W3C home > Mailing lists > Public > www-style@w3.org > November 2007

Comments on media queries

From: Henri Sivonen <hsivonen@iki.fi>
Date: Thu, 15 Nov 2007 18:47:38 +0200
Message-Id: <C9C6EFF0-3E14-4C3B-93DC-64FE26AA8E4C@iki.fi>
To: W3C Style List <www-style@w3.org>

  * The device dimensions seem potentially harmful. Most often authors  
should really be querying the view port. Querying the device  
dimensions implies that the content is presumed to be important enough  
to warrant asking the user to resize the view port to expand to fill  
the available space.
   - If device dimensions querying is kept (I guess it is too late to  
take it out considering that Opera has shipped it already), I think  
the spec should advice authors to query the viewport unless they have  
a really good reason to query the device.
   - Tutorials that I have seen query the device for no good reason.  
That's not a good sign.

  * There's no way to query the aspect ratio of the viewport.

  * Considering that both viewport and device queries are available,  
it would be more useful for the viewport queries to query the page  
area than the page box in cases where page margins are either fixed by  
the UA or have already been computed otherwise at media query time.
   - The reference for page box should be updated to point to CSS3  
Paged Media.

  * The spec should probably say that device queries on print media  
query the paper size.
   - This is a problem considering that CSS can define the page size  
itself, but the spec already has this problem since it refers to the  
page box.

  * Various values don't have any validation provisions about negative  
values. This is fine from the UA point of view for everything except  
ratios of two negative numbers, because non-sensical values fail to  
match anyway. However, it might make sense to treat negative values as  
authoring conformance errors.

  * Similarly, it might make sense to consider grid values over 1  
conformance errors.

  * Resolution should say explicitly "device pixels" to avoid  
suggesting that CSS px is in any way involved.

  * Is it a good idea to treat the interlaced vs. progressive scan  
mode issue as something that the author should care about and be able  
to query? Since interlaced display at low refresh rates in already a  
legacy mode of operation, shouldn't dealing with legacy TV interlacing  
problems be something that a UA--not the author--takes care of?

Henri Sivonen
Received on Thursday, 15 November 2007 16:47:56 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:31 UTC