Use UFO point labels as source for axes

[djrrb]

I’m thinking of using labels com.typenetwork.parama-roundup.XXXX where XXXX is the axis name.

We can use these to fill the data on the input sheet, rather than entering it manually.

[djrrb]

I’m working on implementing the point labels implementation of parama roundup. So the imagined workflow is this:

1. Insert measurements as labels in UFOs
2. Run a script to find glyphs and point indexes and write them to existing Inputs sheet
3. Run a script to parse measurements from Inputs sheet and write them to Measurements and Widths sheets

The point labels are stored in a custom RoboFont lib entry com.typemytype.robofont.pointLabels. They can be written and read by any software, but are not part of the UFO spec or currently implemented in non-Robofont environments AFAIK.  

@dberlow Does all of this sound good? Also, how do you want to handle point labels for phantom points?

[dberlow]

The origin point should be one index larger than the last point of the last contour of the glyph and the width poont should be two index points larger than the last point of the cliff as per the truetype spec.

…

On Nov 17, 2019, at 10:00 AM, David Jonathan Ross ***@***.***> wrote: dberlow Does all of this sound good? Also, how do you want to handle point labels for phantom points?

[djrrb]

Yes, but the question is how do you want to use point labels on points that don’t exist? 🙃 On Sun, Nov 17, 2019 at 11:35 AM David Berlow <notifications@github.com> wrote:

…

The origin point should be one index larger than the last point of the last contour of the glyph and the width poont should be two index points larger than the last point of the cliff as per the truetype spec. > On Nov 17, 2019, at 10:00 AM, David Jonathan Ross < ***@***.***> wrote: > > dberlow Does all of this sound good? Also, how do you want to handle point labels for phantom points? — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#11>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABSGFLM76QVN7Y2K6CHK523QUFXDZANCNFSM4JNOSMVQ> .

[dberlow]

The points and goemetry that got you to The phantom point is the id?

…

On Nov 17, 2019, at 11:36 AM, David Jonathan Ross ***@***.***> wrote: Yes, but the question is how do you want to use point labels on points that don’t exist? 🙃 On Sun, Nov 17, 2019 at 11:35 AM David Berlow ***@***.***> wrote: > The origin point should be one index larger than the last point of the > last contour of the glyph and the width poont should be two index points > larger than the last point of the cliff as per the truetype spec. > > > On Nov 17, 2019, at 10:00 AM, David Jonathan Ross < > ***@***.***> wrote: > > > > dberlow Does all of this sound good? Also, how do you want to handle > point labels for phantom points? > > — > You are receiving this because you were assigned. > Reply to this email directly, view it on GitHub > <#11>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ABSGFLM76QVN7Y2K6CHK523QUFXDZANCNFSM4JNOSMVQ> > . > — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[djrrb]

@dberlow Yeah, before we were just typing ID numbers to a spreadsheet so it was easy to just type in slightly-bigger numbers for the phantom points. Now we are looking to label points directly in the UFO, which requires that a point exists in order to be labeled.

Possible options:

• Add anchors for phantom points (but then they have to be added to all masters)
• Write a script that labels phantom points as a custom lib entry (would not be visible in robofont)
• Hard code results with phantom points into the spreadsheet, bypassing labels altogether for those axes

[dberlow]

Curious, when RF generates a ttf, does it add the phantom points and then throw them away when it reads them? RF, UFO has to store the width (rsb) somewhere?, and technically the location of the origin (lsb) is known to all points;) which is why alignments are a single value. . . so... can't we find values for calculations even without adding them?

…

On Sun, Nov 17, 2019 at 9:14 PM David Jonathan Ross < ***@***.***> wrote: Yeah, before we were just typing ID numbers to a spreadsheet so it was easy to just type in slightly-bigger numbers for the phantom points. Now we are looking to label points directly in the UFO, which requires that a point exists in order to be labeled. Possible options: - Add anchors for phantom points (but then they have to be added to all masters) - Write a script that labels phantom points as a custom lib entry (would not be visible in robofont) - Hard code results with phantom points into the spreadsheet, bypassing labels altogether for those axes — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#11>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAO5VDUBXI237NSCMHZNT23QUH267ANCNFSM4JNOSMVQ> .

[djrrb]

Curious, when RF generates a ttf, does it add the phantom points and then
throw them away when it reads them?

AFAIK there are no phantom points actually added into the TTF file...those index are just understood and can be referenced in the hints. The glyph width is stored in the <hmtx> and the y valuesare stored in the <hhea> and <os_2>.

so... can't we find values for calculations even without adding them?

Yes we can calculate, but we can’t label at least in the same sense that we can label points. I’ll try to come up with a workaround.

[dberlow]

That is correcter, the scaler creates them from the hmtx table. Can we label a value in the hmtx table as/if it were a point?

…

On Mon, Nov 18, 2019 at 4:34 PM David Jonathan Ross < ***@***.***> wrote: Curious, when RF generates a ttf, does it add the phantom points and then throw them away when it reads them? AFAIK there are no phantom points actually added into the TTF file...those index are just understood and can be referenced in the hints. The glyph width is stored in the <hmtx> and the y valuesare stored in the <hhea> and <os_2>. so... can't we find values for calculations even without adding them? Yes we can calculate, but we can’t label at least in the same sense that we can label points. I’ll try to come up with a workaround. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#11>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAO5VDWLYE7UZPZPOEPRJIDQUMC43ANCNFSM4JNOSMVQ> .

[djrrb]

Not using the RoboFont interface, unfortunately. And that was kind of the whole benefit of the labels, to make the process more intuitive.

[dberlow]

Just forget it then. I tried 10 years ago to get phantom points added to RF so we could hint there, I failed to convince our developer or to get them to convince Frederik to do so. I guess RF is now not up-To-date with the needs of people making TT. On Nov 18, 2019, at 4:50 PM, David Jonathan Ross <notifications@github.com> wrote: Not using the RoboFont interface, unfortunately. And that was kind of the whole benefit of the labels, to make the process more intuitive. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.