First impression: the brackets are too heavy in both directions. The right-leaning stroke of the /x and /X also appear heavy.

An accomplished DIN. Have you thought of a double-storey |a|?

I am not a font designer. On the typophile.com screen, the 8 looks thicker than the 7.

Tried fixing the x'es
Played around with a more open aperture for the c and e

Double-storey a is in the works - don't know which will be the alternate.

Great feedback guys :)

Made changes to letters ({[2Xxkce]})78
({[2Xxkce]}).pdf

The question mark looks unbalanced, and I think a more sophisticated form (no corner) may be more appropriate.

More glyphs and a two storey |a| v0.3.pdf

@cerulean - will try with a more sophisticated |?| and to balance the existing one - and then one of them will be the alternate version

Top stroke of |a| looks a bit dark, especially on the right. The bowl could be a bit smoother overall, the counter in particular.

I know - I just had to sleep :)
I'm halfway through fixing it

Updated |a|

Screenshot

|carlsberg|
Screenshot

Hi Andreas... a couple of observations.

The bottom right of the /s/ is very dark.

The strokes of the /S/, /g/ and, most noticeably, /s/ all taper. Possibly the top of he /c/ as well.

/f/ and /t/ look narrow to me, and I would consider spreading apart the bottom vertices of /W/ from each other.

@Steve - The tapering was by design but reduced it and like it better, thx :) not sure about the s - please see screenshot in this post

@eliason - W improved, I have already increased the width of f and t - will maybe add more.

greats

Maybe it's something that's crept in during the creation of your PNG, but it looks like there are a couple of kinks in the /s/ and one in the /a/. The one in the bottom half of the /s/ looks particularly pronounced to me.

@ steve, there was a kink - creating bézier curves without kinks is hard (first font:)

Tried to fix it: ase

Hi Andreas - the /s/ looks much better, but it still looks a touch heavy in the bottom right to me.

I tried to fix the bottom heaviness of the |s| without making i top heavy and also tilted it a wee bit forward.

Been playing around with some special characters.

Hi Andreas - I think the /s/ looks great.

You have a little kink almost exactly in the middle of the /S/. It flattens the first part of the lower curve, which you can see most clearly by comparing the two semicircles within the /$/ sign.

I like the idea of a slightly unconventional '@' symbol, though I have a sense having the stroke end the way it does in the top left will make it look a bit unbalanced. It would be good to see how it works when used within an email address. It might also be interesting to play around something alone these lines (with apologies for the crudeness!).

I like your @ too - appears a bit more conventional. See the current in context below:

Not sure what you meant about the |S| as the S and dollar are identical (except for the | and slight smaller scale) but I've made quite a few adjustments to it. The new |S| is to the left, the |s| are identical.

Thanks again for your feedback. It's really appreciated!

Hopefully, the image below will show what I mean about the /S/. There is a kink in the curve where the red spot is and this 'flattens' the top of the next curve.

I created a short mockup just to test it: Short article mockup (without the updated S though...)

About the [S]: You explain it very well - I think it is gone now?

Earlier someone talked about a more complex [?] - what do you think about this:

I put the whole thing up on github so anyone can contribute there if they like :)

github.com/gidole/sans

Hi Andreas,

I think the /S/ looks much better now.

I like this smoother /?/ as well. I still think it is a little unbalanced though. I think it would work better if you experimented with different lines for the top left part of the stroke, something like the mock-up below, on the right:

You're right, looks better. My attempt:

It looks much better.

:)

Played around with some alternates

I've made quite a few changes so any improvement suggestions on the glyphs and on the spacing is very welcome (note that I haven't done the kerning yet).

See the pdf for higher resolution and more samples.

Looking good. I'm pretty certain the "dotaccent" characters are meant to have a dot above, not below.

Thanks, and you're right - some of those that should be above are currently below (fontforge is not playing nice here...)

EDIT: Just found the combining characters - so now using e.g. U0315 instead of comma when combining - fixes a lot of issues. Will update the pdf soon :)

Just snagged it off the git and immediately tripped over a trap: The Windows descender value (at least on my WinXP machine) has the axis reversed – negative values lie above the baseline.

