Well, just found interesting statistics about font usage of different types.

Typographic Design Patterns and Best Practices by Smashing Magazine.

This may appear to be a statistically valid survey, but as the original choice of sample is somewhat self-selecting ("50 popular websites on which typography matters more than usual"), that rather undermines the results...

totocaster said:
> So please, let’s discuss.

no please, let's not.

you are unable to find "an exact answer"
because there _is_ no "exact answer"...

some people prefer a sans serif face
and other people prefer a serif face...

some people prefer a sans serif face
at small sizes, and serif at big sizes.

some people have a serif face as their
favorite, and a sans serif face as their
second-favorite, and see no conflict.

some people like whatever face came
standard on their current machine...

some people like whatever face is
used most in their local newspaper.

some people can't tell the difference
between a sans serif and a serif face,
even after you've explained it. twice.

many people who do know don't care.

you are an eskimo looking at snow...

so congratulations on your ability to
differentiate 87 different kinds of it.

but ordinary people just call it "snow".

-bowerbird

p.s. i was just kidding when i said
"no please, let's not". if you want to
discuss it, please feel free to do that.

I'm not searching for standard variation of serifs and sans-serifs. I'm just trying to find best font type (for most users of course) which is best for huge articles and books. For example here are some simple questions which I'll try to answer myself later (after an experiment), but I'm also interested in your opinion, that's why I'm here. To share my experience and listen to yours.

So the first park of my questions:
1. Which font type reads faster? Sans-serif or serif? (I'll read a A4 paper in English in both font types and write down time here)
2. Which font type is more comfortable to read? (reading long peaces of text without getting tired)

Please join me if you are interested.

Thanks.

I don't think it's a question of personal preference at all but a question of inherent legibility. Sans serif is good for short bits--a couple of lines of caption material, a few words on signs. It has 3 inherent problems that militate against legibility. 1, It has no serifs, therefore nothing to 'sacrifice' against the erosive effect of high contrast (as in black type against a white page). Hold your index finger up against a strong light source and you will see that the light erodes the outline of the finger, 2, Its lines have no 'finish,' no fixed terminal point, so that they bleed off into nothingness on the page while serifs provide a definite ending to each stroke, so that each letter is 'closed' and self-contained. 3, There is little or no variation in line weight (thickness), therefore none of the 'rhythm' that seems to somehow naturally lead from one letter to the next. All of these failings make sans serif less comfortable to read than Roman.

San serif is frequently found in museum and art-show catalogues, usually set at extremely wide meansures and always hard or even exhausting to read.

I have spent 40 years in magazine publishing and have found little use for sans serif in body type except when the type is very small and the layout cramped. Also, when paper-quality is low, like the rubbish we used for the listings of TV Guide magazine when I worked there too many years ago (the type was Bell Gothic, developed, I think for Bell Tel's phone directories).

Sans serif has been around for a long time yet, for long passages, has found few admirers save for art directors. They like it because it is relatively unasserttive compared to Roman and therefore more easily integrated into the design of the page as a whole.

If you are setting long stretches of type and want people to read to the end of them, choose Roman and don't look back. It will also help to provide compelling prose.

william said:
> I don’t think it’s a question
> of personal preference at all
> but a question of inherent legibility.

you'd think it would be very easy to prove, then.

and you'd be wrong...

-bowerbird

Just look where serif faces are used the most. There's a reason for that.

Then look where sans serifs are used the most. Probably a good reason for that, too.

There you go. That's probably your safest bet.

What I see, Palatine, is Roman used mostly in long stretches of book and magazine prose--and sans serif in small sizes on cheap paper (phone books and the disposable like).

Bowerbird, it IS easy to prove by measuring reflectance/contrast, although that's not very satisfying. Using actual prose pieces would be interesting indeed, but prohibitively expensive and complicated, requiring identical texts in two settings divided between two pretty large groups of people who read often for entertainment and information to read them for say three hours.

The results are perfectly predictable. The Roman readers will begin by immediately congratulating themselves on their good fortune, finish their texts quickly and conclude by devoting their remaining time to amicable and rewarding contextual discussion. Not so the san serif readers, oh no! They will begin by cursing the day they were born. end of Hour One: mass eye-rubbing, painful headaches. Hour Two: bitter recriminations lead to mutiny. Hour Three: fisticuffs and hair-tearing culminate in hostage-taking incidents.

palatine said:
> Just look where serif faces are used the most.
> There’s a reason for that.

the reason is "tradition".

> Then look where sans serifs are used the most.
> Probably a good reason for that, too.

the reason is "tradition".

> There you go. That’s probably your safest bet.

the safest bet is always "tradition".

***

william said:
> Bowerbird, it IS easy to prove
> by measuring reflectance/contrast,
> although that’s not very satisfying.

i'm not nearly as confident as you seem to be
that it's easy to measure "reflectance/contrast",
not in the wide variety of environments needed.

but we do agree that it wouldn't mean anything,
so i guess we don't have to split hairs...

> Using actual prose pieces would be interesting
> indeed, but prohibitively expensive and complicated,
> requiring identical texts in two settings divided
> between two pretty large groups of people who
> read often for entertainment and information
> to read them for say three hours.

well, except that many experiments have been done.
and their aggregated results are anything but clear...

but if you want to do your own iteration, i'd suggest
that -- in addition to randomly assigning the subjects
to a serif/sans groups -- you also should assign them
(randomly, of course) to a "choice versus no-choice"
variable, where the "choice" group selects their font...

my hypothesis would be that the "choice" group will
register most positively on your dependent variables.

notice how i have cleverly dodged the thorny question
about how you would define your dependent variables;
cursing, headaches, mutiny, and fisticuffs are just fine... ;+)

-bowerbird