Re: [css3-values] recommendation for pixel units

Le 20/03/2013 22:32, James Holderness a écrit :
> I'm not sure I follow your reasoning. You've got to make some choice of
> values when making these calculations. So for some reasonable viewing
> distance, you're going to calculate the reference pixel size, let's call
> that R; then given a device pixel size of P, the ideal number of device
> pixels per reference pixel will be R/P.
>
> Now R/P could be anything, but it's unlikely to be a whole number. The spec
> recommendation then is to choose "the whole number of device pixels that
> best approximates the reference pixel". But that can surely only be 1, 2, 3,
> etc. Not 1.5 or 0.75.
>
> Do you disagree that the spec text is recommending a whole number, or do you
> just think it doesn't matter if implementations ignore the recommendation
> and pick a multiple of 0.25 instead? (Or whatever - the Nexus 7
> device-pixel-ratio isn't even a multiple of 0.25)

I agree that this spec text is recommending a whole number, but my point 
is that, in my understanding of the "spirit" of the spec, the "some 
approximation" part is more important than "a whole number", which 
should be more of an example of a possible approximation.


> I understand that a recommendation is not a requirement. I just don't
> understand the reason for having the recommendation if nobody thinks it's a
> good idea.

I don’t think it’s a big deal since it’s not normative. I agree I 
couldn’t find a good wording. (I tried.) Suggestions?

I’m not even convinced that 96 CSS px per physical inch is such a good 
recommendation for typical printed paper. My viewing distance for 
hand-held paper is much smaller than 71cm.

-- 
Simon Sapin

Received on Wednesday, 20 March 2013 22:56:41 UTC