I just publish in TyMS (Typefaces Measure System) website my first thesis to share with the community, this thesis proposes a method so that the different weights (statics, masters and instances) of the fonts will be named according to their optical thickness and not according to the free will of each Tyner (Type Designer).
The names of the weights and the order of the list that you will see in this thesis are still work in progress. As Stefan from Cape Arcona advised me, I opened the discussion to the community to help me fine tune the names, hoping later on to make the final list as consensual as possible. Please give me your contribute to order the final list. See the all thesis here: https://pedromascarenhas.wixsite.com/tyms
Comments
https://docs.microsoft.com/en-us/typography/opentype/spec/os2#usweightclass
First, I agree that the number could be more useful than words in terms of distinguishing members of a large family.
That said...
1) The weightclass numbers are familiar to web designers and type designers, but not to other users. For example, the average user of Word or Google Docs will not know these. (Yet.)
2) Neither 500 nor Medium has any intrinsic meaning, so there is a still an arbitrary aspect.
3) The average user might similarly not know whether Bold or Heavy is heavier, but that problem is more easily and universally solved by apps: they should sorting styles (in part) by WeightClass. Adobe started doing this about 20 years ago.
The Regular weight of my current project, a type with extremely short x-height, which I have adapted from a normally-proportioned typeface by combining its Light weight, for majuscules, and Bold (with extended ascenders), for minuscules. With some other adjustments, of course; nonetheless, this disparity between cases would appear to falsify your theory, Pedro.
I think you also need to avoid reusing distinctively those names that are treated as synonyms in the MS OS/2 weight class specification, much as I wish they were not.
Partial weight steps are almost always understood as relative to Regular. So Semi Bold is less bold than Bold, but Semi Light is more bold than Light. Weight naming moves outward from Regular, not linearly from the lightest.
Nick Shinn,
your post is very relevant.
In my study I understood that are so many "not normal" fonts in the market that a universal thesis will always found some grains of sand in gear. That's why i say in the chapter "What about disruted fonts?" that Some fonts do not follow the rules established by the status quo of font design (and fortunately because it is in creativity and breaking the rules that evolution finds its way into the future), so for these disrupted fonts, Tyner's common sense it's very important to interpret what is the #PST of the font, and then give it the most correct weight name, the name that has the same optical weight as the fonts considered "normal".
With TyMS Weight it will be more easy for "not normal" fonts to find whats the best weight name to give, don't you think?
Thanks for you wised share.
I will take that in mind when i will review the weights names.
Good posts. The main thing for me and other graphic and editorial designer (fonts consumers), is not if the bold is 500 or regular is 400.
What we need is that all Regular looks the same optical thickness and all Bold also, and so on. And for that the most important is that tyners name there fonts with the same thickness logic.
For fonts without optical size variants, even something as simple as the ideal weight of “regular” depends on the intended usage size, and other variables, including ... designer preferences. Some folks might prefer a slightly lighter or heavier “regular,” or prefer it in different circumstances.
One solution for this is making variable fonts. But with existing fonts, the usual solution is to simply choose a different font with the characteristics one desires. Stomping out that variability between font families would be bad, not good, in my opinion.
And that isn’t the only problem. Adopting a NEW descriptive field such as usWeightClass is easier in that there is not as much backwards compatibility issue. (And even that had more than one would have liked!)
While you can publish a manifesto or guidelines, you face the critical problem that your solution is not like weightclass. To make it work, most existing typefaces would need to be substantially redesigned.
Even if one thought your end state of all fonts having the same weight designations was viable and meaningful (and I don’t, for the reasons above, AND the reasons @Craig Eliason stated), retrofitting existing families would be a ton of work and would require convincing all their owners that it was worth doing. That seems highly unlikely.
Thanks, your solid arguments, definitively they will hep me in my journey to pass to tyners the anxieties of the graphic designers.
Just leave here 3 topics
About "Some folks might prefer a slightly lighter or heavier “regular,” — thats why my thesis leave, not an exact number, but a range for each weight.
About "usWeightClass" — this things as you know are not closed, they update with the progress.
About "To make it work, most existing typefaces would need to be substantially redesigned" — No, because i had that in consideration, so almost all of them already fit the metrics of my thesis. Those who don't fit are the most inconsistent ones.
Gill Sans, for instance, has never had a “Regular”, its normal weight was originally what we would now term Medium.
When I updated some of my earlier digital designs, I revised the weight names, making the old Light the new Regular.
There were four factors, I think, involved in the historical lightening of the default:
Seeing the Light—an article I wrote on the topic in 2001.
If that is the goal, then I really think you have to come at it other than through the traditional weight naming or even the numeric weight classes as defined by CSS and recent iterations of the OS/2 table spec. Those systems by virtue of both how they are defined and how they are used are not going to produce optical weight equivalence between different typefaces. Yes, a type designer or group of type designers could decide to use those systems in a specific way relative to the internal design space weight range, in such a way that fonts made in that way would have optical weight equivalence, but this would still be a private convention without any standardisation that would enable interoperability with other fonts or in software controls.
This implies a new system for dialing weight variants that is explicitly intended to produce optical weight equivalence, with strict implementation rules to ensure interoperability. David Berlow has gone some way towards this in his parametric variable fonts, and a project like fauxfoundry has demonstrated the interoperability potential of such an approach, but it still seems to be under-specified in a number of respects, such that it isn’t clear how to maintain the relevance of a measurement system across multiple writing systems that do not share common structures and proportions. As Craig and Nick pointed out, there is a lot more to optically balancing weight than just matching stem weights or considering stem weights in ratio to particular heights.
Let me illuminate, as an art director, the problems I encountered in this.
There is one absolute weight quality that is determined by the type-designer/font, not the typographer/layout application. It is the weight at which vertical stem width equals stem-to-stem distance.
Of course, this is a weight that it is best for type designers (and typographers using variable-weight fonts) to avoid, for reasons of optical dazzle, but nonetheless…