I added an upload button to the add media wizard.
So that when you visit a page like:
then click on the "image" button, then click on the "upload button" in
the add-media-wizard, you should get a list of recently uploaded files
and a few links to inline upload.
From the inline upload you can upload a local file. Then you can click
"insert" to insert the uploaded file to the page. The idea is to make
the process fairly strait forward.
This just got developed over the last few days, let me know if you run
Already on the todo list:
An auto-complete category field and multiple language description
support to the upload description. Ideally I want to avoid a very long
list of things people need to fill out before they can upload content
but that of course needs to be balanced with getting as much description
info as possible at point of upload ;)
Neil and the Multimedia usability efforts are diving into this problem
so I won't focus on that to much.
This builds on mwEmbed AddMedia module, so all the goodies like progress
indicators and inline transcoding with friefogg is supported. ( so you
can select your cameras h.264 encoded file and insert directly into the
article with transcoding, cross site uploading, asset reference and
layout handled by the wizard. )
This uplaoder works via an iframe so it can share credentials across a
white-list of domains. In the future it would not be hard to add in
support for client initiated domain approval ( similar to facebook
connect ) .. this would enable a lot of different mash-ups. Obviously
the proxy is useful for any widget/ gadget that want to do cross-site
api requests that require "post".
Also I should email wikitech-l. If we want to support cross-domain IE
interface interactions we need a p3p policy header to be sent out to IE:
I have proposed a way of organizing our watch of recent uploads on Commons:
This is very much pre-alpha for now, but of course many admins are
already familiar with the job itself. It would only be a tool to
rationalize this task and make it more efficient.
Anyone is very welcome to comment on the page or improve it (it remains
minimalist for now).
Any admin or trusted user is very welcome to register and take a few
watches from time to time.
Any bored botmaster is welcome to volunteer for automating the register
by adding new days and archiving old ones (to be coded only when the
format is stable).
Now that I have your attention :-)
I've cooked up a simple tagging system for Commons. It uses the
category system by prefixing tag "categories" with TAG, e.g.
[[Category:TAG:Flower]]. TAG categories will be "flat", so
intersections are easy. Merging of equivalent tags could later be
possible through redirects, without having to bot-edit all affected
pages. Also, redirects and/or language links could allow for
multilingual tags, or at least multilingual tag search/intersection.
For the frontend, I put the tag section into the sidebar, which will
make it instantly familiar to many web users. On page load, the TAG
categories are removed from the normal category section and shown as
tags in the sidebar. This also means no additional web traffic will
arise for normal page viewing.
Tags can be removed with a single click. One or multiple tags can be
added or removed in a single operation. All editing is done via API,
so the page does not have to reload.
When checking one or more tags, a section containing a link to
"subset" will show. Currently, it links to my CatScan rewrite, showing
all pages in the current namespace that also have all the checked
tags. Linking to CatScan gives you the opportunity to add further
conditions; you can do a combined subset of TAG and "normal"
categories, templates, etc.
Of course, there could also be a dialog window or fake special page
showing e.g. images in the subset, using CatScan only as a backend. I'
holding back on that one...
The script is making extensive use of jQuery, so this script will only
work when jQuery is loaded. This is currently only the case for the
Vector skin, AFAIK. So, don't complain if it doesn't work for you in
I have tagged the total of one example page so far (picture by me, so
noone else unduly harmed:-) Here you go:
Before the traditional (f)lamewar begins, let me say that I want to
thing, or CatScan. I just want to present a working tagging system
within the current software framework, as is. I hope that, now that
there is at least demo code, something more useful than the usual "but
it also needs to make coffee!" discussion will arise.
P.S.: Yes, there are probably a million bugs in the code...