Re: SVG Feedback on HTML5 SVG Proposal

As a quick comment from a somewhat different perspective, I have been
following the SVG Developers mailing list for years.

First Point:
We currently have a series of "why is SVG broken?" or "why is Browser
X?" questions that pop up relatively regularly. Quite a few of these
questions are related directly to sloppiness in the way ASV 3 handled
SVG.

Many members of the list spend a fair amount of time having to educate
people over and over again because they are not doing something
correctly. They refuse to believe it because "it worked over there".

So in addition to the cost to implement, a little thought might be
given to the poor volunteers who support the technology and have to
answer why this pseudo-html/svg stuff works in HTML but doesn't work
when it is real SVG.

Second Point:
As a long time programmer, I've seen the same issue with compilers that
don't follow "the standard". The argument is always "it will be easier
for the user," but the result I've normally seen is bad habits that
persist for years (maybe decades). (And don't get me started on voodoo
programming practices. We've had plenty of that in early HTML.)

Third Point:
I also have a question about the sophistication of the users. If the
casual user is more likely to use a tool to write their SVG, they won't
care about the stricter syntax, they'll never notice. Weirdos like me
who often use vim to develop this kind of document are more likely to
swallow the technical argument (or that's how it seems to me).

G. Wade

On Wed, 11 Mar 2009 10:34:38 -0500
Jeff Schiller <codedread@gmail.com> wrote:

> On 3/11/09, Robin Berjon <robin@berjon.com> wrote:
> > On Mar 11, 2009, at 02:31 , Dailey, David P. wrote:
> >
> >  Just think of the case in which you're trying to teach someone how
> > to make web pages. You can say "if you put a <p>, you'll get a
> > paragraph; and if you put a <rect>, you'll get a rectangle".
> > That'll fly. But now you have to say "okay, so here you can do <p
> > id=foo> but there you have to do <rect id='foo'>". How do you
> > justify that? Learning the stuff is already
> 
> How about we teach them to quote all their attributes?  In my opinion,
> there are already attributes in HTML that we should be teaching new
> developers to quote (@title, @alt).
> 
> Has anyone considered if new HTML developers will find it confusing
> that they need quoted attributes if they want to use spaces or quote
> characters in the value?  Though it seems obvious to me and surely
> obvious to parser coders, I wonder if new users will find that at all
> confusing.
> 
> What are the thoughts on:
> 
> - svg:path's @d attribute
>   - do we teach users to quote them in SVG-in-HTML mode if they want
> to make semi-sane use of whitespace ?
> 
> Also, there are SVG attributes which do not seem to allow commas to
> separate values:
> 
> - @preserveAspectRatio [1]
>   - modify SVG syntax to allow commas to separate the
> defer/align/meetOrSlice values? (spec is not entirely clear on this
> attribute's grammar)
> 
> - @requiredExtensions [2]
>   - modify SVG syntax to allow commas?
> 
> I guess these types of modifications belong in the SVG 1.1 and 1.2
> spec?
> 
> My biggest concern with the quote-less attribute syntax is the
> breakage when copying from a hand-typed SVG-in-HTML document to a SVG
> tool.  There are already quote-less (albeit experimental) examples of
> SVG starting to proliferate around the web [3].  If all desktop
> browsers support a "Save As SVG" or "View DOM" option as Jonas brought
> up then this would be less of a concern to me but that is not the case
> today.  Furthermore, it will still be confusing to developers who are
> used to View Source (another way of thinking about this is that "View
> Source" is no longer 100% reliable and we have to train developers to
> "View DOM" instead).  The other option is to rewrite the parsers on
> all SVG tools (if that's even possible - will Adobe be willing to do
> that with Illustrator?).
> 
> Alas, I already know which way the WG will head on this.  Quoteless
> attributes are just too cool to pass up.
> 
> Regards,
> Jeff
> 
> [1]
> http://www.w3.org/TR/SVGTiny12/coords.html#PreserveAspectRatioAttribute
> [2]
> http://www.w3.org/TR/SVGTiny12/struct.html#RequiredFeaturesAttribute
> [3] http://hsivonen.iki.fi/svg-filters-html5/
> 


-- 
There are two ways of constructing a software design: One way is to
make it so simple that there are obviously no deficiencies and the
other way is to make it so complicated that there are no obvious
deficiencies. -- C. A. R. Hoare

Received on Thursday, 12 March 2009 03:01:21 UTC