[HN Gopher] PNG Specification (Third Edition) W3C Working Draft ...
___________________________________________________________________
PNG Specification (Third Edition) W3C Working Draft 20 July 2023
Author : robin_reala
Score : 45 points
Date : 2023-07-20 12:09 UTC (10 hours ago)
(HTM) web link (www.w3.org)
(TXT) w3m dump (www.w3.org)
| chungy wrote:
| Is this by the same people or is it a competitor to the PNG
| development group (http://libpng.org/pub/png/)?
|
| The latter's website still heavily says that PNG is a still image
| only format, and promotes MNG as the animation alternative. Of
| course, that's not how web browsers have played it out, and it
| also appears to me that the libpng source code still lacks "APNG"
| support. This makes me think it's a competing standard to the
| original developers.
| robin_reala wrote:
| Changes since the 2nd edition of PNG:[1]
|
| - APNG added to the spec
|
| - A new CICP chunk that paves the way for HDR support
|
| - EXIF part of the main spec
|
| - Various other fixes
|
| [1] https://www.w3.org/TR/png-3/#changes-20031110
| Pinus wrote:
| EXIF... If I remember correctly, EXIF uses the TIFF file
| structure. Am I right? Does this mean that a PNG file, with its
| relatively simple format, would have inside it an EXIF chunk
| with a potentially different byte order and absolute offsets
| all over the place?
| mistrial9 wrote:
| > EXIF
|
| .. because we all need UUIDs and device origin on every image
| on the net?
| refulgentis wrote:
| EXIF is an arbitrary metadata schema. Nothing about it
| implies or mandates UUIDs and device origin.
|
| UUIDs, I've never seen in EXIF.
|
| Device origin, if I interpret as "camera model" or "GPS if
| enabled", I have seen.
| somat wrote:
| I guess my question is, PNG already has an arbitrary
| metadata chunk, what does EXIF bring to the table that is
| not already there?
| mattl wrote:
| No but adding attribution and licensing to an image would be
| very useful.
| Daneel_ wrote:
| Reliable date information too.
| MontagFTB wrote:
| The chunk is ancillary, so if size matters it can be omitted.
| codetrotter wrote:
| Size is not the problem. Privacy is.
| pengaru wrote:
| If you're relying on file formats to preserve your
| privacy you've already lost.
| naikrovek wrote:
| then don't add that chunk to your files...
|
| or write a tool which strips it... it's an easy format.
|
| feel free to complain, but also take action, or it's all
| wasted breath and whining. no one wants to hear people
| describe their problems poorly while also taking zero
| action to fix those problems.
| antisthenes wrote:
| So make sure your image processing pipeline strips out
| all metadata?
|
| Seems kind of a non-issue and AFAIK most image hosting
| software and social media already strips out metadata
| anyway for this reason (I know Discord does, not sure
| about others)
| 8n4vidtmkvmk wrote:
| Pretty sure there was metadata in there already.
| Stablediffusion has been saving data into there for
| awhile now. It's not evil in intent. Just strip it out.
| rhn_mk1 wrote:
| XMP metadata: https://dev.exiv2.org/projects/exiv2/wiki/T
| he_Metadata_in_PN...
| KerrAvon wrote:
| EXIF is valuable metadata for many use cases. Location
| data and fingerprinting are absolutely problems and the
| solution is to strip that information before you post
| images anywhere. Having it in the spec is a net win for
| privacy because it provides a standard chunk for it so
| that tools can universally strip it out.
| tracker1 wrote:
| Worked on a small charity project a few years ago, that
| included coordinating images with their geolocation for a
| specific trail to mark signage, etc. Was definitely
| useful. Not useful was the image orientation displaying
| differently in different browsers.
___________________________________________________________________
(page generated 2023-07-20 23:03 UTC)