@ Te: could you try the latest version https://github.com/gidole/sans/blob/master/Resources/Gidole-Regular.ttf?...

Same problem, Andreas. You have your WinDescent set at -200 when it should be 200.

Also, your WinAscent is set at 800, when I believe it should be 667.

Ok, I'm very new to font designing so I misread this guide a bit: http://designwithfontforge.com/en-US/Line_Spacing.html

If I change the ascent, should I change the value in the general value too? Which would give me a new EM size, no?

Sorry, Andreas, forget what I said earlier - I was in a hurry and only looking at the basic characters, not the whole font.

You'll need the WinAscent value a bit higher than the uppermost part of your tallest glyph, including diacritics. The tallest glyph is currently /A ring acute/ (Unicode 01FA) at 1085. So, the WinAscent value for Gidole needs to be a little above that figure.

The WinDescent should be set lower than the glyph with the lowest node, which is currently /g cedilla/ (Unicode 0123) at -385. However, this glyph is usually represented as a /g/ with something like a /comma/ on top. See this page and click on the 'view' link alongside each font to see how they treat this character.

If you change your /g cedilla/ to match those in that list, your /A ogonek/ will become your lowest glyph, with a node at -202. You need to allow a little extra room, but enter the value as a positive one, not a negative one.

As for setting up the metrics in FontForge, I'm afraid I don't use it so can't be absolutely sure about the 'General' value you're referring to.

Ok, was also what I figured out from my quick google search - thanks a lot for the help and link:)

Trying to do a bit kerning at the moment, thought it would be dull but actually fun because it improves the font so much...:)

Just a screenshot of the work in progress, far from done.

Looks better off the rack but for the tightness of lines in Windows's font viewer. I'm going to make the wild-ass guess that most of the time (assuming Aringacute is a rare character anywhere) you could get away with WinAscent/Descent of 960/240. That these figures are the same as em size plus the line gap is just a coincidence, honest! ten fingers up to God.

Funny thing is you could tell me anything and I'd believe since these numbers are french to me...

I've actually changed the diacritics a bit (removed æøå with diacritics as they are never used anyway...). My lowest is now -230 and tallest 883
250/900 and zero lines space looks ok on a mac...

pc? Link: https://github.com/gidole/sans/blob/master/Resources/Gidole-Regular.ttf?...

Hi Andreas - here's a PDF of Gidole in MS Word 2010 on a PC. I don't see any obvious problems with metrics.

http://freepdfhosting.com/4e98370675.pdf

Thanks, good to know:)

I created a github page just for fun.

http://gidole.github.io

Looking good overall! I can certainly see myself using it.

Do make sure you iron out the curve kinks, a clean and monolinear sans is particularly sensitive to those things. There are tools available that help with that.

Is /M somewhat asymmetric?

Have you considered a wider /t, with a more pronounced bottom curve?

One thing I don't like about DIN is the parentheses, which rather look like brackets. Have you tried a simpler, rounder arc shape for those?

@ catharsis - thanks, care to elaborate on the tools available?

The |M| was off by 1/1000 point - but maybe some optical correction is needed?

Only made the parentheses really quick - will play around with it.

The |t| - I kinda like the simple somewhat narrow but could make an alternate.

updated quite a bit

a bit less spacing
tried to improve almost every basic letter with a curve for better rendering on screen
new t,f,() based on catharsis' suggestions
more...

Web preview: http://gidole.github.io
TTF: https://github.com/gidole/sans/blob/master/Resources/Gidole-Regular.ttf?...

I've started working on a hairline version - please see pdf for high-res: PDF

@ catharsis - thanks, care to elaborate on the tools available?

There's Speed Punk and RMX Tools, for example.

Thanks, It's only a hobby project and the font is free so can't really afford to buy tools I'm afraid :) Should anyone with any of those tools at hand want to have a look then please do ;)

What I've done so far is trying to use as simple geometric shapes whenever possible and to model more complex shapes "around" geometric shapes.

I made many small adjustments and getting more satisfied with the overall look. Kinks should be gone in all letters (haven't gone over all symbols yet).

Any feedback is still much appreciated.

Project page with download.

