"Points" are meaningless.

[Mark Simonson]

It does really seem like design software ought to let you specify font sizes by cap height or lowercase height or whatever. The necessary data is already in the fonts, even the old PostScript Type 1 variety. It's just software running on computers, after all.

[John Hudson]

Font size, as expressed in points, is strictly a measure of how far down the page your cursor goes when you hit the return key.

No. The distance that your cursor moves down the page when you hit return will only be equivalent to font size expressed in points if the text is 'set solid', i.e. if there is no additional interline distance. Otherwise, the cursor will either move the point size plus some explicit leading value, if that is the typesetting method being used (default in e.g. InDesign), or will move a distance equivalent to the sum of vertical metrics in the font OS/2 table (Windows) or hhea (Mac) table (default in e.g. WordPad), or will move some multiple of the sum of those vertical metrics (default in MS Word).

Let me try to explain point sizing for vector form digital types as succinctly as possible:

The vector outlines defining the shapes of letters, numbers, punctuation etc. in a font are plotted on a unit grid. The fineness of this grid is defined in terms of 'units per em', in which an 'em' is a square that, when scaled, will be equivalent to the nominal point size of text. The em height is directly analogous to what was known as the body height in metal type. So, for example, the vector outlines of a letter might be plotted on a grid of 2048 units per em (pretty common for TrueType fonts). When you are setting text, it is this em that is scaled to the nominal point size. Think of it this way: the em grid is like a canvas on which the letter is painted, and when you scale the text what you are scaling is the canvas, not the letter directly.

[JamesM]

.

[hrant]

The printer's point in very precise without being needlessly precise.

I'm sorry I misunderstood you, Victoria. I guess you're simply saying that it's a good scale. I can dig that, but would ask: how often do people end up specifying fractional points? Because picky typographers will often specify something like 10.5 point. And when they don't specify fractional points, how often is that an unfortunate result of a "whatever, dude"?

For example if 1/2 point sizes are commonly specified (I mean by designers who think readers are affected by the difference) but anything finer isn't, then I would say the scale should be twice as fine.

And one thing's for sure, the point must be a pain in metric-system-using countries, which is most of them.

hhp

[eliason]

Except that's not as correct as the descriptions above.

[hrant]

Yeah James, that thing is highly misleading in at least two ways, sorry.

hhp

[Nick Shinn]

Good point, Mark.

**

Why points? It’s all about integers:

Picas and points are nicely scaled to typographical distances — which is the point made earlier about kilometers or miles being the right scale for considering geographical distance.

Picas: if you have to make a table 4½″ wide with 6 columns of varying width? That’s 26 picas, which is an easy number to solve this problem.

Points: if you have to decide between different sizes of type for a particular job, choosing between 8 pt and 9 pt is more manageable than comparing cap heights in inches or millimetres.

[JamesM]

[diagram deleted]
Since I apparently screwed up the diagram I've deleted it. Maybe someone can post a better one.

[rcapeto]

Hrant: And one thing's for sure, the point must be a pain in metric-system-using countries, which is most of them.

Indeed.

Pancho Gálvez, in his nice book Educación tipográfica, suggests, IIRC, that one could turn to points/picas for all the measurements on the page. To me that would be impossible.

Just before the worldwide onslaught of DTP software from US companies made it irrelevant, there had been in Europe a proposition of an adjusted Didot point measuring 0.375 mm (instead of the traditional 0.3759), which would make 8 pt = 3 mm, fitting metric pages more easily. I have a nice Faber-Castell typometer using that revised Didot point, and when I used to do this sort of work more often I would reset QuarkXPress's Didot point to this value.

In any case, regarding the thread's subject, we're still talking about the box, not what's inside it. As to this, it's a bit more complex than the original question implies. Even if you mean a plain UC, there may be issues such as hanging J's, Q's tails, etc. And, to be fussy, what about the overshoots.

[timd]

Points are not meaningless if they are used within one type family, outside those parameters they can often become a stumbling block.

In your example, it is true that a measurement system that reflects those of the rest of the grid might be more useful. However, setting a calendar brings a series of problems that you do not frequently face in other jobs.

