New font working title: 'Software Developer'

ebensorkin's picture

This font now has 2 editions, the one originally commisioned ( now being hinted a bit more) & an improved version with a bigger set of targets to hit. My focus is now the new version.

The original font was designed :

- to be used in bitmap mode at 12pt on a pc ( on & off pixels only )

The new design is meant to :

- look okay in print
- work as hited bitnmap ( on & off pixels only )
- and work anti-aliased too ( with grey pixels created by rendering engines ).

Originally I tried to shape my glyphs around the target bitmap at 12pt for PC. But I found out eventually that I was constraining the glyph designs too heavily and creating problems for the bitmaps at larger & smaller sizes. This new font builds on my trial & error experience as well as my new understanding about why monospaced fonts have taken the shapes they have. The latest version does as better job rendering to screen in antialised mode & I think is less ugly when printed.

The latest GIF graphic is simulated PC 12, 24, & 48 pts.

The PDF shows 12 & 72 pts. The 'in typophile' link below doesn't work. This one does.

I have been using Lucida Sans Mono, Lucida Console, Consolas, Courier, Andale mono, Monox, Isonorm, Thordis Mono, Adaptive Mono & Mono 821 as references.

Thanks!

AttachmentSize
example1.gif17.61 KB
example2.gif17.04 KB
example3.gif18.02 KB
sq9.gif24.22 KB
ampersand.gif962 bytes
ampersand.gif1.52 KB
softdevexample.gif25.63 KB
softdev.v3.1.pdf324.64 KB
softdevr6.gif13.41 KB
r6bt.gif2.25 KB
ebensorkin's picture

I was just on your site - microfoundry.com, but I didn't see Mana there.

found this
http://store.adobe.com/type/topics/theampersand.html

Also, interresting for my font is the fact that : In some computer programming languages, the & sign is often used to indicate logical AND. Many computer languages with syntax derived from C differentiate between: && for logical AND & for bitwise AND

from http://en.wikipedia.org/wiki/Ampersand

Is 'open headed' a common description of a common feature or something you have invented?

hrant's picture

Don't look at my site. Please. :-\
You can see Mana at http://www.ultrafonts.com
Click on one of the Manas, hit "Custom Sample", and type it.

"Open-headed" is not a common term I guess, but probably because you don't see many open-headed ampersands. So I guess it's "invented" (in the most lowly usage of the term) but I thought the meaning would be clear: an ampersand with a hook for a head, sort of - the right side open.

Hey, wait! Just like your #8 - but with a conventional lower body.

hhp

ebensorkin's picture

Riiight. I see that now. This is food for thought...

ebensorkin's picture

I haved several new examples of ampersands posted now. I think 22 is maybe the best bet now. But I am open to arguement. (See the links at the top of the post)

Miss Tiffany's picture

Eben, I'm not a programmer ... but, I would prefer to see non-aliased type. This is one thing I actually do miss about OS9.

While I can see the old-style figures as a nice _design_ feature I don't see how they could be useful to a programmer. I suppose you could have two different sets for tabular settings and paragraph settings both.

The para-setting which you've given is set very tight for the mono-word spacing. Is this typical for programmers? If it is am I wrong to ask if the word-spacing can be tightened up? That sounds wrong now that I've written it. If it is mono-space you shouldn't play with that. Can we see a setting that isn't so tight? I have a feeling that lining figures will still look better with this mono-space in a paragraph.

How much control do you have over the aliasing?

Did you tighten the spacing up in your ampersand samples? Overall I'd say all of them need a pixel of space on both sides. 9 16 17 22 all strike me as possiblilities. 9 seems the most logical. Hmm... 6 is good too.

ebensorkin's picture

Hi Tiffany,

Thanks for the feedback!

I should probably clarify about a couple of things.

Aliased or not is a decision made by the operating system and or the program you are running. The fonts I am making ( now there are two - kind of big & small) are meant to be used by programmers whose tools use the non-aliased rending of the font. That said the fonts can be used in an aliased fashion if you want- and my goal is to make them look as good as possible there once their core use ( on screen un-aliased ) is addressed.

Originally the forms were just blocks to correspond to pixels at a given point size but now they are smooth shapes - like Courier, or Lucida console - and I am using True Type hinting to try achive good results at sizes 9-16 or so. After 22 points the font is going to be aliased by the programmer tools/OS/video card anyway.

