RE: Minutes, June 9 2011 SVG WG telcon

With regard to the text positioning effects I've been raising, I think that
three things are important concurrently: the ability to see text in SVG as
text, so that it remains searchable and accessible, to enable considerably
more control over the shape and alignment of glyphs, and to keep it all as
vectors rather than as bitmaps.


The reasons for doing these things, I've talked about some in the past, but
basically lots and lots of text in the wild (logos, advertisements, signs,
map labels) require text to confine itself to a curvilinear region. Consider
what happens when one draws some text in Adobe Illustrator and then applies
an "arch warp effect." While still in ai format, the text remains text and
is mouse-selectable. Upon export to SVG however, the appearance is preserved
but the glyphs are converted to paths and accessibility is gone.


Suppose, for example, we had the following ability: 

<text id="T" style="font-family:ariel;font-size:16">

<textShape topalign="path1" bottomalign="path2">Hello, here is some text
shaped by two bezier curves.</textPath>

< /text>


I think this would handle most of the effects allowed in either Illustrator
or Microsoft WordArt, together with a much broader array of flexibility.
Like <textPath>, text would remain text, and vector based, but with
considerable flexibility in styling, shaping and confining to geometries
such as geopolitical regions on maps.


Clearly there are hard cases, where truly complex curves (path1 and path2)
might set up nasty edge cases, but subject to certain constraints (like that
the associated line segments that identify getPointAtLength(Q) as Q
traverses path1 and path1 never intersect) I think the majority of the
common textual-graphic effects that designers use could be embraced.


In the interim, if the call that several people seconded here a few months
back for exposing glyphs shapes to javascript, then a proposal for how
something like a textShape might work could be sketched out as Cameron had
recommended. For building something like Fontographer as  a web app, or for
mobile designers to know where rendered text actually appears on the screen
so that adjustments could be made, seems quite natural.


Another small step in the right direction would be (as Israel Eisenberg
pointed out) to have some function that acts like Opera interprets
getExtentofChar in
(hit retype and then topjustify) . I think maybe Opera got it right and the
spec got it wrong in this case.


Currently in aligning text to a (baseline) curve in SVG, the characters are
never distorted, merely positioned and rotated. To accomplish the full
fledged dual-path approach to alignment, one will probably need to apply 3D
transforms to the glyphs, but that will be a part of SVG2, anyhow, so it is
not too far a stretch perhaps?





From: [] On Behalf Of
Rik Cabanier
Sent: Thursday, June 09, 2011 4:56 PM
To: www-svg
Subject: Minutes, June 9 2011 SVG WG telcon


