Re: @image for pdf and html

[Gavin Smith]

On Tue, Nov 24, 2020 at 11:11:33PM +0100, daniela-spit@gmx.it wrote:
> > Sent: Tuesday, November 24, 2020 at 10:55 PM
> > From: "Gavin Smith" <gavinsmith0123@gmail.com>
> > To: "Christopher Dimech" <dimech@gmx.com>
> > Cc: "help-texinfo gnu" <help-texinfo@gnu.org>
> > Subject: Re: @image for pdf and html
> >
> > On Tue, Nov 24, 2020 at 10:11:09PM +0100, Christopher Dimech wrote:
> > > It has been figured that we cannot use the same file for both
> > > pdf/dvi output and html output.  Because if the file has too many
> > > pixels, the image will be too big for html.  We could end up with
> > > many calls to image with different files using @iftex and @ifhtml.
> >
> > You could equally say it's too big for pdf, if the pdf is going to
> > put online.  It's the same image either way.
> 
> It is html that is limiting resolution, thus capability must be discussed when
> compared to pdf output.  Html provides serious limitations, unlike in pdf
> where you can select the width and height.  In geology based manual, the 
> resolution
> is the most important aspect of any image.
> 
> > I don't see anything wrong with using conditionals for different
> > output formats to specify different image files, if that is what
> > is desired.
> 
> There is nothing wrong.  However geology based manuals inherently
> have many images, and defining for both is an extremely cumbersome
> proposition. Besides the fact that images in html have limited use
> because of the number of pixel limitation.  Otherwise you cannot
> see the text.

What might be a good feature in texi2any for your usage is to display
the image in HTML with a low-resolution version (both for page
layout, and to save bandwidth), but have a link to higher resolution
versions.  This is what happens on Wikipedia.  This would be quite a
bit of work to implement, though.