The spacing bewtween letters is an interesting issue. Having two pixels between letters when they get as small as my smaller font isn't practical at 12 pt - the space looks too huge and interferes with reading words and code. But eventually as the font scales up I have found it that having two pixels between at certain sizes makes sense.

For the bigger size I am using two pixels at 12pt ( one on each side ) but only at certain sizes because as the font shrinks it starts to need to consume one of the pixels in order to kep the forms legible. From 12 (2pixel) say to 9 (one pixel) this happens. At small sizes there are all kinds of trade-offs to make.

The Ampersand example is from the smaller font hence the tightness.

I am convinced now that the lining figures may not be for everybody. So I will be making versions with & without. I have found that they are nice in texts but don't do much for coding... but that's a matter of taste.

As far as other kinds of control over aliasing - I suppose I could do what Hrant is doing & make a version that has individual pixel shades - but that doesn't suit my client's purpose, they want no aliasing & multiple pt sizes; so I won't be going that route with this font.

I did not tighten the spacing in the samples no. See above & you would not beleive how much people disagree about the Ampersands. Everbody likes a different one best! Twenty people & twenty different answers. I ended up going with 22 because it shrank to smaller sizes the best and was still fairly popular. The small gaps in the head 17 for instance are murder to hint as the font shrinks. The example I was showing is 12pt on a mac so you can see there are at least 2 smaller point sizes I was still targeting.

BTW - On my powerbook running OSX I can go into System prefs & Appearance turn aliasing off from 12pt down to see what I am getting in this environment. You could do the same - It doesn't effect all of the OS & programs but it does have an impact. You might want to try it. Unless your not on OSX. I thought you were for some reason...

I hope I have explained this well enough ... Let me know if any of this doesn't make sense.

ebensorkin's picture

This font now has 2 editions, the one originally commisioned & an improved version with a bigger set of targets to hit.

The original font was designed :

- to be used in bitmap mode at 12pt on a pc ( on & off pixels only )

The new design is meant to :

- look okay in print
- work as hited bitnmap ( on & off pixels only )
- and work anti-aliased too ( with grey pixels created by rendering engines ).

Originally I tried to shape my glyphs around the target bitmap at 12pt for PC. But I found out eventually that I was constraining the glyph designs too heavily and creating problems for the bitmaps at larger & smaller sizes. This new font builds on my trial & error experience as well as my new understanding about why monospaced fonts have taken the shapes they have.

The latest graphic is at 12, 24, & 48 pts.

A full character set will appear later.

ebensorkin's picture

The 'in typophile' PDF link at the top below doesn't work. This one does.

Mark Simonson's picture

Eben, This is shaping up quite nicely. I think the short numbers are an improvement over your original old style ones. This is the same solution Frutiger came up with for OCR B way back when, and it still makes a lot of sense.

There a lot of spots here and there where your curves could be refined without affecting how it works at low res. For instance, look at the top and bottom of the "O". These curves should be symmetrical vertically as well as horizontally. Or the "S": The stroke should be evenly modulated as it snakes around. As it is, there are some abrupt changes in the curve, like the inside top left and the top of the middle of the stroke, and this causes it to modulate a bit unevenly. The ends of the strokes on the "S" and some of the other curved characters seem a bit too tapered.

Now that you've got the basic shapes worked out, you should go through all the characters--especially the ones with curves--with an eye toward perfecting their shapes. Curves should flow evenly; symmetrical shapes should be perfectly symmetrical (taking into account optical effects as needed if they still don't look symmetrical).

The "A" seems a bit too wide--I measured it and it is wider than the "U". Perhaps that's intentional; if so, and it's noticeable, you've gone too far.

I can see that you've made your horizontals thinner than the verticals, but in some of the characters (the "I" and the "8", for instance), the horizontals look heavier for some reason. You may need to cheat some of these thinner to compensate.

That should give you enough to chew on for a while. I know you've spent a lot of time on this, and I think you've made a lot of progress. It looks pretty good at text size. You've still got a ways to go, but it's mostly a matter of refining things now so that it also looks good close up.

cerulean's picture

I like ampersand 3 best, the x-height one. It's classy. 22 is nice too.

ebensorkin's picture

Mark,

Thanks very much for the detailed crit! I have some questions still...

