Working title "Exactica"

ValKalinic

Hi all, 
I've been working on this idea I had about combining a strict grid with rules for radia; from a few specified ones, a radius is chosen based on the specific situation (end of stroke, lowercase curve, lowercase intersecting curve, uppercase curve, uppercase intersecting curve).
For example, look at m, B, or 8, and the multiple different radia occurring inside those single characters. 

The goal of such a system would be creating an impression of precise industrial machine-etched type. 



This is of course very much a work in progress and still early at that, please consider it as such. I will increase the weight difference between upper and lower case; some characters are not quite there yet; I have to think about the way MNVWXYZ fit in with the rest. 

I've had to pause the work on this for two months and now I'm resuming it, so I thought it would be a good time to hear what advice or criticism someone more knowledgeable than myself could offer. Thank you.

Here are some current samples:








And a chunk of text:

ValKalinic

Notice any glaring mistakes I've missed overall?
Is anyhing sticking out too much?
Some other advice on taking this further? Any suggestions for a name?
Thank you!

K Pease

My strongest recommendation is to design W anew to not resemble M in any way.

The crossbar of F may have to be shortened, as it is giving the impression of being longer than the top. You might carefully consider whether to drop the center of P down to where A is.

Christian Thalmann

Spacing looks a bit too tight IMHO, except for the right side of /l/. I would probably lower /p/q/ to the baseline. The idea certainly makes sense for /g/y/, but feels a bit exaggerated there. Might profit from breaking the strict grid... you don't need to tell anyone.

Theunis de Jong

The /g and /y may work better with a flat bottom instead of the current curved one. Then you can also shorten /p and /q by a single one of your units.

Admittedly, it would make them go less deep than the /j, but the /p and /q are a bit too deep for my taste anyway.

Christian Thalmann

Theunis de Jong said:

Admittedly, it would make them go less deep than the /j, but the /p and /q are a bit too deep for my taste anyway.

Or shorten the /j/ too. Shallow descenders are a sign of good taste, just ask Hrant.

Paul Miller

The X and the x look a little ambiguous, they may be confused with H.

K Pease

It's natural for us, when examining the descending letters up close, to suggest the obvious ways to make them less unusual, but I find that, as they are, this pronounced differentiation is most effective for reading text. But only if you expect that to be how the face is used.

Adam Ladd

I like the attempt for a unique take on the /4 but to me it feels uncomfortable to view/read... perhaps making it a little more conventional? And I wonder if the tail on the lowercase /l is a little long? When read in the word "failing" in your above chunk of text, it creates a pretty large gap between the /l and /i.

John Savard

In general, I think it looks nice. My first reaction was, oh, there are so many typefaces in this general category these days - but that is true of any category of typeface, so I quickly woke up and realized you should not let that stop you.

However, the general philosophy of this specific typeface has led you into trouble in one place. The letter X, both upper-case and lower-case, is not swiftly recognized, reducing legibility. Can this be addressed without reneging on the rules for curve radii that give this typeface its temporary name?

I see in both upper and lower case V, and capital W, there are diagonal lines allowed, so I have hope you can make a more legible X yet be true to the philosophy of the typeface.

I attempted to illustrate what I was thinking of, however, and showing your original X on the left, followed by my attempts, it is clear they are all dismal failures.

ValKalinic

Hi @John Savard
Thanks a lot for your insight

 there are so many typefaces in this general category these days

What would be that category, the overall look? Is there any name/tags for it? Asking so I can find examples and learn more about it.

The letter X, both upper-case and lower-case, is not swiftly recognized, reducing legibility.

Thanks for trying it out. Absolutely! It bugged me a little, but after @Paul Miller pointed it out, I noticed how strong the impression is. Working on it right now. I think a big clue to solving /X/ of this font is what the natural way of writing it is; two separate diagonal strokes, instead of two curves. So the letter should be more in tune with the Z than the U. Tuning and trying out! Thank you both!

ValKalinic

@Theunis de Jong @K Pease @Christian Thalmann
Thank you for your suggestions! I'm now tuning the ascending and descending characters:
 

Any thoughts on this direction? I feel these shorter and simpler descenders are much more pleasant! Still looking at lowering /p/q/g/y to baseline, not sure yet.

ValKalinic

Adam Ladd said:

I like the attempt for a unique take on the /4 but to me it feels uncomfortable to view/read... perhaps making it a little more conventional? And I wonder if the tail on the lowercase /l is a little long? When read in the word "failing" in your above chunk of text, it creates a pretty large gap between the /l and /i.

