While such a feature sounds quite nice to have, it poses several problems:

• It would be a violation of CC and GFDL to offer such an archive without

including information about license, creator, uploader, history, etc. Basically,
the image description pages would have to be packet along with the files.

• Compression does not work well on bitmap files like png/jpeg/gif (because they

are already compressed). It would be a huge waste of cpu cycles to try. Perhaps
a non-compression format like tar could be used, or such files could be stored
in the zip-archive with no compression.

• Creating such an archive takes quite a bit of resources, and may take to long

to be done "live" in an HTTP request. It would probably be a good way to
implement some type of throtteling and/or caching; Perhaps it would be best to
do this as an "asynchronous" service: registered users could request an archive
and would receive an email with a download link when it is done.

• Traversing subcategories recursively may have unintended effects - one

erronous categorization could cause several thousand files to be included, even
though they are not really wanted. It may be a good idea to show all categories
that would be included, along with a total number of files and a total
uncompressed size to the user to confirm - but that may already be too expensive
to compute "live". But there must also be a hard limit - someone *will* try to
download *everything* - for a large wikipedia or commons, several terrabytes, if
I recall correctly.

So, the request is reasonable, but not trivial to implement.

hugo.lpz wrote:

Enable this function at least for administrator seem good.

Copyright : a table with [ License | Uploader | ImageName ] seem to be need.

hugo.lpz wrote:

An Idea : [ ImageName | License | Uploader ] can be merge in a new dowloaded
file name, such as :

• On commons : [ Image:France_map.png | CC-sa-fr | Uploaded by Piom ]
• downloaded on you computer : France_map-CC-sa-fr-Piom.png

Problem, I really have no idea how do such thing.

hugo.lpz wrote:

what about this request ?

hugo.lpz wrote:

Commons really need this functions, at least for admin, otherwise categories such http://commons.wikimedia.org/wiki/Category:Bw.png_stroke_order_images or http://commons.wikimedia.org/wiki/Category:Coats_of_arms_of_France , where users have involve so much energy to accumulate this images, are IMPOSSIBLE to download, and so totally useless.

Bryan.TongMinh wrote:

Since this requires extracting data to obtain author and license information, it is more a task for the toolserver.

Another possibility is to add all revisions of the image description page to the archive, and have a static set of files added to the image (licenses).

Raising priority, this is a very requested feature which prompted the creation of the partial workaround https://toolserver.org/~platonides/catdown/catdown.php (linked from Commons default sidebar).

(In reply to Nemo from comment #7)

Raising priority, this is a very requested feature which prompted the
creation of the partial workaround
https://toolserver.org/~platonides/catdown/catdown.php (linked from Commons
default sidebar).

Sorry, but I'm going to re-lower this to low priority. Well it would be cool, it would be quite complicated to implement, both technically (Some categories have over a TB worth of images in them...), and socially with regards to license compliance. At the same time this hardly seems like a critical feature to commons

See also this related suggestion at meta.