v2: Regular quick diff #9

Activity

1. mjlagattuta commented on Aug 3, 2018

   

   mjlagattuta

   on Aug 3, 2018

   Contributor

   More actions

   • No the UPM has not been changed, the source ufos I started with had a upm of 2048
   • Yes, I added anchors and changed the position of accents to harmonize with the height of the tittle, additionally some double accent composites were modified because they originally appeared to be mechanically scaled down versions of the relevant diacritics
   • Any glyphs that are missing are most likely do to converting to a glyphs file, where I converted glyph names to allow for the autogeneration of features. This includes .smcp to .sc as well as using strokeshortcomb instead of crossbar
2. mjlagattuta commented on Aug 6, 2018

   

   mjlagattuta

   on Aug 6, 2018

   Contributor

   More actions

   @davelab6 After chatting with @m4rc1e on a private chat, I wanted to get confirmation on what levels of tolerance/matching the original we are shooting for, especially given Roboto Slab had a set updates made that never got pushed.

   A couple of notes on the setup and why there may be a number of differences:

   • The ufos up on github were updates that had never been pushed, and I used those as the starting point for the VF remastering
   • Given the addition of a black weight (extrapolated and corrected) we had discussed the potential for a 3-master setup (see Parity with Roboto’s weights #7). The file in the repo uses Thin, Regular, and Black as masters (effectively making Bold an instance)
   • In working on the file, I performed some cleanup where possible. With kerning, I removed unused classes and merged identical values into classes. I also believe some metrics and kerns were modified in the ufos I started with (see Metrics and Kerning Inconsistencies #8)
   • Additionally I set all relevant glyphs and marks up with anchors and added components where possible. While doing this I replaced some glyphs with those used in a typical glyphs setup (replaced /crossbar with /strokeshortcomb, /tail with /tail-cy, etc.)
   • One area I am concerned with are changes that were made in the ufos that are less visible

   Additionally, we had discussed doing the hinting for these projects in VTT. @m4rc1e informed me that the current static Roboto Slab fonts served use ttfautohint so I am not sure if I should hold off until ttfautohint can handle VFs, or if you want to move forward with VTT
3. davelab6 commented on Aug 6, 2018

   

   davelab6

   on Aug 6, 2018 via email

   Member

   More actions

   On Mon, Aug 6, 2018, 7:48 AM Mike LaGattuta ***@***.***> wrote: @davelab6 <https://github.com/davelab6> After chatting with @m4rc1e <https://github.com/m4rc1e> on a private chat, I wanted to get confirmation on what levels of tolerance/matching the original we are shooting for, especially given Roboto Slab had a set updates made that never got pushed.

   I believe we should be making stuff better quality, and anything that deviates from the current version but is better is fine. Even if it's very disruptive; the future is longer than the past. The fontdifferenator and diffbrowsers tools give us insight into what changed, and confidence that what changed is better, and didn't unintentionally change anything. A few families we want to be much more careful about changes because they are system fonts, but even Roboto was radically changed in 2014/15 and there was less outrage than Montserrat last year. Roboto Slab is pretty popular, but it isn't set in stone. A couple of notes on the setup and why there may be a number of differences:

   - The ufos up on github were updates that had never been pushed, and I used those as the starting point for the VF remastering Yes, good - Given the addition of a black weight (extrapolated and corrected) we had discussed the potential for a 3-master setup (see #7 <#7>). The file in the repo uses Thin, Regular, and Black as masters (effectively making Bold an instance)

   Sure, should be good

   - In working on the file, I performed some cleanup where possible. With kerning, I removed unused classes and merged identical values into classes. I also believe some metrics and kerns were modified in the ufos I started with (see #8 <#8>) Yep I think they were, and moving to classes is great - Additionally I set all relevant glyphs and marks up with anchors and added components where possible. While doing this I replaced some glyphs with those used in a typical glyphs setup (replaced /crossbar with /strokeshortcomb, /tail with /tail-cy, etc.)

   Great

   - One area I am concerned with are changes that were made in the ufos that are less visible If less visible, why more concerned? :) Additionally, we had discussed doing the hinting for these projects in VTT. @m4rc1e <https://github.com/m4rc1e> informed me that the current static Roboto Slab fonts served use ttfautohint so I am not sure if I should hold off until ttfautohint can handle VFs, or if you want to move forward with VTT

   It seems VF support in ttfautohint is coming along so I'm ok to have Roboto Slab 99% Ready To Go except for hinting and then circle back to it when we get to publication time and decide if ttfautohint is acceptable. Of course I prefer it as a libre solution.

   …

   👍React with 👍1Reacted by Steve Powell
4. mjlagattuta commented on Aug 6, 2018

   

   mjlagattuta

   on Aug 6, 2018

   Contributor

   More actions

   If less visible, why more concerned? :)

   In converting from ufo to .glyphs there was a portion of user data that as far as I know goes unused by Glyphs, as well as extra custom parameters such as "OS2strikeoutPosition" and "OS2strikeoutSize" which I believe are also ignored by glyphs.

   I was unsure how important such information was and if there was a way I should be implementing it into the glyphs source properly
5. davelab6 commented on Aug 6, 2018

   

   davelab6

   on Aug 6, 2018

   Member

   More actions

   Ah yes, right. I guess that stuff is in the current TTFs too, and there are font level Glyphsapp Custom Parameters you can use to set those values if you want to set them manually. I wonder if the Glyphsapp defaults for them are even different from what's in the TTFs
6. mjlagattuta commented on Aug 6, 2018

   

   mjlagattuta

   on Aug 6, 2018

   Contributor

   More actions

   The original file I started with imported those values for the first ufo I opened in Glyphs (Roboto Slab Thin) but much of them were greyed out:

   

   When cleaning up the file I removed the ones that were greyed out. I believe Glyphs generates these values and they seem to be different as shown in first diff above comparing attribute numbers
7. davelab6 commented on Aug 7, 2018

   

   davelab6

   on Aug 7, 2018

   Member

   More actions

   much of them were greyed out

   Not sure what that means. Have asked (https://forum.glyphsapp.com/t/greyed-out-custom-parameters/9525)

   Basically I think its fine to forget about that stuff.
8. mjlagattuta commented on Aug 7, 2018

   

   mjlagattuta

   on Aug 7, 2018

   Contributor

   More actions

   Thanks

   I was thinking the same, based on the assumption that it being grayed out meant that Glyphs was just storing that information but it was not actually going be used in the creation of the ttf.

   However, since the font is generated with fontmake I'm not sure how that deals with the greyed out custom parameters
9. davelab6 commented on Aug 7, 2018

   

   davelab6

   on Aug 7, 2018

   Member

   More actions

   Have asked (forum.glyphsapp.com/t/greyed-out-custom-parameters/9525)

   Have answer: greyed out CPs are inactive, their key's string is not used by GlyphsApp, but because you can have such "private" CPs, they are kept in the file
10. davelab6 commented on Aug 7, 2018

    

    davelab6

    on Aug 7, 2018

    Member

    More actions

    since the font is generated with fontmake I'm not sure how that deals with the greyed out custom parameters

    Right; also not sure.