W3C home > Mailing lists > Public > public-i18n-bidi@w3.org > January to March 2010

RE: [html-bidi] Feedback on Additional Requirements for Bidi in HTML

From: CE Whitehead <cewcathar@hotmail.com>
Date: Thu, 18 Mar 2010 20:53:38 -0400
Message-ID: <SNT142-w27003DDABC681360E8CAB5B32A0@phx.gbl>
To: <ehsan@mozilla.com>, <public-i18n-bidi@w3.org>, <ntounsi@gmail.com>

Hi.
 
> From: ehsan@mozilla.com
> Date: Wed, 17 Mar 2010 17:43:17 -0400
> Subject: Re: [html-bidi] Feedback on Additional Requirements for Bidi in HTML
> To: cewcathar@hotmail.com
> CC: public-i18n-bidi@w3.org
> 
> It's not about secrets! If HTML specifies a default algorithm, all
> browsers will be using that
>From what I gather, from your discussion with Vladmir and Fantasai, word-count should not be the default algorithm
because it costs a lot of processing -- but if it were set to be the default algorithm,
instead of first strong, 
in your opinion, would all browsers/users/developers then be using word-count as the algorithm of first choice?
 
> (and at the worst case, authors can just
> test how the browser handles things, and infer the algorithm used.)
Possibly.  
> It's just that AFAIK, nowhere else in the HTML spec do we see
> algorithms being requested in source code, so the mere notion of that
> will be new to web developers.

I think elsewhere the algorithms are transparent in html.  
 
 
(I can't see any reason for not making these two algorithms transparent though if you have plans to modify them significantly . . . then I see a case where you might not want to specify exactly one or the other; maybe developers should be able to specify word-count-mod for modified word-count, or simply auto if nothing else works, but I hope they will specify which.  I am not sure about Vladmir's autodirtype: 
 what about just auto, which would go to a default -- auto-word-count say;

you might also have auto-first-strong, auto-word-count-mod [modified], auto-IMADEUPMYOWN)

 

For the bidi isolate name:
I like bdi-sce o.k. Or just plain sce. Actually the name is not so important; you all are accustomed to our strange names anyway)
That's my two cents.
 

 

I completely agree with Najib that whatever it is named, elements enclosed by links should also be bidi isolates!  (Good catch, Najib.)


Best, and best wishes,
 
 
--C. E. Whitehead
cewcathar@hotmail.com
> 
> --
> Ehsan
> <http://ehsanakhgari.org/>
> 
> 
> 
> On Tue, Mar 16, 2010 at 8:56 PM, CE Whitehead <cewcathar@hotmail.com> wrote:
> >
> >
> > Hi.  Again, I really think it's fine to have dir rtl auto-first-strong or
> > similar in source code (I think I understand the issues correctly too).
> > More comments are below!
> > From: Ehsan Akhgari <ehsan@mozilla.com>
> > Date: Fri, 12 Mar 2010 19:55:21 -0500
> >
> >>Also, I'm not a huge fan of specifying different algorithms as values
> >> for the dir attribute.  I think relying on web authors to figure out
> >> what algorithm to use can be very fragile, and it would be safe to
> >> assume that if they understand the issue well enough to determine
> >> which algorithm to use, they can probably come up with their own
> >> implementation anyway.  I think in practice having a single attribute
> >> value of dir=auto is much more useful, especially given the fact that
> >> a large portion of web developers have very little understanding of
> > ? the issues existing with supporting bidi text.
> > I think web developers can understand these two (though of course there are
> > a wide range of web developers, but those developing applications for bdi
> > already understand some of these issues it seems).
> >
> > The issue is when someone else ends up processing input text; for example
> > when I input text, it's displayed with "dir-auto-first-strong" or some such,
> > and then some other application picks up the displayed text; it might be of
> > interest to have as input which algorithm was used.
> >
> > And it's interesting to know why your text displays as it does too.
> >
> > (For most applications implementation details are transparent enough; in
> > Word for example, it's obvious that setting the font-color to auto displays
> > default color choices to text and links; no further details are needed;
> > similarly, for html, I know right away that the style codes set as css
> > properties are going to be displayed on elements with the appropriate names
> > and need no more information to interpret the code;
> > however this algorithm, since there are two choices, would be an exception
> > to this rule.)
> >
> > In addition, I personally like being able to view as much source code as
> > possible -- that -- with tutorials -- is how dummies like me learn.
> >
> > I suppose we are forcing browser makers to give away secrets, but when there
> > are only two choices, how much of a secret is there?  If it is needed so
> > other applications can process text, then it's useful.
> >
> > What we are really doing by making these two algorithms transparent, is
> > encouraging improvement -- and the users can say either "first strong
> > worked" or "first strong did not work" -- and that's helpful.
> >
> >
> > Best,
> >
> > C. E. Whitehead
> > cewcathar@hotmail.com
> >
> >


* * *


 		 	   		  
Received on Friday, 19 March 2010 00:54:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 19 March 2010 00:54:13 GMT