> I assume over 600 pairs become unmanageable
Actually that's a pretty modest number these days, not even counting class kerning! Even just for basic Latin support I tend to go around 1500 "manual" pairs (and not because I'm unhappy with my "base spacing"). BTW, some years ago certain apps had a problem with over ~2000 pairs, but that seems to have died off.
Sorry, Hrant: I meant 6,000 pairs (buffering to 5,120).
"I tend to believe that web crawling will return much more typical pattern redundance"
I suspect that it's relative to a degree. If the font will be used online, then online language usage would be "typical". If it's a font for the next luxury edition of the Bible, maybe not. (Not that I'm implying it is!) :-)
But the influence of context/media on actual letter combinations may well be so minute that it doesn't really matter, at least not for locating kerning pairs. Dunno.
"If the font will be used online, then online language usage would be “typical”."
I was speaking about "neutral" flat web crawling, neither font nor content oriented (you know, "for a semantic web"). My unspecific induction is that because of the relative degree of standardization in language used in web pages, and the over redundancy of certain formula (let's say, : "home", "Google", "search"), if not filtered, may drive to quirky results.
About the Bible (which version?), there are numerous comprehensive term searching by there: surprisingly, some words one could suppose usual in it, are not.
Beginnig by the g, look at where we arrived... :-o
Here it goes the last version of my balanced g, spaced in context.
First of all, the basic "nngnnoogoo" for fixing –and fitting– a standard across the font.
Then, the beloved "login burger" (sorry, I'll release more characters soon).
Just for testing, the double g by repeating the glyph (not the gg glyph itself I'll publish tomorrow).
I realize that thanks to Hrant influence about the open loop, and following the comments and suggestions of valued critics I don't need to mention (see above), I arrived to an unsuspected g I like the most.
Thank you, guys!
It looks fantastic! Now you'll need a name for the typeface that has several g's in it, to show off your handiwork.
Looks nice! Please show it around to the open-"g" nay-
sayers - I think you'll have more of an effect than me.
Some people just need to be knocked over the head with
good hard proof; they can't just smell a good idea.
Spacing: it's hard to be sure onscreen, but I think you
need to tighten the left a hair, and kern the "g"-round
Name: Well, at least one "g" for sure.
What about "Apogee"? Or "Pedigree"?
Apogee's already taken, and Pedigree sounds too much like a dogfood. Besides, it ought to be a Spanish name of some sort. Perhaps Perigeo (it's more down-to-earth than apogee, anyway).
Finally, I can write this:
Here there are some preliminary sketches of the lowercase alphabet, just to show you the concept:
I guess I have a lot of things to adjust yet in that rough version, but wanted to introduce all small sisters at once.
So, what's the name of the font? ;-)
Work in progress...
Forget the anemic y I pasted above. That's the one I drew:
Coming along nicely. Maybe those tails of a, d, and u will become the next little detail that spawns a long thread.
Craig: Let's see the funny tail in d-tail...
I listen comments.
PS: I must standardize the lenght of these tails if I wanna go on.
> I must standardize the lenght of these tails
Not really. Standardization is a means not an end.
Right, Hrant. It was a mistake (not a mistype).
I meant to standardize the usage I did, not the length: for some reason I can only blame to my distraction, in the first "dua", the tail of the d and the a actually don't align with the stem; it is fixed in the second one and I think it works well now.
BTW, have a slightly corrected version of "dua" I guess is somehow improved.
I thought I'd return the favour Marcelo.
This is looking good. It has a touch of Slimbach about it, but that's no bad thing. I can see why you altered the g - numbers 3 and 4 on your initial post are very Minion Medium. The latest incarnation bothers me but I can't quite put my finger on the reason. I quite like the idea of the tail serif, but maybe if it was not so angled; more to the vertical but not completely.
Slimbach? Hmmm... It's hard for me to find similarities.
Mi first post was only an attempt to catch some attention on the quirky loop, where I had a couple of uncertainties. The broken link appears in many faces. Hoever, it was not my aim to post an original new shape, but to trigger unconscious of forgotten questions, both in me and in the critics.
After all, I think I succedd in that matter.
I'll make some tries on a quasi vertical serif in the tail (actually, in my post of the 19.Feb.2009 6.20am you can look at a previous one).
Thanks for your ideas.
The UPPERCASE set @ its current state, which is to say, budding:
The next step is softening the strong hardness of the strokes, though without losing the sharp look I want to preserve in the font.
(Going to rest)
Those 'dua' improvements are good, I think.
On the caps, I suspect a challenge with such bold serifs will be to get them to work on the thin strokes without overwhelming them. AKMNUVWXY (Actually, the diagonal thins make it easier, as they flow into one side of the serif, so the real challenges are MNU.) Maybe cupping the serifs more will help here and will also "soften the strong hardness of the strokes," though perhaps that would lose too much of the sharpness you want to retain.
Is the slight overshoot of the left side of the tail of the Q intentional? Also, due to the concerns Craig raises, I wonder if the M shouldn't have splayed legs.
In the UC:
- "J" is too hefty in the bottom.
- That thing jutting out in the "Q" is cool, but way out of character here.
- "L", "S", "U" and "Z" are too wide.
It's nice to see your y is on a similar path to mine! :)
Did you change the bowl of lc a as well? I prefer the last two above.
I think L might be too light? And the lower leg of the K.
>softening the strong hardness...without losing the sharp look.
An interesting difficulty to solve!
I must confess that when I encouraged you to pick the y I liked more, I was secretly validating my concerns on the bottom of mine. Anyway, I guess both are very personal and have their own character for not being related, except as "cousins in law" ;-)
Your "democratic" y leans on a slanted serif while my "republican" one is abosultely conservative and sets its whole foot on the floor.