Coming along nicely!
Have you considered to add Amharic/Ge'ez script support?
http://en.wikipedia.org/wiki/Ge%27ez_script
http://en.wikipedia.org/wiki/Amharic#Alphabet

Thanks:) Learned how to use cornu spirals which helped a lot.

And I have but I've unfortunately only learned to read/write it whilst in kindergarten so it's quite a big undertaking but could be inspired by noto and other version as a start.

Cyrillic and greek will probably come first though.

PDF for resolution independent preview: ReadMe+TestPrint.pdf

Thanks, It's only a hobby project and the font is free so can't really afford to buy tools I'm afraid :)

You know, people sometimes do spend money to support their hobbies. ;o)

Your newest letters are indeed quite smooth; good job. The only thing that still catches my eye is the curve in /n/m/u/h. I believe it would look better if you moved its apex a bit away from the main stem. Don't be afraid to break the symmetry of the arch; it's not symmetric anyway, with a straight stem on only one side. The arch might actually look more symmetric if you bend it away from the stem a bit.

The /8 looks top-heavy. The stress on the ogonek looks reversed. Maybe lighten the bowl's connection to the foot in /a a bit? Is /4 a bit too wide?

How complete is your character set? Do you have /Þ, /ð, /ß, etc...?

Oh, there’s all of Latin-1; no worries. It’s complete up to (but not including) the long s.

@Catharsis:
the apex on /h /n /m /b /d /p /q is 5 pt away from geometric centre already but I've tried moving it 5 pts further away (not /u as I currently use a stemless one) - see /h vs /n below.

/8...maybe a little bit but I want it to look as symmetrical as possible so it's a fine balance...

/4...maybe, I've narrowed a little bit.

@Té: Is the long s used by anybody these days?

(right click and open image in a new window to see side by side)

and about the hobbies - I know but I have too many hobbies and they all cost...and I'm still studying so can't really allow myself to buy expensive software at the moment when Fontforge works ok-ish... :)

Updated ogonek:

I'm still not 100% happy with the /n arch. I realize you want it to appear circular, so moving the apex even more to the right is probably not in your interest. But the stroke still feels a bit wobbly to me, like perhaps the curvature of the upper right outer outline is too low, making the stroke locally too thin:

/8: I understand your goal, but the current /8 does not meet it in my opinion. The upper half looks larger to me.

As for the long s, I'm quite fond of it. In fact, I use it in the official orthography of one of my conlangs, Jovian. :o)

Ogonek: The stroke thins down on the bottom left diagonal, which breaks the monolinear feel of the font. If anything, the stroke should be slightly thicker there since it's a "strong diagonal" in pen rules.

The ogonek was updated too quickly so I accidentally flipped the thickness of the stroke.

About the /h the stroke was already slightly thicker to make the curve more round but I have increased the width of the stroke a bit.

I think the long s is very rare nowadays, but I still stubbed an 'f' to take its place in my copy of Gidole, just in case I wanted to use it somewhere.

Is the long s used by anybody these days?

In faithful transcriptions of old texts or in discussions on the origins of |ß|, to give just two examples ;-)

Andreas: Looking at your /h up close, I still get the impression that the right side of the arch is wider than the left side (it isn't, but to my eye it looks that way) and that it would profit from moving the apex a bit more to the right. The stroke also still appears to me to be thinner at its 2 o'clock position than at its apex, probably because I intuitively expect a "heavy" diagonal there. The outer upper right curve segment could probably profit from a bit more outward curvature.

At small sizes, the shape looks quite good, though.

@Catharsis: I haven't uploaded the changes to /h, /n and /m yet so it's only the screenshots. Will update in a minute. :)

@Té you can always send you additions back to me on Github (although I just added the stubbed /f) so everyone will benefit from any modifications - that's why I've put the source up there in the first place:)

I was talking about the screenshots.

Ah - apologies, I just assumed you used the font file since you mentioned small sizes. Here's a pdf for better resolution: https://dl.dropboxusercontent.com/u/3019777/gidole/0.9.5.pdf

And the diagonal is as thick as I'll let it be since making it thicker makes it look too heavy.
But making it slightly thicker has definitely improved the look so thanks for the suggestion:)

I'm still unsure about the /4 - I think it looks a bit narrow now and a lot of fonts have a /4 that is relatively wider - I'll probably leave it for now though.

