Earth am I NOOOOW??" lost.

Anyway, do not use powdered forms of coke. They foul up the result. Badly.

Metafont: thoughts so far.

Metafont uses an interesting language, containing elements of PostScript and Mathematica.

But implementations seem to be a nightmare. What I envisaged was:
• there being somewhere accessible a .mf file specifying a parameterised sans-serif Helvetica-like typeface.
• there being an application, as user-friendly as compilers/developments applications can be, which can convert .mf files to some widely-used font format.

Instead:
• I can’t find a parameterised sans-serif Helvetica-like typeface, nor indeed, any others.
• The Metafont application seems to be a pure command-line thing.
• And the only route to PostScript is via a bitmap.

This seems like old information. For one
thing I'm sure you can export PS outlines.

Try asking this local guru:
http://typophile.com/node/73827

hhp

You sure, hrant, you're not thinking of MetaPost?

And now I’ve just noticed a node problem with the counter of the ‘6’ of Santana-RegularCondensed. Sigh.

Picky people simply can't save as much money!

It's very hard making a high-quality font, so it's rare to find a good free one, and extremely rare to find a good free one that fits a tight design brief. Sure, somebody could get lucky, but as others have said there comes a point where the time you spend trying to get lucky isn't cost-effective.

hhp

OK. I need ‘6’, ‘3’, ‘7’, and ‘0’.

• Pairs of digits (63, 66, 70, 77), properly kerned, which I can control, must approximately fill an A4 page less 30pt margins, a space in a ratio close to 111:76. More exactly, if you care, ( (2^-1.75)*72/0.0254 - 2*30 ) / ( (2^-2.25)*72/0.0254 - 2*30 ) to 1.

• Lining, sitting on baseline, matching heights.

• Upright: non-italic, non-oblique.

• Diagonal 6.

• Not fussy. That probably means sans-serif, but small serifs might be OK. It also means not odd or mal-proportioned, in any of the various senses in which a font can be so.

• Bowl height about half total height.

• With counters.

Contact details at www.jdawiseman.com/author.html.

• No foreshortening.

• The 111:76 ratio is for two digits. ✘✘✘ This does not mean that one digit is 111:38. ✘✘✘
◊ In ‘66’ the digits touch at the bowl, so need a little space there.
◊ In ‘70’ the space, in some typefaces, is negative, especially relative to ‘77’.
Repeat: the 111:76 ratio refers to pairs of digits.

Draft placemats for the 11th October tasting updated to use digit pairs drawn by Frode Bo Helland, in whose direction I doff my hat. Thank you.

There isn’t anywhere centre-stage to add a credit, but names and links appear in the crowded last page. Not perfect—sorry. If wording unsatisfactory, please request different.

Looks good!

Why does the "7" have -what looks like- a trap?

hhp

Because it’s trippy.

I hate to mention this, but Optical Center is normally perceived somewhat above Mathematical Center; as a result, your 6s appear malformed, IMHO...

If I were paying somebody a squillion monetary units to be perfect, then there would be a few small observations.

But I’m not, so instead, again, doff hat at the creator. Thank you.

They are above, but maybe not enough. Thanks for the comment.