Present: Cameron, Doug, Chris, Rik
Chair: Cameron
Scribe: Rik

                               - DRAFT -
                   SVG Working Group Teleconference
09 Jun 2011
   See also: [3]IRC log
     * [4]Topics
         1. [5]SVG 1.1 Second Edition
         2. [6]Seattle F2F form
         3. [7]CSS WG testing
         4. [8]Text layout
         5. [9]FX taskforce
     * [10]Summary of Action Items
   <trackbot> Date: 09 June 2011
   <shepazu> [11]
   <heycam> ScribeNick: cabanier
   <heycam> CM: let's get started
SVG 1.1 Second Edition
   DS: did you catch the out of date reference?
   CL: I did not. didn't think it was prudent to change
   ... can you still change?
   DS: yes you can
   <ChrisL> can change in response to ac review comments
   <heycam> CM: I saw philippe mentioned our CSS2.0 reference, did you
   explain that to him?
   <heycam> CL: yes
   CM: there are 4 weeks to comment and then will become recommended
   there weren't too many people commenting. There is a culture where
   if people don't comment, they imply yes
   don't panic if there are no comments but it would be better if
   people do
   (that was doug schepers)
Seattle F2F form
   CM: form is supposed to be done at the end of the mont
   ... send a reminder to get topics for the venue
   ... one of the agenda items is the FX topics but we will wait until
   Vincent calls in
CSS WG testing
   <heycam> [12]
   DS: in the testing taskforce there is a consensus to use the stuff
   that Peter Linss has done for CSS
   ... he wrote a framework to aggregate the tests
   the framework allows to run a subset of the tests. when running test
   you can run all tests or just a couple
   DS: I want him to come talk to us because we feel strongly to have
   unified testing
   CM: should one of us call in?
   DS: I'll follow up with him
   ... he responded off-list that he wanted more information
   <scribe> ACTION: shepazu to follow up with Peter Linss about testing
   [recorded in
Text layout
   ds: we have to wait for Vincent to be here
   CM: I made a recent proposal for positioning
   ... I didn't read Rik's email reply yet
   <heycam> RC: it seems that svg is trying to do two things at once
   <heycam> ... it's doing layout for you, but also wants to provide
   you with the ability to position glyphs yourself
   <heycam> ... it's a little bit the same thing with ligatures
   <heycam> ... it'll create ligatures for you, but then also wants to
   give you the ability with positioning not to create ligatures
   <heycam> ... so it seems a bit messy
   <heycam> ... I propose that maybe we should have a new type of text,
   more like pdf, specifying individual glyphs taht you want to use
   <heycam> ... so you can choose the ligatures
   <heycam> ... and you always specify glyph positions
   <heycam> ... and the existing <text> will be automatically laid out
   <ChrisL> sounds like a glyphlist rather than text - so would need an
   'alt' for accessibility
   <heycam> ... and it can create ligatures on the fly for you
   <heycam> ... so people who want a simple interface can still have
   that, but people like us, who want more control over glyph selection
   and positioning can have that as well
   <heycam> CM: we may need to have glyph indexing
   <ChrisL> the interaction between altglyph and opentype layout tables
   is underspecified
   <heycam> CM: but we do have <altGlyph> and <glyphDef>
   DS: do we have properties to control if ligatures are formed?
   <ChrisL> CSS3 has a way to disable discretionary ligatures
   CM: true
   ... but, I don't think that turns of mandatory ones
   DS: ok
   CM: I'm not completely well versed in the area.
   DS: seems like a pain
   CL: I understand that the idea behind the 2 modes
   CM: do we want 'abc' or indexes into the font file
   CL: not sure
   RC: going the PDF way might be very difficult
   CM: not if just using glyph ids
   RC: true
   CM: SVG is not ideal for absolute position but there is a need for
   positioning on character based input.
   ... Rik do you have an opinion
   RC: I was looking at it from absolute positioning
   CM: we need to keep allowing positioning in the current SVG text
   ... and make it as accessible as possible
   CL: one of the reasons that was added was for adding more
   complicated text on a path
   RC: and every glyph could have its own matrix
   CM: yes by specifiying a rotation
   ... maybe we should present this at the face-to-face
   ... I need to implement it anyway and I will start off with my
   ... I will give more detail at the F2F
   ... for the use case of absolute positioning, we should try to
   support better and someone should start on a proposal
   ... rik, do you want to take this on?
   RC: I will discuss with Vincent but Adobe's interested.
   CM: using glyph indices is probably a good way
   RC: using glyph indices requires you have the font
   CM: I think that is reasonable.
   ... that the font is always there when you do absolute positioning
FX taskforce
   DS: want to discuss the CSS and SVG charters
   CM: should we wait for vincent
   DS: we'll talk about it via email
   RRSAgent make minutes
   RSSAgent, make minutes
Summary of Action Items
   [NEW] ACTION: shepazu to follow up with Peter Linss about testing
   [recorded in
   [End of minutes]
    Minutes formatted by David Booth's [15]scribe.perl version 1.136
    ([16]CVS log)
    $Date: 2011/06/09 20:50:05 $

Received on Friday, 10 June 2011 00:58:11 UTC