RE: New tutorial for REVIEW: Creating Bidi XHTML/HTML Pages

Hello Mati,

Thanks for your comments !  I've been working on them for some time.  Better
late than never, as they say.

See below...

Richard Ishida

> -----Original Message-----
> From: Matitiahu Allouche [] 
> Sent: 16 August 2005 10:09
> To: Richard Ishida
> Cc:
> Subject: RE: New tutorial for REVIEW: Creating Bidi XHTML/HTML Pages
> A few comments. 
> 1) On slide 3 (and others like it), I originally thought that 
> the blue lines were links to specific topics.  In fact, 
> clicking anywhere in the image leads to the next slide.  This 
> is ok, but fooled me at first.  You may want to warn the readers. 

Hmm. I'll try to fit this into the template later.

> 2) On slide 6, you could note that specifying dir="rtl" does 
> not affect the paragraph direction of the title bar.  
> Depending on the browser and the platform, adding <RLE> ... 
> <PDF> around the title text may give it an RTL paragraph direction. 


> 3) On slide 8, you write "dialog boxes, such as the one shown 
> on the slide, will be mirror imaged." 
> As far as I know, this applies only to alert message boxes 
> created with Javascript.  I don't know what other types of 
> dialog boxes there may exist for an HTML page.
> Maybe the example would be more piquant if it displayed a 
> wholly or mainly English message, whose ending punctuation 
> would appear on the left side (I am not sure that in your 
> current example, naive users will notice that the Hebrew and 
> Arabic strings have swapped places around the "W3C" string). 

I changed the wording slightly to try to capture these points.

> 4) On slide 10, it might be useful to remind the reader that 
> the example shows text with RTL paragraph direction.  You 
> might also explain the meaning of the digits below the 
> arrows, or remove them. 
> In the last paragraph, the reference to 2nd and 3rd line are 
> inverted: the logical encoding order is shown on the *second* 
> line and the visual encoding order on the *third* line.

All fixed. Thanks for spotting.
> 5) On slide 11, you may want to mention the ISO encoding by 
> its precise IANA name, "ISO-8859-8". 


> 6) On slide 13, "the that the" => "that the" 
> I also suggest to add an hyphen between "ISO" and "8859" in 
> the list of encodings facing "Visual" and "Logical". 


> 7) In slide 16, if the example included more than one word 
> for each language, you could add that the relative order of 
> words in a "language run" is also preserved. 

True, but we haven't explained that stuff yet, and I think it may confuse
some people.

> 8) In slide 20, I don't see the benefit of pinpointing that 
> reducing bandwidth "may be an important consideration in 
> countries where Arabic is spoken".  It is equally applicable 
> to countries speaking any language, but somehow implies that 
> Arabic speaking countries might have scarcer bandwidth 
> resources than other countries. 
> Also, the goal is not to reduce "bandwidth" but to reduce 
> "bandwidth requirements". 

Ok. Changed.

> 9) On slide 26, I suggest to replace the sentence "Space and 
> punctuation...either type of script." by: 
> "Spaces and punctuation are not strongly typed as either LTR 
> or RTL in Unicode, because they may be used in either type of 
> script."

> 10) On slide 27, you should add that the last sentence ("Note 
> also that alongside...") applies only to European digits. 

Hmm. Investigating...

> 11) On slide 30, the sample HTML code (in red) shows the 
> exclamation mark on the left of the Arabic words.  I don't 
> know *any* HTML editor that would show this in source editing 
> mode, so it may give false expectations to readers.  On the 
> other hand, showing the exclamation mark on the right side 
> seems to contradict what we are trying to explain.  Whatever 
> solution is chosen, a note should explain the potential 
> difference between text as displayed in source mode and text 
> displayed in WYSIWYG mode. 

Hmm. Yes, this is one of those very difficult things to deal with when using
examples to explain content authoring.  I added a note.  

Btw, Martin Durst has been working on getting editors to do the right thing
in this kind of situation, and I'm crossing my fingers in the hope that the
upcoming release of Xmetal with bidi support will also fix this.  We'll see.

> 12) On slide 31, same remark: the exclamation mark and the 
> entity (&rlm;) or numeric reference are likely to be 
> displayed on the righ side of the Arabic words by a real-life 
> source editor. 

Another note added.

> 13) On slide 34, the last sentence ("Putting markup around 
> the comma...") is not correct: the markup solution would be 
> to put markup around each of the Arabic words.  The 
> egg/hammer metaphor is thus less appropriate. 

Hmm. Good point.  I removed the sentence.  I'm not sure that markup works at
> 14) (removed) 
> 15) In slide 39, same comment as 11 and 12 above: the HTML 
> code in red shows the content of the span as if the editor 
> was capable of applying the markup while displaying the 
> source, which is over-optimistic and gives false expectations. 

I added text to the last note saying "This will apply to all following
examples of this type.", since it seemed a bit like overkill to keep saying

> 16) In slide 41, I believe that the red text should say 
> "</span>" (and not "<span>"). 

Correct. I'll change it.

> 17) In slide 41, the last paragraph says "This is not a bug". 
>  True, this is not a bug in the display agent, but it is a 
> bug in the source text.  The space is supposed to separate 
> the Hebrew part from the following English part, so it should 
> not be included in the Hebrew span. 

Ok. Added 'in the display algorithm'.  

> 18) On slide 43, the code example is correct in the last 
> line, but the dir attribute is inverted in the red text. 


> 19) On slide 50, the codes for RLE and LRE are swapped (both 
> in the image and in the table within the text). 

Fixed. Thanks !

> 20) In slide 51, "Unicode control character have" => "Unicode 
> control characters have" 


> 21) On slide 52: is there a rule against referencing 
> documents other than W3C's?  If not, a reference to the 
> Unicode Bidi Algorithm (Unicode UAX #9) seems in order. 

Added in the text.

Once again, Mati, many thanks for such a thorough review.  Lot's of useful


> Shalom (Regards),  Mati
>           Bidi Architect
>           Globalization Center Of Competency - Bidirectional Scripts
>           IBM Israel
>           Phone: +972 2 5888802    Fax: +972 2 5870333    
> Mobile: +972 52 2554160

Received on Monday, 8 May 2006 19:45:33 UTC