nice work.
one thing to think about is parts of the description field that could be broken out into the medium field, or title field
(for example as i manually did in your example)
https://commons.wikimedia.org/w/index.php?title=File%3ABodyguard_of_Ranjit_Singh-_1838-1839_-_BL_Add.Or.1385.jpg&diff=118413205&oldid=118380423

it's probably going to be different for each institution, how they input their metadata, and how we structure it.

jim hayes


On Fri, Mar 7, 2014 at 8:30 AM, James Heald <j.heald@ucl.ac.uk> wrote:
A quick update.

I've been able to a find ways to help me clean up the layout of the wikitext on the description pages using the semi-automated AutoWikiBrowser tool; and also a less miserable approach to getting the page renaming done; so that I am *not* now planning any longer to do a full re-upload of the set, or indeed any re-uploading.

(In fact my hands were tied, because people were already starting to use and edit the pages, which a re-upload would have wiped).

So the wikitext layout on the pages is now all pretty much corrected, and I should have worked through renaming the remaining filenames by the end of Sunday.


A typical diff can be seen eg at:

https://commons.wikimedia.org/w/index.php?title=File%3ABodyguard_of_Ranjit_Singh-_1838-1839_-_BL_Add.Or.1385.jpg&diff=118380423&oldid=118137724

>From the top I have made the following changes:


*  Added an =={{int:filedesc}}== header above the Artwork template

*  Re-ordered the fields in the artwork template, and added blank ones

*  Moved the gwtoolset fields into a separate template, {{Uploaded with GWtoolset}}, which currently produces no output, but could be adjusted to output whatever you wanted.

*  Added whitespace before and after the new {{Uploaded with GWtoolset}} template

*  Split the categories directives each onto their own line

*  Added whitespace before the commented-out Metadata sections

*  I have also turned all instances of &#39; back into apostrophes.
(A single apostrophe has no significance for wiki-markup, and so does not need to be escaped.  A double apostrophe may well be intentional).


I didn't get the change perfect -- an extra newline got in at the top that shouldn't have been there; and the artwork template is nicer with a single space before the pipe character, which I forgot.  But it's good enough, and now feels to me like a proper WikiCommons page should.


A final thought about the inclusion of all the commented-out metadata. It's not ideal, because it can lead to category information being split between two places.  The natural place for categories is soon after the description, so that an editor can quickly read down in the wikitext from the description to the categories.

However, a lot of the visual tools to assist in adding and editing categories tools assume that this will be at the bottom of the page -- so simply add new categories at the end of the page.

In this case, however, that would lead to the description page having category information in two different places -- some above the big metadata comment, some below it.  It's not good for the information to be going to be split in this way.

So -- if the metadata is useful (which it may well be), a better place to put it might be in a separate sub-page.  On a separate page, it would also be safe from automated edits -- for example my edits with AWB here.


My apologies that I got into a bit of a state about all this last night (and my relief that it's not the blocker I thought it would be). These issues may seem trivial, but in my view they are important (to me, a difference between acceptable and unacceptable output), so IMO they are things that *need* to be tidied up before any big launch.

All best,

   James.



On 06/03/2014 19:28, James Heald wrote:
David,

Thank you so much for this.

For me the most pressing issues are:
  *  allowing punctuation in the filenames
  *  the layout of the Artwork, so that the fields occur in their usual
standard order & missing fields are included
  *  moving the 'gwtoolset-title-identifier' and
'gwtoolset-url-to-the-media-file' fields out of the artwork template, eg
into a template of their own

I hope these are all fairly small changes, almost cosmetic, that can be
sorted out quickly.

But they would make a huge difference -- I've already had a sharp note
on my Commons talk page that the images have filled up the automatic
"Artwork template with incorrect parameter" maintenance category,
https://commons.wikimedia.org/wiki/Category:Pages_using_Artwork_template_with_incorrect_parameter

making the category useless for identifying other user's genuine
mistakes because it's full of the 430 images that I uploaded.

As for the filenames, and the template fields, I really really want to
get these sorted.  Really the only sensible way for me to fix them is to
re-run the entire upload, once the tool is patched.

But until I've done the re-upload that's blocking me from doing a lot of
essential plumbing -- eg properly categorising the images; wikilinking
their subjects, adding them into articles (including swapping the new
images in instead of a lot of existing inferior versions -- which are
exactly the things that are needed to make the upload look good, if the
upload is going to be cited in the official release at the end of the
month.   But at the moment I'm blocked, because there is no point in
doing any of those things, if I know that I'm going to do a batch
re-upload that will wipe all of those things out.

So I hope these key things aren't big fixes, but if it could be possible
to get a patched version of the tool up and running I'd be incredibly
grateful.

All best,

    James.




_______________________________________________
Glamtools mailing list
Glamtools@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/glamtools