Thanks @Adam Ladd. Absolutely, among the numerals, /4/ sticks out from the set.
Trying out a few paths now, but not quite there still.



As for the /l/ tail problem, I didn't notice it before you pointed it out, but now I see how distracting it is. Trying out a shortened one now. I've also matched the /t/ to it, but not sure about that one:

K Pease

A concern for numerals, which also tends to inform the aesthetic of "machine" styles such as this, is that they should be different enough never to be confused for each other in conditions of poor visibility, damage, or even deliberate falsification. So, as fine as your 3 looks, try laying both its terminals flat so that it doesn't fit exactly into 8.

Christian Thalmann

The /p/q/ cause the baseline to dance up and down; very distracting. The /g/y/ might have a similar problem, but it's less pronounced — would have to see it without /p/q/ enhancing the effect.

As for /t/: Try making the left spur shorter and extend the crossbar and the foot to the right instead.

Adam Ladd

The /l seems better now. Agree with Christian about the /t... feel it is ok to have wider/longer.

Hmm, yeah the /4 is tough, not sure any of these are clicking yet like you said. Maybe something less vertical on the left stroke and perhaps with more diagonal to it? Would take some playing. The right stroke may need to be a little taller in some cases.

Paul Miller

John Savard said:

I attempted to illustrate what I was thinking of, however, and showing your original X on the left, followed by my attempts, it is clear they are all dismal failures.

I had a little think about what I would have done given this problem and came up with this.

ValKalinic

Paul Miller said:

John Savard said:

I attempted to illustrate what I was thinking of, however, and showing your original X on the left, followed by my attempts, it is clear they are all dismal failures.

I had a little think about what I would have done given this problem and came up with this.

Thank you both very much @Paul Miller and @John Savard
This is the current state of /X/ and /x/, both intended to look as composed of originally two diagonal strokes, that are interpreted in the style of this typeface (straight beginning and end, and matching radius):

and in context of XYZ:


What do you think about it now?

Paul Miller

A big improvement.

ValKalinic

Paul Miller said:

A big improvement.

@Paul Miller thank you for your help!

For those interested, here is the current state of the basic set for further criticism, as well as a map of all the glyphs so far in the attached pdf:

Thomas Phinney

The top right bar of the “t” should stick out as much as its bottom foot.

The new K and R are not really working as yet.

ValKalinic

Thomas Phinney said:

The top right bar of the “t” should stick out as much as its bottom foot.

The new K and R are not really working as yet.

Thank you for the advice.

Agreed, I'm trying out a few different approaches.

Ray Larabie

When you're going with a technical style you're showcasing a rule. The reader looks at the words and perceives the rule. In this case "a strict grid with rules for radia". When you need to break those rules, I recommend thinking about bending those rules within a consistent threshold of imperceptibility. How much are you willing to go off grid and to change the radius or a corner? Figure out how far you're going to bend the rules and then do that consistently across everything. If the rules are being consistently bent to a certain threshold and you need to go beyond that, it's more likely to go unnoticed. If you stay rigid to the rules and bend them too far, it jumps out at the reader.

For example; if you have a low resolution pixel font and one of the letters goes "off-grid" like an M or a W, that rule break could be easily noticed. If the rest of the font is barely perceptibly already off-grid, the M or W is more likely to harmonize.

Another example: picture a classic monospaced typewriter font. Except one letter isn't monospaced. There's a wide m. The gaps in the m are identical to those in the n. That's going look odd and stand out to the reader. Now look at ITC American Typewriter. The monospace rule is broken all over the place so we don't even notice that the m is wide. The rules are bent just enough that the reader still perceives it as a typewriter font.

When I look at this, the and H and U look wider than the O. The top of the B looks larger than the bottom. and the "crossbar" looks thick. If you optically adjust the crossbar like Helvetica, that's probably bending the rule too far. But you can give it a subtle nudge higher and thinner to a degree that it doesn't spoil the effect of the rule. If you keep making those adjustments all across the typeface within a consistent perception threshold, you'll have more flexibility in dealing with troublesome characters like K, R, V and X.

