New Microsoft size-specific design selection mechanism

Sandy Cerovich

Thanks Georg. I'm getting old...memory has to be upgraded. I know I had a discussion with someone there back in 2008 and was told not to use weights less than 250. The reason escapes me, but you could be right. Struck me as an odd thing to do, since you could use the same weight, but call it 250 and it would be fine...but if you called it 200 it caused problems.

Ofir Shavit

It's the same thing to this date
I've encountered horizontal font smearing of less than 250 ("reported") weights in several MS software. You better fake the weights values below 250.

James Puckett

You better fake the weights values below 250.

Set weight values below 250. It’s not your font Microsoft lets incompetents design and code software. If users complain tell them whose fault the bug is.

Deleted Account

I think you mean it’s not your fault, if users complain tell them whose bug the font is?

Thomas Phinney

> This should probably be a different topic, but it does tie into the Microsoft specs discussion. Weight - at one time the MS specs only allowed weights above 250. Has that changed? Do they allow 100 or 200 now, for instance?

This was not a spec failure, but an implementation failure by many developers, including MS Office, who often simply assumed that there were only normal and bold fonts and wrote code that made that assumption.

Years ago, when Ernie tested this issue at Adobe (at my urging), we found that about half of all Windows apps we checked made the relevant poor assumptions. Adobe considered that releasing fonts that smeared bold in half of all Windows apps was unacceptable.

> It’s not your font Microsoft lets incompetents design and code software. If users complain tell them whose fault the bug is.

Yes, how dare they let arbitrary third parties develop software!

Remember, it is not only Microsoft apps that did/do this, and it was not the OS's fault as far as I could ever tell. Rather it was the fault of assuming the base member in a style-linked family would have a weight of 400, and asking the OS for a "regular" weight font for the base member. The OS would then synthesize such a font if it could not find one. (weightclass < 250, then faux bold by adding 300 would be closer to the "desired" weight, hence it should do a fake bold. Ugh.)

SiDaniels

I don't think there are any restrictions on weight class in the OT Spec. The limitations are with apps and OS's. Related: CSS weight classes have to be multiples of 100.

Nick Shinn

The zeroes have always struck me as being rather unnecessary. Although the extra room must have seemed like a good idea to the software engineers at the time, considering how they had goofed with the “a font family has only four members” thing.

Deleted Account

Oh, know they strike you as unnecessary, Nick. But may I introduce you to my buddy biáng, and then you'll perhaps think more of OS/wt's zero(s) than as a stumped appendage to OT?*

http://blogs.adobe.com/CCJKType/2014/03/ids-opentype.html

Nick Shinn

That’s a bit beyond me David, and doesn’t appear to have anything to do with weight.

This Typophile thread came up with a maximum of 15 weights in a typeface family:
http://typophile.com/node/59141

Thomas Phinney

Nick, if there can be up to 15 weights in a family, how are the extra potential digits unnecessary? After all, without them we have only nine possibilities.

Equally problematic is the particular distribution of those possibilities. For example, typefaces today often have a number of lighter-than-regular weights. If one keeps with the specs and restricts to single digits, there can be at most three such. Similarly, at most two weights heavier than bold.

Deleted Account

Jeeze, Sorry Nick, Iost some of my paste and left a hanging reference mark.

I know this link isn't about size, or even weight. Yet, if one looks at the heaviest stroke of the heaviest biáng in the six-biáng range Adobe is so kindly illustrating, that's 500-600 by my noodle weight reckoning. What that means is although biáng might stop gaining weight 1/2 way through the scale†, the rest of Chinese needs a lot more than 5 weights to reach Black in a 1-stroke Chinese glyph, much less including what I discuss in †. So 1,000 discrete values are only just enough for all weights of all glyphs, imho, and I'd like more.

The impact of this OS/2-not-enough-wt-thing, and it's relationship(s) to the OS/2 not-enough about size-thing, globally, imho is barely inadequate for a Latin standard, and as one gets into the depths of kanji canyon, well, knowing there is still a river running at the bottom, cutting ever deeper into the glyph repertoire served by this standard, doesn't help lend a positive view to anything less than a global solution asap, as that inadequacy globalizes.*

*Or not. maybe there should be two or five 'standards' for glyph, style, family, size, and etc.

† which of course it would not have to, if it was 80 feet tall, for your 30 foot away amusement, and those little white spaces would like to range from the size of a cherry pit to the pinky of a spider, taking OS/2 well past 'Eleven'.

Nick Shinn

You are both no doubt correct, but I still find the idea of 100 possible weights a bit strange, as I have not been able to make above eight. I guess it’s a consequence of the decimal system, that in order to have eleven, you also get 99.

Jan Schmoeger

you also get 99

Good thinking!

Thomas Phinney

Actually, it's 1000 possible weights. I will freely grant that 100 would likely be enough, but I don't feel hard-done-by that it is 1000. (I do feel hard-done-by that CSS decided to only recognize the 100-increment values and ignore anything else about the system, but that's another story.)

Deleted Account

Is zero possible?

James Puckett

A conceptual font with zero stroke width is possible, but only useful if you’re a Dutch student looking for a lot of free publicity.

kupfers

http://sourceforge.net/adobe/adobe-blank/wiki/Home/

Nick Shinn

https://www.p22.com/family-Cage
This begs the question: what is the difference between a font with zero stroke width and one with zero strokes.

John Hudson

...what is the difference between a font with zero stroke width and one with zero strokes.

A zero-width stroke could be be made visible in a number of ways (parallel paths, dynamic stroking, plotter pen).

Nick Shinn

Yes, there are no images in digital fonts, only instructions which are dependent on, or interpreted by, rendering devices.