I'm not really happy with the overall shape of the /8 so I'll try to fix that now.

I would likely have to set up on Github first (haven’t looked into it, really). Most of the mods were just basic cheap adjustments towards my preferences, anyway: adding some morse (Unicode spaces, dots and dashes) and setting the en and em dashes flush to their canonical widths.

@Té if you could just send me your modifications so I can inspect and maybe use some of them?

Updated the parenthesis, curly brackets and more:

Right now it’s just spaces and dashes; haven’t muddled with the dots yet. The spaces are a semi-fab kit that’s renumbered, given values for punct/number/thin/hair and then plunked straight into the SFD file at the end. FontForge is smart enough that I don’t really need to adjust the totals in BeginChars and the maxp table, but I do so anyway.

Also, I set the char. numbers to start at 600 so I wouldn’t have to renumber them for a while yet as I transfer them from a previous version to the next. This set is from Gidole 0.9.

Edit: There are some dots now, in the en leaders.

StartChar: endash
Encoding: 8211 8211 472
Width: 500
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
0 284 m 1,0,-1
 500 284 l 1,1,-1
 500 218 l 1,2,-1
 0 218 l 1,3,-1
 0 284 l 1,0,-1
EndSplineSet
EndChar

StartChar: emdash
Encoding: 8212 8212 473
Width: 1000
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
0 284 m 1,0,-1
 1000 284 l 1,1,-1
 1000 218 l 1,2,-1
 0 218 l 1,3,-1
 0 284 l 1,0,-1
EndSplineSet
EndChar

StartChar: uni2000
Comment: En quad - half an em quad
Encoding: 8192 8192 600
Width: 500
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2001
Comment: Em quad - as wide as tall
Encoding: 8193 8193 601
Width: 1000
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2002
Comment: En space - half an em space
Encoding: 8194 8194 602
Width: 500
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2003
Comment: Em space - Usually, an em quad
Encoding: 8195 8195 603
Width: 1000
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2004
Comment: M/3 space - one-third of an em space
Encoding: 8196 8196 604
Width: 333
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2005
Comment: M/4 space - one fourth of an em space
Encoding: 8197 8197 605
Width: 250
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2006
Comment: M/6 space - one-sixth of an em space
Encoding: 8198 8198 606
Width: 166
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2007
Comment: figure space - as wide as a tabular figure or a zero
Encoding: 8199 8199 607
Width: 534
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2008
Comment: punctuation space - as wide as a period/comma
Encoding: 8200 8200 608
Width: 202
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2009
Comment: thin space - here, 3/4 of a word space
Encoding: 8201 8201 609
Width: 187
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni200A
Comment: hair space - here, 1/4 of a word space
Encoding: 8202 8202 610
Width: 63
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni200B
Comment: zwsp - zero-width space
Encoding: 8203 8203 611
Width: 0
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni200C
Comment: zwnj - zero-width non-joiner
Encoding: 8204 8204 612
Width: 0
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni200D
Comment: zwj - zero-width joiner
Encoding: 8205 8205 613
Width: 0
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni202F
Comment: non-breaking thin space, equal to uni2009
Encoding: 8239 8239 614
Width: 187
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni205F
Comment: mmsp - medium math space, 2/9 of an em space
Encoding: 8287 8287 615
Width: 223
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uni2060
Comment: word-joiner - non-breaking zwsp or zwj
Encoding: 8288 8288 616
Width: 0
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: uniFE00
Encoding: 65024 65024 617
AltUni2: 00fe0f.ffffffff.0 00fe0e.ffffffff.0 00fe0d.ffffffff.0 00fe0c.ffffffff.0 00fe0b.ffffffff.0 00fe0a.ffffffff.0 00fe09.ffffffff.0 00fe08.ffffffff.0 00fe07.ffffffff.0 00fe06.ffffffff.0 00fe05.ffffffff.0 00fe04.ffffffff.0 00fe03.ffffffff.0 00fe02.ffffffff.0 00fe01.ffffffff.0
Width: 0
GlyphClass: 2
Flags: W
LayerCount: 2
EndChar