ValKalinic

Ray Larabie said:

When you're going with a technical style you're showcasing a rule. The reader looks at the words and perceives the rule. In this case "a strict grid with rules for radia". When you need to break those rules, I recommend thinking about bending those rules within a consistent threshold of imperceptibility. How much are you willing to go off grid and to change the radius or a corner? Figure out how far you're going to bend the rules and then do that consistently across everything. If the rules are being consistently bent to a certain threshold and you need to go beyond that, it's more likely to go unnoticed. If you stay rigid to the rules and bend them too far, it jumps out at the reader.

For example; if you have a low resolution pixel font and one of the letters goes "off-grid" like an M or a W, that rule break could be easily noticed. If the rest of the font is barely perceptibly already off-grid, the M or W is more likely to harmonize.

Another example: picture a classic monospaced typewriter font. Except one letter isn't monospaced. There's a wide m. The gaps in the m are identical to those in the n. That's going look odd and stand out to the reader. Now look at ITC American Typewriter. The monospace rule is broken all over the place so we don't even notice that the m is wide. The rules are bent just enough that the reader still perceives it as a typewriter font.

When I look at this, the and H and U look wider than the O. The top of the B looks larger than the bottom. and the "crossbar" looks thick. If you optically adjust the crossbar like Helvetica, that's probably bending the rule too far. But you can give it a subtle nudge higher and thinner to a degree that it doesn't spoil the effect of the rule. If you keep making those adjustments all across the typeface within a consistent perception threshold, you'll have more flexibility in dealing with troublesome characters like K, R, V and X.

@Ray Larabie thank you for taking the time to explain that view, very insightful. I will thoroughly evaluate all the glyphs again with such an approach in mind.

John Savard

Personally, in line with Ray Larabie's comment, I thought that unlike the X, the original K and the R weren't broken, and so "fixing" them led to the typeface no longer embodying its rationale of being designed according to constant rules as much.

Thus, he has a valid point, but to apply it will require thought, the right way not always being obvious, as the question is, where to strike the balance between keeping what makes the design unique, and ensuring people can easily read text in the face.

But he also makes a different point: that if the typeface makes only one exception to a rigid rule, instead of relaxing it everywhere a bit, that will be glaring. Looking at your typeface from that point of view, but only briefly, I saw only one thing that appeared to clash seriously with that rule: capital M and N as against lower-case m and n. And of course, it really didn't, since the forms of the capital and lower-case are genuinely differnt, and the rules were applied correctly to each - perhaps there is something this hints at that could be adjusted, whether in those or other letters, but as likely not.

My view is that instead of changing the K and the R to attempt to embody this principle after having changed the X... the thing to do is to stick with the original K and R, keep the changed lower-case X as is, and change the upper-case X so that the crossbars are angled like those in the lower-case X, thus keeping a bit more to the original "grid" aspect, while also not having the legibility issue of the original X.

I found Thomas Phinney's insights spot on. In my own simplicity, I'd have been as happy to have the top right bar of the "t" shortened to match the top left bar as to have had it lengthened to match the bottom, but I would defer to someone clearly far more knowledgable than I am.

ValKalinic

John Savard said:

Personally, in line with Ray Larabie's comment, I thought that unlike the X, the original K and the R weren't broken, and so "fixing" them led to the typeface no longer embodying its rationale of being designed according to constant rules as much.

Thus, he has a valid point, but to apply it will require thought, the right way not always being obvious, as the question is, where to strike the balance between keeping what makes the design unique, and ensuring people can easily read text in the face.

But he also makes a different point: that if the typeface makes only one exception to a rigid rule, instead of relaxing it everywhere a bit, that will be glaring. Looking at your typeface from that point of view, but only briefly, I saw only one thing that appeared to clash seriously with that rule: capital M and N as against lower-case m and n. And of course, it really didn't, since the forms of the capital and lower-case are genuinely differnt, and the rules were applied correctly to each - perhaps there is something this hints at that could be adjusted, whether in those or other letters, but as likely not.

My view is that instead of changing the K and the R to attempt to embody this principle after having changed the X... the thing to do is to stick with the original K and R, keep the changed lower-case X as is, and change the upper-case X so that the crossbars are angled like those in the lower-case X, thus keeping a bit more to the original "grid" aspect, while also not having the legibility issue of the original X.

I found Thomas Phinney's insights spot on. In my own simplicity, I'd have been as happy to have the top right bar of the "t" shortened to match the top left bar as to have had it lengthened to match the bottom, but I would defer to someone clearly far more knowledgable than I am.

@John Savard thanks a lot for your insight, it helped me sort some things out. Still working on the font and improving things, will make an update when it gets to a certain point. 
Thank you!