For improvements, I would consider kerning and indents and baseline shifts on your dashes.

Tim

[oldnick]

@dumpling

Really? All this fuss because you can’t create calendar pages as easily as you would like?

As you may have discerned from the discussion above, there are very good reasons why things are done the way they are done, and not the way you would like to see them done to suit your particular needs…

[hrant]

Rodolfo! Nice to see you back.

Nick, come on. When you're designing display matter and you want the caps (or less often the lc x-height) to be a certain size it takes some extra work that should be unnecessary (considering the data is there in the font). Also, if you want a block of text to harmonize with the rest of a metric-system layout, the point is non-ideal; just like serious fields like science don't use the Imperial System any more, we should switch too (to Q).

hhp

[quadibloc]

@John Hudson:
No. The distance that your cursor moves down the page when you hit return will only be equivalent to font size expressed in points if the text is 'set solid', i.e. if there is no additional interline distance.

This is true, but it is also somewhat irrelevant, since his statement was simply a complaint that point size has no fixed ratio to x-height or cap height and so on.

Also, there is another exception, at least in some word processors. Sometimes, two fonts will have different baselines, and mixing them on the same line at the same point size will result in the line taking up additional space corresponding to the difference in baselines.

[hrant]

But the physical page (the thing the stuff ends up on, so has to work with) will be either inch-based or meter-based, so...

hhp

[russellm]

The unit of measurement you choose makes no [material] difference. It's just different numbers for the same thing.

I like to measure with barley corns whenever the opportunity arises.

[russellm]

So, I can deal with decimals.

Or rather, my computer can.

:o)

My point in saying that the kind of unit used is irrelevant is that I understand the original question to be about how to determine the dimensions of the visible portion of type and the space above and below.

Ultimately, as designers, why care? If it looks right to your eye, then it is right. We ain't tool and die makers when we make calendars. None of this matters until you have to replicate a layout or build something with real materials, in which case all the data you need is already there. You can round your dimensions off to fit an existing standard or you if feel that that makes life easier, or you can live with something cumbersome like 19.976 barley corns, because you don't need to remember it. (Or even know it.) Your computer does that for you.

While it is possible to obtain that data from the font itself this is already done to a certain extent. llustrator and CorelDraw both recognize the base line.

I design signs for a living and am, always having to set sigle lines of type to specific cap heights and place them in precise locations on sign faces - using the methods described above. When spec'ing for sign shops and design consultants, I'll define the cap height and, if I am concerned they might have limited skills in the field, I provide the size in points that will produce the size of text I want. I.e, to get a cap-height of 25mm, and in brackets, I specify 106 points* (for ClearviewADA).

* 0.149 points, or 0.053 mm or 0.0021 inches too large, which is an acceptable magin or error

[dumpling]

It goes like this:

I'm trying to automate the process of generating the calendars. The key word here is "automate".

My goal (it remains to be seen if I will reach it) is to set up an online service to generate just this sort of calendar on the fly. I intend to use software called FPDF for this. (Maybe this is the wrong software. I don't know. But I'm not going to spend hundreds of $$$ for something that will turn out unusable anyway.)

Now. I'm in America. We do not use A4 paper here. I originally designed my calendars for A4 paper, and I found out the hard way that nobody around here has it. Another thing I found out is that dealing with available paper margins is a pain, and a rather expensive pain if you go to get your calendars printed and get them all off-center because you went closer to the margin than their machines are set up to support.

So. I am planning on setting up a service where the user specifies what paper size she has available, and how close to the edge of the page it will print. I will probably have a test page or two so she can see how her printer deals with margins. She will of course also specify what time period she wants the calendar for, which (if any) region's holidays she wants on it, etc. Maybe there will be an option for her to choose the typeface. (I like DIN, but maybe she doesn't. Any ideas, folks?*) And as the calendar will be generated without human eyes to look it over, I need a way to get character size and position from point size and position.

