Re: libpng complexity (Was: [PNG] Cancel meeting?)

On 2024-01-22 19:37, Leonard Rosenthol wrote:
>
> FWIW – Adobe uses libpng for all of our desktop and mobile 
> applications (for the web we let the browser do the work for us).
>
Yes, libpng (or a patched version therof) is in widespread use despite 
the known issues of complexity, efficiency and security. There needs to 
be a path forward for libpng-using applications.
>
> As John says, libpng is VERY old code designed to run on any platform 
> (not necessarily to do so optimally).  The original author/maintainer 
> (Glenn) passed away about 5+ years ago, and I am not aware of who (if 
> anyone) picked up that responsibility.
>
Cosmin Truta has been doing that.
>
> I have no objection behind review existing implementations and seeing 
> if one of them could be updated to be considered ”reference”
>
In general, W3C tries to avoid "reference implementation" preferring 
instead "reference specification which is proven to be implementable by 
two or more independent implementations".

The corollary of "reference implementation" is "in the case of a 
difference between the spec and the implementation, the implementation 
is correct" and we do not want that. I recall, for example, that the 
license for Macromedia Flash had words to that effect.

We should btw remove mentions of "reference implementation" from the PNG 
spec, because IIRC there are some, and it is being used incorrectly there.

> – but I think that a starting point would be for the maintainers of 
> any such implementations to join the W3C and this group.
>
Indeed. We would also welcome independent individuals who maintain such 
implementations to contribute to technical discussions, which are open, 
and people who contribute in a substantive, consistent and positive 
manner would be welcome as Invited Experts as well.

-- 
Chris Lilley
@svgeesus
Technical Director @ W3C
W3C Strategy Team, Core Web Design
W3C Architecture & Technology Team, Core Web & Media