Does anyone understand how Microsoft Word and Adobe Acrobat share information about font size?

Very interesting. I just tested this myself. Does the same thing.

Replication details:
[ul]
[li]Mac High Sierra public beta[/li][li]Word for Mac (O365) 15.39[/li][li]Acrobat Pro DC for Mac, part of my CC subscription, version number too long to paste[/li][li]Arial in TrueType format, origin unclear[/li][/ul]

11 pt in Word, 10.8 in Acrobat. Digging in now. Should be fun.

ETA: @DPRK

Could well be. Typography is not my field even a little bit. Altough I have done battle with trying to control pixels vs. displayed content size on a web page pre-CSS. Talk about a fool’s errand in service of a bigger fool of a customer.

You and I simulposted last time. I thought about ETAing that I wasn’t trying to rebut you, just commenting on the OP’s most recent post. Sorry if I sounded disagreeable.

Good to see Pork Rind come up with some independent confirmation. Sounds like the problem isn’t a quirk of the OP’s specific versions.

Here’s a potential immediate win…

If I create the PDF from Word by printing the doc and selecting “Save as PDF” from the PDF drop down in the bottom left of the dialog, I get 10.8 pt type as viewed in Acrobat.

Instead, in Word if I go ‘save as…’ and select “PDF” in the export section of the file formats list, I get 11 pt type in Acrobat. Note that I also left the default choice active “Best for electronic distribution and accessibility”.

I’m happy to keep digging, but perhaps this does the trick?

You’re not the only lab that’s struggled with this, mischevious.

edit: I see Pork Rind has discovered the same work-around as in my link. Oh well, I’ll leave it as it may provide comfort to know they are not alone in this fight.

Interesting further data point:

If in Word, in the ‘Save As’ dialog mentioned above, after selecting PDF, if you pick “Best for printing” it all goes to hell again.

Running a preflight diagnostic on these files shows that files with …

Creator: "Microsoft Word" 

… come out fine, but files with …

Producer: "Mac OS X 10.13 Quartz PDFContext" 

… come out gooched.

Yes, we also tried the Word “Save As” function, as opposed to “Print->Save as PDF” function. However, we still ran into problems:

  1. My version of Word does not allow me to choose between “best for electronic” and “best for print” options. Which version of Word are you using? The only option I have gives me 10.8pt. I can switch versions, of course, except:

  2. My PI’s version of Word does allow you to switch between “best for electronic” and “best for print” when saving as a PDF, and as you noticed, the font registers as 11pt when you save with the “best for electronic” setting. However, the images come out looking fuzzy and their colors look like mud. It’s particularly obvious on diagrams, less so on pictures. It looks from Inner Stickler’s link like other people have noticed that too. It’s not terrible, but not especially professional looking.

I’m starting to remind myself of certain posters, who shall remain unnamed, who ask for advice and then reject every possible solution because it wasn’t exactly what they wanted. I do very much appreciate the workaround options.

It’s so frustrating! The font takes up the exact some amount of space on the page, it just registers differently on the PDF. If we could trust that the NIH functionary is printing it out and using a pica gauge, that would be fine, but they’re probably just looking at the font size in Acrobat.

That is comforting, yes. I keep feeling like I’m doing something wrong.

I’m gonna keep working on this. It bothers me personally now.

Conspiracy theory time: Someone in the Microsoft-Apple-Adobe triangle is trying to burn the NIH down.

If it turns out to be a problem with Apple (this Quartz PDFContext component) rather than Microsoft or Adobe, then I owe Microsoft an apology. I did say it was only a hypothesis that it was Word’s fault.

Pork Rind: did you try the trick of opening the actual PDF file to check the size declarations?

This is why I love the Straight Dope. Perfectionists.

Disagreeable perfectionists!

I tried opening the PDF in TextEdit, but I couldn’t find this information in any way that I could understand. There were 78 lines that looked like the one that you posted (<</Type/Page/Parent 2 0 R/Resources<</Font<</F1 5 0 R/F2 7 0 R/F3 10 0 R/F4 12 0 R/F5 17 0 R>>), and it wasn’t clear which ones referred to what text, and they all contained many font sizes. Interestingly, even though the entire document is supposed to be 11.0pt only a few contained “11 0”, which I assume is 11.0pt, and none contained “10 8”, which I would guess is 10.8pt.

