6/5/19 core meeting notes

Thanks to Brian for taking notes this week!

The call was recorded
https://benetech.zoom.us/recording/share/KHbDF9pIC0DkmH_h-m0_K4QRawSxhbDZ6Ggoq0xpRhWwIumekTziMw
If you have questions about the notes or want more details, please check
the recording and audio transcript, imperfect though it may be.

Workflow/issue resolution:
ns: suggested we add labels based on resolutions
fw: posted suggested labels "resolution needed"/"spec changes
needed"/"polyfill needed"/"implementation update needed"/"tests needed".
Every time something is finished, the label gets removed. When all labels
are gone, the issue is closed.
fw: I am fine with labels suggested there
ns: Any objections?
None
ns: resolved we will use the labels to track this and see how that goes


Issue 85: Additional Spec Data
bk describes
ns: I guess we can tentatively resolve to task David C to at least search
for a label and connect it to part of the spec. It would be great to link
to something helpful about how to write a test and locally what needs
testing
ns: spec could point to github and a landing page that bk is working on for
testing


Topic: trying to align with html5/css - 5 issues linked to this.
fw: we resolved some of them I think in last meeting
ns: it seems like Rob brought up some new points around colors/case
insensitivity
fw: we agreed to rely on HTML or CSS, but it seems that the implementations
are not interoperable currently
ns: rob you've been doing some research?
rob: I have been looking, I wanted to see what they did in SVG... I've
noted down my findings in #85 and what our experiments gave us, I think we
go with CSS and it is so far working very well.
ns: so are we pointing to CSS 3 colors?
rob: yes, it has all of the modern colors/etc
ns: I'm concerned that some systems won't support all of CSS. Maybe it
should just say 'authors are discouraged from' for (specified) things that
are outside of the web?  In the spec.
all: resolved to do that


Issue 62: we agree to go with color attributes

Issue 63: there are some length attributes that only make sense in a
browser.
ns: we should caution against using them in non-web contexts


fw: Numbers like "5." is allowed in MathML currently, but not in CSS. Can
we align? Resolved to align. Since the values are parsed by CSS, we should
follow the CSS spec (as opposed to HTML)

Issue 22 - agreed to make attr values case insensitive

fw: when we use the CSS parser for attrs, it is case insensitive in core
and full. Both specs should proably recommend it is lowercase to remain
backward compatible with existing tools. Agreed to

ns: we have a general goal of producing a shorter spec. A little concerned
about adding a lot of notes.
bk: can we just put a thing in core? 2 sentence note.
fw: seems ok
resolved to add quick rec notes on these in core encouraging back
compatibility when authoring

Issue 23
ns: isn't this just aligning with CSS
fw: mpadded is an issue
ns: yeah, it has all sorts of special cases. The values allowed for mpadded
are special kinds of lengths.
fw and ns: we can agree to disallow named spaces
fw: it looks like scriptsizemultiplier is the only non-length, non-integer
value outside of mpadded.
fw and ns: we need to check the spec to see if there are other numbers we
missed
all: lets create a separate issue about mpadded attrs, more general than
#81.
ns: Can we say for 23: 'with the possible exception of mpadded lengths, we
will use CSS for parsing values of attributes?'
fw: the relaxng schema doesn't use number, so I'm not sure we can search
for number to know which attributes use number... someone should go through
it carefully.
ns: ok, we're not resolving 23 at the moment, but we think what we said
above +  scriptsizemultipler/scriptlevel in full, and someone look for
others

Issue 28: Whitespace collapsing and trimming in content (not attrs/values)
ns: What do you end up getting from the parsers
fw: we get the spaces from the parsers
ms: Word strips them -- spaces before/after mi and mo usually isn't
sensible; done other ways
bk mentions leaving it seems consistent
  /fw shows copy paste example and describes that whitespace is preserved
there and innerText(), etc. /
resolution: further investigation of how the other stuff works


Issue #1 mstyle
ns: at the last meeting we agreed dropping some general attrs and mapping
other general attrs. What should happen to the rest that are allowed?
ns: fred can you describe your proposal?
fw: my proposal is to remove all attributes from mstyle
/ debates about what to do with this, we can't deprecate 'it' but what can
we do /
ns: recommend we say that is taken out of core and stays in full.
objections?
fw: notes that the polyfill is complex in that it needs to know what
attributes belong to what element and add them in to those elements unless
overridden.

bw: we should have a general meeting about strategies to get 'full'
implemented in browsers. Maybe we should be in discussion with ShadowDOM
group or others because currently they can only be used with names that
contain a '-' unless it is part of a whitelist.
ns: can you write something up and then I'll create an agenda around it
bw: agreed to try to get something written this week.

We will meet again next, same time.