Recent Posts

Pages: [1] 2 3 ... 10
I can say that about a month ago I got a Python/FontTools script written to "fix" fonts for compatibility with Word Mac 2011, based on Adam's guidance earlier in this thread. I've used it on six large families totaling over 100 fonts, and I have yet to have any problems with it.

The only problem I know of so far is that the "newly fixed" fonts are not necessarily compatible with the old ones insofar as the naming has changed. But that problem is unavoidable, as far as I can tell.
Erm, I'm afraid currently TransType 4 is doing some GPOS kerning "optimization" that is not always good. In essence, it always rebuilds kerning classes based on the kerning value, and redistributes the GPOS kerning pairs to those new classes.

We thought this was a good idea but it turns out there are problems with this approach. We will make this "optimization" optional and work on it some more.

Also: if you have made complex families that you know work well, send them to us snd tell us what you've done, and we'll very likely documdnt these methods and model our app behavior after your best practices. I'm really not claiming that I *know* how to do this right, and I'm constantly trying to find that out, and document the process :)
Ah, that must be it. :) I would *love* to finally get this done right in our apps, and publish some info about this. Unfortunately, we often get only "signals" that something's not working right. It's a fairly complicated world and I have to refer to you guys — because you're the ones making these fonts.

Please report these problems, and please be as specific as only possible (OS versions, apps names and versions, OS and app language settings, FontLab tool names and build numbers, font export settings, screenshots, development and final font files, Python code, your workarounds, notes and comments). I realize that it takes time for you to document such things but I hope this is time well-invested. You'll have more time when this will be resolved and your lives will be more stress-free (and so will ours :) ).

It's simply impossible for our small team to test through all possible combinations with all kinds of tools. We develop these guidelines in good faith, hoping that they'll help people fix things before we put them into final app code.

Note: TransType 4 and the upcoming next-generation FontLab Studio do all this automatically, of course.
TransType 4 !? Can you please explain what TransType 4 is really doing to kerning? :-)
don't forget to use 400 and 700 as US weight class

I don't think this is unusual at all, many typographic design solutions use high contrast, and users want convenient shortcuts. In Adobe fonts, Black was often style-linked to Light. Of course, that was in the days of FOND resources (Mac PostScript), where style linking was much much more flexible then today.

within 2 weeks after accurately following your method, customers told us fonts were not working, we went back to our old production method and everything was fine again. The reason you weren't able to replicate the problem that I was reporting, is that I never precisely reported the problem :-).
Ps. I'll add that a number pf font vendors and I myself have found the methods I described above to work more correctly than the older methods. I know that Lucas feels otherwise but we haven't been able to replicate the problems he's reporting.
Agreed that this is a very unusual move, and is likely going to cause problems somewhere.
The community recommendation is that the styling link name is always one of "Regular", "Italic", "Bold" and "Bold Italic". If you stick to this recommendation, you can rely on 10+ years of testing fonts made in such a way by many vendors on many platforms. You can also choose not to follow this community recommendation, and it's possible that your fonts will behave correctly in all, most or many enviroments — but you need to do all the testing yourself. That's how community recommendations work. They're the "conservative" side of things: "do like others have done and you're safe".

But of course you're encouraged to be progressive and experiment around. Community recommendations are not the "only" way, or even the "best" way — they just aim to be the "safe" way, so in a way, they're the opposite of progress. It's very healthy if someone chooses to challenge them, but then it's that person's task to do all the testing work. If you choose that path, we will, of course, be happy if you share your findings! :) This is how progress happens!
I would have to look at the font itself to be sure, but that sounds good so far. Besides what you mentioned, one might also be well advised to test in:

- a "normal" Mac app such as TextEdit
- an Adobe CoolType app such as InDesign
- one or more MS Office apps on Windows

Finally, you could run the family through the CompareFamily test app from the Adobe AFDKO tool kit. It checks a bunch of known issues about both how individual fonts are built, and how they should relate in terms of style linking, etc.


Pages: [1] 2 3 ... 10