StartChar: figuredash
Comment: Set as wide as a zero
Encoding: 8210 8210 618
Width: 534
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
65 284 m 1,0,-1
 469 284 l 1,1,-1
 469 218 l 1,2,-1
 65 218 l 1,3,-1
 65 284 l 1,0,-1
EndSplineSet
EndChar

StartChar: uni2015
Comment: Quote dash, here a 1.5-em flush dash
Encoding: 8213 8213 619
Width: 1500
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
0 284 m 1,0,-1
 1500 284 l 1,1,-1
 1500 218 l 1,2,-1
 0 218 l 1,3,-1
 0 284 l 1,0,-1
EndSplineSet
EndChar

StartChar: uni2E3A
Comment: Two-em dash
Encoding: 11834 11834 620
Width: 2000
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
80 284 m 1,0,-1
 1920 284 l 1,1,-1
 1920 218 l 1,2,-1
 80 218 l 1,3,-1
 80 284 l 1,0,-1
EndSplineSet
EndChar

StartChar: uni2E3B
Comment: Three-em dash
Encoding: 11835 11835 621
Width: 3000
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
80 284 m 1,0,-1
 2920 284 l 1,1,-1
 2920 218 l 1,2,-1
 80 218 l 1,3,-1
 80 284 l 1,0,-1
EndSplineSet
EndChar

StartChar: longs
Encoding: 383 383 622
Width: 359
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
115 513 m 2,0,1
 115 566 115 566 130 597 c 1,2,3
 145 632 145 632 178.5 652.5 c 128,-1,4
 212 673 212 673 256 673 c 0,5,6
 283 673 283 673 321 663 c 1,7,-1
 303 599 l 1,8,9
 274 607 274 607 256 607 c 0,10,11
 211 607 211 607 194 571 c 0,12,13
 185 553 185 553 185 513 c 2,14,-1
 185 0 l 1,19,-1
 115 0 l 1,20,-1
 115 415 l 1,21,-1
 35 415 l 1,22,-1
 35 477 l 1,23,-1
 115 477 l 1,24,-1
 115 513 l 2,0,1
EndSplineSet
EndChar

StartChar: uni2023
Encoding: 8227 8227 623
Width: 325
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
SplineSet
80 167 m 0,0,1
 80 333 l 1,-1,1
 245 250 l 2,-1,1
 80 167 l 3,-1,1
EndSplineSet
EndChar

StartChar: uni2024
Encoding: 8228 8228 624
Width: 500
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
Refer: 17 46 N 1 0 0 1 149 0
EndChar

StartChar: uni2025
Encoding: 8229 8229 625
Width: 500
GlyphClass: 2
Flags: W
LayerCount: 2
Fore
Refer: 17 46 N 1 0 0 1 24 0
Refer: 17 46 N 1 0 0 1 274 0
EndChar

Thank you Té - I've added most of them in the next update :)

It surprises me how rarely I see this in fonts, given how little it costs. I’m sure the old codgers here can give you much better advice on how to tune it for your font than I ever could.

I guess they're rarely used but you're right - takes up no extra space so one might as well add it.

I've updated to 0.9.8 now with many small fixes. Should be even cleaner curves and changed + added some glyphs. Managed to reduce the file size relative to the number of glyphs.

@Té: a small tip - use github gist for sharing raw data in the future so it doesn't take up so much space in a thread.

Example

gist.github.com

I've updated many of the diacritics and other special characters.
Page 5+6 in this pdf.

I'm very unsure about the /ß (haven't made capital ß yet). Have tried different things but not quite happy yet - suggestions welcome. I have changed the /ð so it has correct x-height but I can see many fonts use a smaller circle. I'm also unsure about the cedilla (¸) - any suggestions?

I'll probably make the oe and æ a bit wider than the current is a bit too narrow.

Alternative ß:
I like this glyph better but not sure if it "fits" with the rest of the font - i.e. too curvy and not geometric enough.

Yet another ß

The upper counter of /ß is much too small.