* I cannot use typefaces with weird licensing requirements. I don't want the user of my site to have to specify how many copies of the calendar will be printed, whether their use is commercial or non-commercial, etc. (What if the calendars are being used by a business for the business's internal use? Is that commercial or non-commercial?) I'm not a lawyer and I can't afford a lawyer. So I must use fonts without strings attached.
In an ideal world (ideal for me, at least), the typeface designer would be paid a fair, flat fee for their labor, and then the typeface would be released into the public domain. But this is not an ideal world.

[JamesM]

Why not just offer a selection of several fonts, and then predetermine (by manual testing) what size works best with each of those fonts?

[Nick Shinn]

In an ideal world (ideal for me, at least), the typeface designer would be paid a fair, flat fee for their labor…

Ah-ha, Dumpling is a communist!

@Hrant: …if you want a block of text to harmonize with the rest of a metric-system layout, the point is non-ideal…

I always set the page size units to be in picas, whether metric or imperial.
Then the margins and gutters come out to be manageable integers.

Here is a Letter-size layout I did with horizontal metrics in picas of:
3 + 14 + 1½ + 14 + 1½ + 14 + 3 = 51

A comparable layout for A4, in centimeters would be:
1 + 6 + 0.5 + 6 + 0.5 + 6 + 1 = 21

Not bad, but needs a bit more width in the margins and gutters, I would say.

The body type is 9/12 and the headline 24/30 in points.
That’s 3 mm on 4 mm (2 on 4 if one goes by cap height), and 8 mm on 10 mm.
Not bad, but the difficulty arise for the “next” body sizes.
If I had decided that the leading should be less, it’s easy to take it down a notch (or half) in points—9/11 (or 9/11½)—but that would be 3 on 3.8 (or 3 on 3.9), which makes figuring out vertical metrics of the page more complicated.

Perhaps I am wrong on this; I’d be interested to hear from designers who work in metric how they handle similar layouts—I might even theorize that the units used have some influence on cultural style.

[John Hudson]

I always define my page layout using pica/point measurements, because I work outwards from the text rather than inward from the page dimension. If the paper size is metric, resulting in fractional points, I absorb these in the inside margin so that my text block and outer margins are integer measurements.

[JamesM]

I too find that picas work well for page layout.

[hrant]

Robert, it sounds like -if you want to do this right- you need to write a plug-in for InDesign that accesses a font's metrics information.

In an ideal world (ideal for me, at least), the typeface designer would be paid a fair, flat fee for their labor, and then the typeface would be released into the public domain.

That does happen sometimes (it's happened to me, with Arasan) and it's called commissioning a custom design! But it's not cheap (because it shouldn't be).

I work outwards from the text

When you have freedom of paper size (like I had when I made my letterpress business cards) that's a wonderful thing. But more often the paper size is pre-determined, and you can't just ignore that.

hhp

[quadibloc]

@russellm:
I like to measure with barley corns whenever the opportunity arises.

Well, as there are two (Selectric Composer or laser printer) picas (and, hence, 24 points) in a barleycorn, that isn't too confusing.

[John Hudson]

Hrant: But more often the paper size is pre-determined, and you can't just ignore that.

Of course you can't ignore the paper dimensions. But you can start from those dimensions and work inwards, as typified by modernist grid systems, or you can start at the level of letters and words and work outwards. Practically, there's always some back-and-forth, but as first principles they embody different design philosophies. If the first question I am asking is 'What size should the text be in order to be comfortably read in X typeface?', this implies something quite different from first asking 'How should I best subdivide this A4 sheet for the purposes of this document?'. Both are important questions that need to be asked, but the order in which you ask them is also important.

[George Thomas]

@Robert --
* I cannot use typefaces with weird licensing requirements. I don't want the user of my site to have to specify how many copies of the calendar will be printed, whether their use is commercial or non-commercial, etc.

Such comments tell me there are a lot of things besides the point system you don't understand. Personally I think you are setting yourself up for an expensive FAIL because of all the things you don't know.

Forget lawyers; hire an experienced production artist, even as a consultant -- someone with about five years or more of experience. If nothing else you will learn a lot from them.