[WIP] [designspaceLib] designspace docs edits #1750

justvanrossum · 2019-10-18T10:50:46Z

misc clarifications and edits from Erik
proposed addition to allow rules be applied after other features. Relates to [designspaceLib] [varLib] Allow FeatureVariations to be processed *after* other features #1747 and addFeatureVariation doesn't work on a glyph that's also being swapped by another opentype feature #1371

- added docs for DesignSpaceDocument object, methods and attributes - removed comments on validation, localisation and generating UFO instances. - added note that axis minimum, default and maximum are in userspace coordinates. - added clarification to map input (userpace!) / output (designspace!) values. - added note that sourceDescriptor location is in designspace coordinates. - moved comment on rule subs to rule descriptor object. - added proposed "processing" flag to rules element

- move note on sub element

Add comment

justvanrossum · 2019-10-18T10:54:35Z

Would the new processing attribute in the rules element need a format version update? Say 4.1?

(See also #1749)

…ttools#1750, as well as the related doc.rulesProcessingLast flag

justvanrossum

Apart from the import statement in the example, this looks good to me.

Doc/source/designspaceLib/readme.rst

belluzj · 2019-10-21T08:43:21Z

Doc/source/designspaceLib/readme.rst

-Default font
-============
-
-  The source with the ``copyInfo`` flag indicates this is the default
-   font.
-  In mutatorMath the default font is selected automatically. A warning
-   is printed if the mutatorMath default selection differs from the one
-   set by ``copyInfo``. But the ``copyInfo`` source will be used.
-  If no source has a ``copyInfo`` flag, mutatorMath will be used to
-   select one. This source gets its ``copyInfo`` flag set. If you save
-   the document this flag will be set.
-  Use ``doc.checkDefault()`` to set the default font.


Is this bit of information going to be moved to another part of the document? Or is it removed completely?

Same for other deleted bits, to me they were interesting information

Maybe these need to moved to a paragraph marked deprecated, rather than be removed?

I'm sorry for removing it without proper discussion, I guess I was trying to clean the document up. The motives for removing this text were:

Finding or assigning the default font in a designspace used to be more informal and automatic, so there was code to support that. But the role of the default has become much more prominent and it should be assigned explicitly by setting its location and the axes to specific values.

This was formalised a while ago when the default is understood to be the font located on the (mapped!) default value on each axis. The author of the designspace has to make sure this is the case, there are no reliable methods of guessing what the default should be.

Assigning the default by setting the copyInfo flag then only creates ambiguity.

I could add a paragraph explaining that there may be older designspaces in which the copyInfo flag could mean that the source was intended to be the default.

I could add a paragraph explaining that there may be older designspaces in which the copyInfo flag could mean that the source was intended to be the default.

That sounds perfect, thanks.

Fine, we'll keep the copyInfo flag, but remove the note about that it assigns the default.

The default master in the axis is defined in variable font coordinates but that masters can have a different range. So I need to first go through all axis, collect all default coordinates, transform them to the master coordinate space and hope that I find a master with that coordinates. This is to complicated and introduces to many possibilities for failure.

I agree this is not great, the decision was made a while ago by @behdad. It is outside the scope of this PR.

On the other hand, using the axes default makes .designspace potentially more useful. We can potentially have software that uses .designspace files plus a compiled VF to produce “slices” (subset) etc. I know it’s a bit annoying at first but I think the axes default method is conceptually cleaner.

The default master on the default values of the axes is very clear. It can be a problem that these values are in userspace coordinates. As long as there is no axis map this is fine because user space == design space. But adding a map can easily move the expected location of the default. That just means editors need to keep track of that. I updated mine.

To come back to the second part of Jany's question, after the decision to keep copy* in:

The copy* flags:

if absent, copy * from the default font.

if present, copy * from the source that has the flag if you feel like it, or don't, but don't change the source for the default outlines (which is always determined by the axis defaults).

Is that correct?

anthrotype · 2019-10-21T12:41:35Z

Doc/source/designspaceLib/readme.rst

+----------
+
+-  ``processing``: flag, optional. Valid values are [``first``, ``last``]. This flag indicates whether the substitution rules should be applied before or after other glyph substitution features.
+-  If no ``processing`` attribute is given, interpret as ``first``.


in the other thread @twardoch raised a valid point about the undefined meaning of processing=post when generating instance UFOs from a designspace. Should the spec say something here?

The "UFO instances" section about rules currently doesn't say anything about other glyph substitutions, so I think a general clarification of how to deal with that would be good.