Agreed, changed it for all three versions...still undecided about which style suits the rest of the font best. I'm leaning towards a combination of 1 and 3.

First or third. Third one is perhaps a bit too wide.

I think the combination between 1+3 works quite well - or?

My first go at /ẞ (I've never really used this character so it's a bit "foreign" to me so unsure how to do it

Thumbs up for |ß|. |ẞ| is tricky. I think yours is in the right direction, but isn't yet completely convinced it belongs to the uppercase ;-)

Possibly useful resources:
http://opentype.info/blog/2011/08/03/how-to-draw-a-capital-sharp-s/
http://opentype.info/blog/2013/11/18/capital-sharp-s-design-review/

Very useful - thanks again both of you :)

Looks more "capitalesque" with less curve. Can't decide if it's too wide though.

Yes, /ẞ should be wider than /B in many fonts, otherwise it can look cramped. Your design is a good start, but a bit unbalanced. I would make the top wider and move the tail end further to the right so as to make more space between it and the stem.

Your /ß works for me now.

Revisiting the alternates - haven't done anything to zxkZXK yet - suggestions?
Unsure about the /N

Experimenting with the /ẞ as per your suggestions

I like the alts, they turn this into a whole new font. I feel like /G and /y could use some improvement, though. How about a /u-shaped bowl for the /y?

The new /ẞ looks good, but you'd have to put it in context to see how well it works. The curves on the bowl look bumpy, and there seems to be a pinch around its 2 o'clock position, though.

... And a curved leg for the /K/k?

Yeah, wasn't happy with the /G or /y either - I have reverted the G and altered the /y.

I tried many different kinds of curved legs for the K's yesterday - didn't really work but flipped the curve and think it looks better. Altered the /e and made some other small changes.

I think the /ẞ works well now - there were no kinks but the antialiasing in fontforge is bad:)

Latest draft of the alternate version. Will probably name it Gidole Play if I don't come up with something better.

The /ẞ still feels a bit unbalanced to me in that /FUẞBALL sample. The hard corner is at odds with the rest of the letters, the top counter is small compared to the wide belly below, and the bowl's stroke ends uncomfortably close to the stem. I would make the top slightly wider and the downward angle slightly steeper (bringing the central junction further away from the stem), and end the bowl's stroke farther away from the stem.

I like Gidole Play a lot. I was going to suggest narrowing the /V/v/y, but I think you did that on your own.

I'm wondering whether the diagonal curve at the bottom right of /a/d couldn't use a bit more curvature to fill out that gap where traditional typefaces would have a foot and/or an outstroke. Currently they look like they might tip over to the right.

The spacing still looks a bit uneven to me; look at the wide spaces around /c or the irregular rhythm of /pqrst...

You're spot on as always Christian - and I can tell I'm improving because I have already narrowed them and I have also tried to tip a,d,b so they stand more balanced but haven't been able to create something balancing that still looks visually pleasing...yet:)

Spacing and kerning is definitely my weaker side and Gidole Play's spacing is more or less just copied from Gidole so I'll need to look more into it.

The ẞ also needs more work - and I'll probably follow your recommendations and end up with something between the current version and museo sans version:

What do you guys think about this /a and d ? (Right-click and open in new tab to see it better)
I'm torn but inclined to go with the new versions.

Have you tried a curve like the one on top left of |P|?

I'm getting too much of an "o with a stem"/"ð with no crossbar" feel off the new d. Perhaps a slightly flattened bottom would help?

Played around with different solutions - making the font somewhat different but probably also better - what do you think?

I think the work on /a /d might have gone a bit too far — the bottom right now looks more like a corner than a curve. Something also bothers me about the bottom curve of /p, not quite sure what it is... maybe you'll be able to use your work on /d on /p as well.

Are /f /t too narrow? And /y still too wide?

Hehe...I have already reversed the a and d so they are more curved :)
I'm playing around with p,q,g,r,n and m using the same corner as on b,d...not sure how it will turn out

Playing around trying different forms. Some letters such as /g are only half done as I want to get the bowl shape right first.