This is completely plausible. However, I think it’s slightly more complex than that: most biologists preferentially use Macs*, but in order to afford Macs, you need funding. If you already have a Mac, you likely received at least one NIH grant in the past, and people who have received a previous NIH grant are substantially more likely to get another one**. So Adobe is trying to level the playing field by getting Mac-users grants rejected, which gives new investigators a leg up. It’s totally logical.

*I think that Mac use is a legacy of when Windows was shit for image processing. If your professional life depends on clear images, there used to be a real difference, but isn’t so much anymore.

**This can be viewed as “rewarding high-quality research” or “protectionism from jerks who already have enough already”, depending on how what medications you’re on today.

Not yet, as I am at least in theory still at work. But it’s on my list of things to do.

Back in the early 90’s, I was a pre-press guy, and back then that often meant reading PostScript files to find the source of an error. It’ll be interesting to see if I still got it, so to speak.

But to be honest, I’m more interested in why Quartz is doing this, as I’m pretty confident that the PostScript itself in both examples is clean. The preflight that I ran on both examples is clear that there’s no internal scaling going on, so I’m confident that the declaration will be 10.8 or 11, respective to the bad and good outputs.

Interesting observation #1:

Created a file that contains a phrase, set six times. Once at 10 pt, once at 11pt and once at 12pt, and once with at True Type font (Arial) and once with an Open Type font (Hobo, if you must know).

Output both as “Print: save as PDF” and “Print: save as Adobe PDF”.

10 -> 9.84
11 -> 10.8
12 -> 12

What the fuck?

Okay, I’m sorry, this made me laugh like a loon. Welcome to the land of What The Fuck, my friend! I really appreciate the company.
Oooh, additional testing following on that observation: 6, 12, 18, 24, and 36 point fonts render exactly, while 6.5, 8, 10, 14, 22 points do not. Hypothesis: there is something special about multiples of 6.

LSLGuy is correct. As goofy as I think twips are, at least they are defined as 1/20th of a point – whatever that is.

And, as DPRK says “there are American points, Didot points, new Didot points, metric points, and so on.”

This is why I said “there’s some history” behind the point. I have no idea how Gutenberg defined a point, nor how it was defined in 1885. But in 1886 Ottmar Mergenthaler invented the Linotype machine. The industrial revolution was great, but machining tolerances at the time were not capable of five, much less six, decimal point accuracies. So in 1886 Mergenthaler’s company arbitrarily defined the point to be .014 inches. That’s right: .014 inches. It was that value from then until the mid 1960’s when phototypesetting machines came into being. I have no idea how the point was redefined again in the 60s, but it was and became .013837 or as DPRK indicated, 72.27 points per inch. Adobe is in fact just the next guy who had a reason to redefine a printers point to some convenient value.

For what it’s worth, 1/72 inch is a good number. It’s an integer number of units per inch and is thin enough to be a useful distance without fractions. I always hated 72.27.

Just as a fun way to close the loop on this while I continue to try to look for a cause/solution, up until about three years ago, I had a working Linotype in the garage. Which meant I had three different pica poles. One with Adobe points, one with ‘standard’ points and one with Linotype points.

My partner’s ever-so-useful input:

"So what I don’t get is why you aren’t using a proper professional publishing tool to prepare these documents…

…a decent Linotype hot metal secondhand is not that expensive these days."

It’s not! Say the word and I can have one on a truck headed your way in no time! I’ve managed to save five from the scrapyard so far. I’d be glad to add to the total.

The guy the directed the Linotype moviefigured that mine was one of fewer than a hundred still operational. Let’s get another one restored!

And yeah, you’d have a much easier time with InDesign, that’s for sure. I used to weep with fear whenever a sales rep would come back to my office with Word files to be set up for the press. I know things are a little different today, but that trauma runs deep.

I always liked those Monotype machines. I don’t like lead poisoning, though.