An important detail is (under "5.0 rules element") "The rules are evaluated in this order". This is currently not true with varLib.featureVars, and I'm not sure it should be. The order is only relevant if rules react on each other (eg. rule N+1 acts on the output of rule N).

While this may be fixable (generate substitution lookups in the order of the original rules), I'm not sure what the use case is.

(Perhaps this discussion deserves its own issue.)

- implementation differences - varlib vs. mutatormath - older versions - rules and generating static ufo instances Updated the description of the `copyInfo` flag of the sourceDescriptor.

…ttools#1750, as well as the related doc.rulesProcessingLast flag

Various changes in the designspaceLib readme.rst - added docs for DesignSpaceDocument object, methods and attributes - removed comments on validation, localisation and generating UFO instances. - added note that axis minimum, default and maximum are in userspace coordinates. - added clarification to map input (userpace!) / output (designspace!) values. - added note that sourceDescriptor location is in designspace coordinates. - moved comment on rule subs to rule descriptor object. - added proposed "processing" flag to rules element - move note on sub element - implementation differences - varlib vs. mutatormath - older versions - rules and generating static ufo instances - Updated the description of the `copyInfo` flag of the sourceDescriptor. - Change the example import. - Remove additional mention of copyInfo.

LettError added 5 commits October 18, 2019 11:48

WIP

ab543c7

- move note on sub element

WIP manual line wrapping?

963e87a

WIP

c723439

Add comment

WIP rename the branch

d7b25c1

justvanrossum added a commit to justvanrossum/fonttools that referenced this pull request Oct 18, 2019

Implement <rules> attribute 'processing' according to spec update fon…

d42b7d7

…ttools#1750, as well as the related doc.rulesProcessingLast flag

justvanrossum commented Oct 20, 2019

View reviewed changes

Doc/source/designspaceLib/readme.rst Outdated Show resolved Hide resolved

belluzj reviewed Oct 21, 2019

View reviewed changes

anthrotype reviewed Oct 21, 2019

View reviewed changes

LettError added 3 commits October 28, 2019 17:06

Collected notes that were removed in the previous edit in section 8

93e2e23

- implementation differences - varlib vs. mutatormath - older versions - rules and generating static ufo instances Updated the description of the `copyInfo` flag of the sourceDescriptor.

Change the import.

e4d0f3a

Remove addional mention of copyInfo.

bb74ac8

anthrotype approved these changes Oct 28, 2019

View reviewed changes

justvanrossum and others added 3 commits November 7, 2019 19:04

GPOS was meant here, GSUB should be left alone

2579829

WIP change 2 headlines, remove sentence about older files.

7e1edbb

WIP change 2 headlines, remove sentence about older files.

c18af69

justvanrossum merged commit b8fafdd into master Nov 7, 2019

justvanrossum deleted the WIP-designspace-docs-edit branch November 7, 2019 18:30

simoncozens pushed a commit to simoncozens/fonttools that referenced this pull request Mar 10, 2020

Implement <rules> attribute 'processing' according to spec update fon…

5b2067d

…ttools#1750, as well as the related doc.rulesProcessingLast flag

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] [designspaceLib] designspace docs edits #1750

[WIP] [designspaceLib] designspace docs edits #1750

justvanrossum commented Oct 18, 2019 •

edited

Loading

justvanrossum commented Oct 18, 2019

justvanrossum left a comment

belluzj Oct 21, 2019

belluzj Oct 21, 2019

LettError Oct 26, 2019

LettError Oct 26, 2019 •

edited

Loading

justvanrossum Oct 26, 2019

LettError Oct 28, 2019 •

edited

Loading

twardoch Oct 28, 2019

LettError Oct 28, 2019 •

edited

Loading

madig Oct 28, 2019

LettError Oct 28, 2019

anthrotype Oct 21, 2019

justvanrossum Oct 21, 2019

justvanrossum Oct 22, 2019 •

edited

Loading

justvanrossum Oct 22, 2019

[WIP] [designspaceLib] designspace docs edits #1750

[WIP] [designspaceLib] designspace docs edits #1750

Conversation

justvanrossum commented Oct 18, 2019 • edited Loading

justvanrossum commented Oct 18, 2019

justvanrossum left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

LettError Oct 26, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

LettError Oct 28, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

LettError Oct 28, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

justvanrossum Oct 22, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

justvanrossum commented Oct 18, 2019 •

edited

Loading

LettError Oct 26, 2019 •

edited

Loading

LettError Oct 28, 2019 •

edited

Loading

LettError Oct 28, 2019 •

edited

Loading

justvanrossum Oct 22, 2019 •

edited

Loading