Latest test:
http://imgur.com/a/1YKVz

Current version on behance:
https://www.behance.net/gallery/23391815/Gidole-Play-free-typeface

I’m not sold on the ∩-shaped |N|. Have you considered an asymmetric design (raised left shoulder)?

Actually, I found that /N rather charming...

I kinda like the ∩-shaped N too...made it a bit more squarish in latest version.

http://imgur.com/a/iP9ip

I don't like the squarish version. It breaks with the pill-shaped element that all other characters are based on. The inconsistency with /U is especially notable. If you're going to commit to an /n-shaped /N, go all-out and make it analogous to your /n.

Point taken on the squarish N.

I've tried curving the straight parts:

Major update underway - design is almost final now. Although always open to suggestions:)

I've narrowed the /e a bit - not sure if it's too much or just right...?

High res: http://i.imgur.com/hktKDCP.png

New small changes to a, b, c, d, e, f, g, h, j, m, n, o, p, q, r, t, u

http://typophile.com/files/gidole_5723.png

I've adjusted the width of many of the letters and also changed a bit spacing
http://typophile.com/files/aQuickBrown_3944.png

Updated many of the curved letters - some minor changes and some bigger. Any suggestions?

Unsure about:
2 - should I make the shift from curve -> straight line softer?
6 - tilting to the left? Unsure how to fix without 96 looking weird.
8 - Too wide?
9 - Currently an inverted 6 - but should it be less wide?
e - I made the foot a bit longer - too much?

I decided to do a complete redraw with 2048 UPM for better detail and easier interpolating later on.

In the process I've made some minor changes: increased the overshoot and contrast, vertical metrics ≈ Roboto, and the curves follow the guideline circles a bit more.

Thoughts?

https://dl.dropboxusercontent.com/u/3019777/gidole/gidole2.png

It starts to look like a different typeface.

I know...but was actually something like that I'd wanted to create to begin with :)

I'll post "the quick brown" soon for a better view.

The quick brown with quite a few adjustments:

https://dl.dropboxusercontent.com/u/3019777/gidole/lazy3.pdf

I've now revisited all basic characters, any suggestions?
https://dl.dropboxusercontent.com/u/3019777/gidole/gidole%20rev%202.pdf

Looks really good, Andreas, but I still think you have an issue with the shoulder of the /h/, /n/ and /m/. The right side is very 'flat' and I think the problem originates from using the /o/ in their construction.

The /o/ has a slight diagonal stress, with the top-left/bottom right parts of the curve being thicker than the top-right/bottom-left. If you flip the /o/ before you start construction of the other letters, I think you'll find it easier to eliminate the 'flat' spots.

I've made a rough mock-up to show what I mean. The middle /n/ is my attempt and was made from the /n/ and /o/ from your most recent upload. I needed to tweak the junction a little after merging the two shapes, but you'll notice the shoulder is 'fuller'.

The /o has no diagonal stress - but should it?
I haven't used the /o when creating the /n but I've used circles for guidelines for both.
I have tried making the shoulder as "full" as possible without looking too distorted or thick - it definitely looks better but is it enough?

Thanks for the feedback:)

EDIT: Image deleted, see below

I've updated to try and make them look even fuller (b untouched but included for comparison)

Just redid the numerals - slightly narrower now to fit the other characters better.

added some more conventional alternates

I agree with Steve, the arches of /n/m/h/u could still use some work. The apex of the top curve is too far to the left, and the inner curve is too horizontal. The stroke looks heavy at the apex, then lighter in the diagonals, then heavy again in the stem.

Forget circles. You'll never get these curves to look circular if you actually make them circular.

@ catharsis - I forgot everything about circles a long time ago, I've used a "squircle" as guidelines but spirals to draw the curves:)
I get your point though and had already done some of what you've guys have been suggesting but now that I've looked at it again it wasn't enough.

I've tried making the h,n and m better as per your suggestions - thoughts?

http://typophile.com/files/hmn_3990.png

I think the joinings are perhaps a bit too thin, but it's something one shold look at text sizes to see how it really works.

@riccard0 I didn't make the thinnest point thicker but made the top curve of the joining a bit more rounded:

https://cloud.githubusercontent.com/assets/3859425/6920225/899fe836-d7bd...