> This is the same solution Frutiger came up with for OCR B way back when, and it still makes a lot of sense.

I thgink I wouldn't mind making an altrenate font later with altrenate altrenate numbers but for my main purpose I think this is the better solution.

> look at the top and bottom of the “O”. These curves should be symmetrical vertically as well as horizontally.

I think I wanted to suggest the same kind of movement Luc(as) De Groot has in his monospaced fonts. I am probably just doing a poor job of it or doing it too subtly for it to seem deliberate. Have you seen the fonts I mean? If so, what do you think of that approach?

> The stroke should be evenly modulated as it snakes around. As it is, there are some abrupt changes in the curve, like the inside top left and the top of the middle of the stroke, and this causes it to modulate a bit unevenly.

Again, I wanted to emulate other type but am not doing it well enough, for instance monoco or Lucida Console have this modulation... Do you not think it appropriate for this kind of font?

>The ends of the strokes on the “S” and some of the other curved characters seem a bit too tapered.

Agreed.

> The “A” seems a bit too wide

No, I just missed this.

> (the “I” and the “8”, for instance), the horizontals look heavier for some reason

Agreed! Thanks!

> You’ve still got a ways to go, but it’s mostly a matter of refining things now so that it also looks good close up.

Thanks!

I have started making the fixes you describe but would you have a look at the diagonal axis thing in the o ( & the pbdg etc.) , and the modulation? I would like to see what else you have to say about this before I change this aspect of the face. It would be much easier to make the face more DIN-like & symetrical but I thought the most successful monos I had seen were degroot's and I wanted to bring some of his influence into my design... Again, What do you think?

I can send you some PDfs if you don't already have the references. I don't want to make you dig look things up!

Cerulean, Thanks! What do you think about the PDF?

ebensorkin's picture

Mark, here are some references for the modulation & the diagonal axis question.

http://www.fontfabrik.com/fofafon4.html

...and some screen capured letters from that page:

on
Op

Mark Simonson's picture

I see what you mean. But even in TheSans Mono there is perfect symmetry (or at least the appearance of it). It's just not 2-way mirror symmetry. Instead, it's got "pinwheel" symmetry--the kind of symmetry you see in a pinwheel, a swastika, or a yin-yang symbol. In your zero, for example, I see left/right mirror symmetry, but top/bottom asymmetry--sort of like Hobo.

Trying to emulate De Groot--you've certainly set a high goal for yourself. The Dutch school is worthy of emulation, to be sure. However, I don't know that I would be the one to tell you how to do it. As much as I admire it, it's not my bag. Sort of like asking a blues musician how to write electronic dance music.

ebensorkin's picture

Yes, it is a high standard that I am falling short of. And I am falling well short of it. Talk about hubris; I want to learn from them & see if Notan might have some role in monospaced fonts!

I just have to keep at it. Maybe I can get there in time & with enogh looking & thinking. I doubt I will make it with this font. I haven't quite decided how far to take it in that direction anyway. I don't want to make a consolas-alike!

I haven't seen the pinwheel symetry you are talking about yet. I'll have to look again!

Unfortunately, that rendering engine issue I was on about in the other post is related in the sense that working out the issue had left my shapes are in more of a shables than I'd like. It was with the sense that I might have made some progress since last time & knowing that I had gotten in too far too it it all again that I posted my stuff again.

Hopefully the next post will be tighter, more tested & more pleasing.

Thanks for helping. I really appreciate it!

ebensorkin's picture

Mark, by the way, how would you define your bag? Or would you? Why exclude the rigorous text face?

Mark Simonson's picture

I didn't mean "rigorous text" face--that is a much broader category. I was talking about the Dutch school of type design, which I admire, but don't follow closely. I suppose some people might equate the two. I wouldn't.

I guess what I mean by my "bag" is that I tend to work intuitively and not from a rational/analytical basis. I'm not saying one is better than the other. Different strokes for different folks, that's all. It's not that I don't have rules, but the rules I have are internalized to the point I'm not consciously aware of them most of the time. I just know what looks right. It's part of the reason I don't usually participate in these critiques, I suspect. Too "left brain."

ebensorkin's picture

I guess was way too quick & imprecise when I wrote that. What I meant by rigourous wasn't 'thorough' or 'careful' or 'precise', but had instead to do with a aesthetic of looking stiff, analytic, deliberate, or rational. Proxima Nova is certainly plenty rigourous by the 1st standard but doesn't have the perceptable rational feeling I associate with a lot of dutch type.