Probably the L is more wide than light. The leg of the K needs to be more N, R alike.
I was going and coming around the cupping: actually, my first try was a deep cupping that seemed to my eyes very distractive. I think that because of hinting (or unhinting) the current one does not look as curvy as in hardcopies (it will be a defiying matter for the display version).
It's a quiz in which I'll have a hard battle until arriving at the fair point.
About the Q, I was kidding for a while, making time to decide if I melt the tail in the bowl or split the stroke away.
My M will be definitely legs-splayed, a decission I was lazily avoiding to make.
My J rised its bottom until reaching the baseline.
I'm not sure about the width of the entire UC, since they are in an very early approach.
J: Very nice, but I worry about the "LJ" pair (which is admittedly rare - but Ljubljana is a capital! :-). I actually prefer the descending form, it's just that the one you had was too aggressive. Or maybe make OT alternates?
Q: I'd certainly encourage you to make it as little like an "O" as possible. This might sound impossible, but it's actually achievable: use the tail to cut into the "O". Here's a design I drew long ago, but have yet to actually turn into a font (not feeling qualified yet):
In the meanwhile, let me know what you think about that couple:
I like the B slightly overshooting up and down.
I have a concern about solving accute joins in order to prevent optical distortions:
Oddly I like both options. :-(
Leaving a gap (like in the "B") is much less work, and some would argue is good enough. But I personally prefer what you did in the "v" (just more of it), especially in such an organic design.
Sorry, Hrant: I was writing while you posted the last comments.
About the LJ, that's what happens with the J as is (my keyboard doesn't allow me to enter both Cacute and Ccaron):
Thanks for the link. The Microfoundry is always an exciting experience.
Let's see the Q.
Here I made some quirky sketches I don't know why I find interest in.
[Thanks to previous strong Hrant influence ;-) my counters now tend to be open...]
I think the first two match the character of the typeface better. (And open up the P now :-)
Having fun with the Q in context:
There are some rought adjustments in stresses and junctions I'll return on later.
"And open up the P now :-)"
Well, the "Q" can afford to be flamboyant - but of course it's possible to go too far, and I think the 4th one maybe does. Or not. :-) The first one can work, but that gap is too tentative - you might try curling the right stroke inward. The second is of course safest, but maybe too safe - a lost opportunity; but if you go with it you might try pointing the terminal upward, moving the cupping to the right side? If that works, you might try that on the first one too.
As you might suspect my favorite is the third - but it needs work: the lefthand join is too abrupt, and the tail needs weight and definitely stroke contrast, most likely in the form of making it increasingly thick (and then tapering quickly).
I guess I'll walk on this Q design to toe the line of my font:
Nevertheless, I'll keep my faithful to this alternate one (may be even shy a bit):MarS
I personally think the Q in REQUEST is rockin'!
I have to agree with Steve (especially for all-caps setting).
Or I'll add another alternate Q:
(Just kidding: I wanted to keep records of my undeliberate thoughts)
[Must not abuse OT features]MarS
"I personally think the Q in REQUEST is rockin’!"
Do you mean it's rocking gently from side to side as a boat? ;-)
I think that will be the main Q of my font, with minors optical adjustments.
It seems to work fine in both UC and lc contexts.
I'll keep the two alternates I posted latter.
How does it look?
I mean rockin' as in Soda Stereo. :-)
Your Q from 2:01 would be interesting in a Preissig-style typeface. (Keep track of these ideas for future typefaces!)
I don't know if this is intentional, but the Q (and the O too) is leaning to the left. :P
I find the 2:01-o-clock "Q" strangely intriguing. It looks like it's going to jump away. :-)
Though the one you have now is really nice I think.
I don’t know if this is intentional, but the Q (and the O too) is leaning to the left.
That's typical of Garalde letterforms.
"I don’t know if this is intentional, but the Q (and the O too) is leaning to the left. :P"
Jelmar, actually, the counter in curvy shapes is intentionally rotated 3 degrees to the left, to keep some calligraphic taste and to avoid "hyperelliptization".
Though it's more noticeable in the O and Q, you can find the axis is also slanted in the C, D, G, b, c, d, e, g, o, p, q, for example.
Usually Humanist (Jenson Pro by Adobe) and Garalde (Caslon Titling by Monotype) faces share this feature; I somehow prefer the criterion adopted in Jenson, while Caslon has the counters rotated inversely, that is, 3° to the right.
It looks like it’s going to jump away.
Yeah, I thought of this guy when I saw that tail.
Must be cartoon week on Typophile. :->
[Don't tell anybody, but I hate when, after reached 80 comments, a forum is split in pages, losing some usable features (new comments are no more marked with the red-backgrounded "NEW", clicking in the "xx new" links in the forum doesn't bring you to the new, but to the start, and so on), and making it hard to go to the intended place. Even the RSS feeds don't work as they are supposed. Or may be it's me? :-S ]MarS
Same here, Marcelo. At least it's back to 90 or so; up until recently the page split happened at something like 30 posts to a thread. I think it's a server performance thing, but it's still kind of annoying.
Yeah, I thought of this guy when I saw that tail.
Must be cartoon week on Typophile. :->
My Queuekachu is ready to jump the queue. So, what?!
After refining the animé, here they are, numbers:
I'll need a name, sooner or later.
> actually, the counter in curvy shapes is intentionally rotated 3 degrees to the left, to keep some calligraphic taste and to avoid “hyperelliptization”.
Though it’s more noticeable in the O and Q, you can find the axis is also slanted in the C, D, G, b, c, d, e, g, o, p, q, for example.
Yeah, I did see it in those other characters, but it isn't so apparent. The O and Q do jump out a bit. I just thought I'd point it out, but apparently I'm somewhat ignorant about the fact that it is normal for Geralde typefaces. :)