The Model M and Model F Font

[micrex22]

Decided to break this in a dedicated thread as to not spam the Model F remake group buy as it's going off topic. I don't know how in-depth anyone has ever bothered to investigate the specific font used on Model Fs and (most) Model Ms. With the exception of the special characters the font used is indeed 'Helvetica' as many say, but a variant that I could not find anywhere. In fact, it may very well not exist in a modern digital font format. I was wondering if the Selectrics had font golf balls with this specific variation of Helvetica but I'm not a Selectric aficionado and couldn't find a convenient chart outlining every single font golf ball.

This particular rendition of Helvetica is notable as it's a lot easier to read and has some departures from the original Helvetica as it's based on Helvetica Light (which was created by a different typographer). The Helvetica IBM uses is a crossover between Helvetica Light and Helvetica Rounded, but not exactly the same as either. IBM even used this font on some Personal System/2 documents as shown here:


What's confusing is that they switch from this unique Helvetica to Helvetica Regular in documentation from the period. And because I am certain it's not a priority for anyone to pay attention how many times IBM rapidly jumps through Helvetica variations, it's subtle at times. Not to mention IBM later flat out adopted Helvetica Regular on subsequent keyboards after the Model M--but kept the special characters and arrow keys. Sadly, they dropped the cool Model M / F thin-tailed shift arrows in favour of conventional ones.

First here's an example of IBM's Helvetica (bottom) versus plain regular Helvetica (top) Unicomp uses:

It's bolder and has prominent edges, whereas IBM's variant is very rounded off. You can see how much attention to detail was put on the Model M, the original team even customized the spacing between the lettering for optimum visual readability: whereas Unicomp (and even IBM's modern keyboards) just used the default character spacing you find with Helvetica Regular.

I've attempted a somewhat similar effect by using Helvetica Light and applying a 'rounded' outline to achieve more boldness and roundness. While pretty close the ratios are still off as they're baked into the font (has nothing to do with how it's printed on either paper or key caps). I then attempted with Helvetica Rounded and while that's more along the lines of what we're trying to achieve, it's a little bit too bold and squashed vertically for some reason, therefore my last mimic key attempted to stretch Helvetica Rounded out a bit.

Top shift has been placed off-centered to mimic Unicomp's uneven printing as a joke

It would be cool if anyone had any more information than I was able to dig up--it's possible this is a documented form of Helvetica.

Now that I recall... "IBM Helvetica" was included in some of their operating systems as a default system font. However, none of them have that soft roundness. It should be possible to recreate The Model F/M "Helvetica" based on more scans of documents with it and references.

Why does it matter? I'd have to say it's my favourite variant of Helvetica, and I definitely prefer it over some of the more visually 'busy' fan made prints:

[emdude]

Fascinating post, thanks for writing it up. I would certainly like to see if there is more documentation on IBM's Helvetica as well.

This would probably be of great interest to Ellipse since he's trying to reproduce XT-quality key caps.

On another note, the standard Helvetica that Unicomp uses makes my eyes bleed, and it doesn't help that they can't position their legends to save a life. They did seem to use the IBM Helvetica typeface for their key caps into the late 2000s from what I've seen, so I guess they must have been using the old dye-sub stencils until just recently.

EDIT: For example, a Unicomp M5-2, from 2008:

s-l1600.jpg (129.85 KiB) Viewed 4108 times

[Hak Foo]

I wonder if it's less an explicitly "built" version of the font and more just artifacts of the technology of the time. It might have been actual physical type somewhere.

[fohat]

The lack of spacing between letters is what makes me nuts.

I learned "graphic design" in the 1970s and we were taught that generous spacing was important for both esthetics and readability. For example, we learned that on something like signage there should be space for the capital letter "O" between words.

These days, presumably to "save space" everything seems to be squeezed together ruthlessly.

[Scarpia]

First here's an example of IBM's Helvetica (bottom) versus plain regular Helvetica (top) Unicomp uses:

Spoiler:

Interesting post. I do have one objection to that example though: while I have to applaud the effort to add letter spacing for improved readability, the gaping chasm between the 'f' and the 't' is just plain bad kerning. To me, that makes it look just as bad as the squished-together (but at least consistently kerned) regular Helvetica on the Unicomp legend.

Or is it just me?



There, I fixed it:



(still not perfect, but the 't' is no longer fleeing the scene)

[emdude]

First here's an example of IBM's Helvetica (bottom) versus plain regular Helvetica (top) Unicomp uses:

Spoiler:

Interesting post. I do have one objection to that example though: while I have to applaud the effort to add letter spacing for improved readability, the gaping chasm between the 'f' and the 't' is just plain bad kerning. To me, that makes it look just as bad as the squished-together (but at least consistently kerned) regular Helvetica on the Unicomp legend.

Or is it just me?



There, I fixed it:



(still not perfect, but the 't' is no longer fleeing the scene)

Whoa, déjà vu.

[micrex22]

I wonder if it's less an explicitly "built" version of the font and more just artifacts of the technology of the time. It might have been actual physical type somewhere.

I thought this initially too, however the design clearly has characteristics which cannot be achieved exclusively by artifacts: it's still not possible to get an exact recreation by simply distorting Helvetica Light (if you look at the original 'Shift' key and pay attention to the tail, it has a curve which is more extreme than Helvetica Light), although you can come a lot closer than Helvetica Regular!

This font seems to have been used quite heavily in this era, especially in printed documents. I've discovered IBM was not the only one to use it heavily. However, whatever (devices) responsible for printing it remain a mystery. People involved with printing of that generation would know... but at this point it's lost information.

I happened to notice the 'WorkPad Z50' box has used Helvetica as the copyright text as well. Here it clearly has artifacts from the bleeding of the ink and gives us a good idea that the 'Model M' font has a fair bit of clarity going for it:

Interesting post. I do have one objection to that example though: while I have to applaud the effort to add letter spacing for improved readability, the gaping chasm between the 'f' and the 't' is just plain bad kerning. To me, that makes it look just as bad as the squished-together (but at least consistently kerned) regular Helvetica on the Unicomp legend.

I am certain this was a deliberate design choice to make the characters on 'Shift' more readable; the logic behind it would be so that the strike on both 'f' and 't' wouldn't collide or be visually congested (which is quite the attention to detail). On the pad printed M13 labels this specific character spacing persisted and since they had to create a new system for pad printing this re-affirms that it was deliberate and not simply bad kerning. HOWEVER, the later Model M dye-sub re-tooling after '89 seems to have omitted the easier to read spacing:

Why the pad printed labels (bottom) kept the original character spacing but the second gen dye-sub (top) did not is anyone's guess (maybe part of the design team was allowed to dictate 'proper' character spacing was present with M13 pad printing production, but when the new redesign happened with thinner legends the manager at the time decided to revert it). The M13 pad printing would have been at the Maxiswitch factory too, so original IBM equipment from '86 wasn't involved.

[y11971alex]

That wayward 't' may not have been quite so uncommon.

[Ratfink]

Interestingly, on my Displaywriter beamspring, left shift is "SHIF T" and right shift is "SHIFT".

[Scarpia]

Interestingly, on my Displaywriter beamspring, left shift is "SHIF T" and right shift is "SHIFT".

Q.E.D.

Attention to detail my ass. That wayward 't' is and always was bad kerning, and if it is indeed intentional on IBM's part (I agree that the consistency across different legend printing technologies makes this seem likely) then it could be a copy protection device for telling genuine IBM buttons apart from unauthorized clones. I'm only speculating here of course, but I can't believe those legends were set like that for either aesthetic or legibility reasons.