If you don't mind, I'd like to ask you about two other things that I think relate in some ways to your work:

- What do you think about 'Legato'? To me it seems to break with the dominant dutch direction in some ways. Do you agree? Do you think it relates to your design better than the other dutch type? Or at all?

- At one point I was trying to convince Hrant that if the caligraphic quality found in dutch type could be married to a carefully made set of conditional substitutes in open type font, then an opportunity for a better interletter spacing or 'notan' could be reached - similar to what a latin caligrapher naturally would do. Or what arabic & chinese/japanese caligraphers do. What made me think about it was the way in which japanese/chinese charcters have to be written out. Each of the complex characters is made of an accumulation of simpler parts. A bit like a word is made with letters but arranged with more freedom of relative placement. When a kanji/character is written it is made *very* contextually. When I look at your logo work I see a lot of that contextual sensitivity. That is one of the things that deeply impressed me about your work. So far I don't think Hrant agrees with my 'more caligraphic' idea - for long texts anyway. I think that he thinks that the best inter letter spacing and legibility will be achived not by a bigger set of conditional shapes but instead by better 'notanic' design of the glyphs and the spaces between them.

What Evert said was,

"FF Legato abandons the idea of a sanserif consisting of simple rigid forms. The mirrored image below shows how (different from typefaces like Gill Sans and Syntax) movement and direction is expressed by outer- and inner-forms behaving relatively independent, and is maintained through all weights. It makes characters refer to previous and next letters to build better words and lines."

And also : "Not only are arabic characters connected, their shape changes according to their context. "

Both these quotes come from:

http://www.evertbloemsma.nl/

Okay, thats probably enough for one post.

ebensorkin's picture

I have an update. Here are the images. They are on the 1st page too.

* at 12 point *

* at 48 point *

I am trying to make it crispy at 12pt. The numbers are not tweaked yet. The Uppercase is slightly tweaked. This is how it renders on OSX. XP/Vista will be different. OSX turns out to be the most difficult to perfect so as soon as I am done with this I will make the TrueType b/w version. Then I will make a Cleartype version that should look alot like this on XP/Vista but at 10pt.

Any comments? Thanks!

antiphrasis's picture

Eben,

It's looking really sharp at 12pt. Very well done, almost looks like an ultra font. I like the big x-height, and how the ascenders are taller than the cap height. Dots inside the zeros are great. The apostrophe is a little bit blurry.

The only letter that might be a little hard to read is the lc M. But there's really nothing you can do about it when you have a monospaced alphabet. I wonder how it'd look if you lowerered the right side a little bit (the second arc)?

I just noticed that lc L, is pretty similar to uc I. The difference is more distinct in earlier versions. Also since the x-height is big, I'd be curious to see uc and lc C, S, U, X, and Z next to each other to see how easily you can tell them apart.

Small comments on the 48pt version, it'd be nice if lc T's leg looped upward a little bit, lc Y's stems connected a little bit higher, and if you could pull in the lc B's bowl at around 1-2 o'clock.

I'm really impressed with what you've done so far. So keep up the good work!

ebensorkin's picture

Lauri, thanks for the crit!

The punctuation hasn't been optomized at all. It will get crisp like the rest soon. I came up with more extreme shapes to try to get two clear bites out of the top of the 'm'. I may need to revisit that again. I am not sure what you mean by the right side. Do you mean drop it down a pixel? I'll have to look at that lc 'L'. Maybe a more traditional monospace shape is needed... I'll post the next sample with some code & with successive placement eg Aa Bb etc.

Re the 48 pt: I wanted to bring the tail of t up too but i thought it made it too blurry at 12 pt. I may put it back after I do some print testing... I'll look at the Y & B. I may need to keep the shapes if they blur out wsith changes. Still, sometimes things work that you would not have expected. The biggest trick in these shapes is the narrow joins on the leg of the r,n,h etc. Doing the join that way took out alot of blur! I am also considering having the shapes loose a little weight to cut blur as well.

Syndicate content Syndicate content