Tool testing report
BaGLAMa and GLAMorous ported to Wikimedia Labs
Over recent months, many of the tools regularly used in GLAM-Wiki contexts have become increasingly inaccessible due to the Toolserver being overloaded (symptom, background). The recommended solution for maintainers of such tools is to port their code to a new environment on dedicated servers, Wikimedia Labs (details).
There is no comprehensive listing of the tools to be ported and the progress of the porting, but Magnus Manske has such an overview for his tools, and he invites feedback from users on what to port and how to prioritize. Following popular demand, he has already ported some of the most widely used applications. As a frequent user of several of his tools, I am following the process and will attempt to share my observations on the matter in this and upcoming issues of the Tool testing reports, starting with Quick Intersection, BaGLAMa and GLAMorous.
Category intersections provide a listing of wiki pages that are in two specific categories. Magnus' Quick Intersection is one of several tools available for that, and one of the first to be ported. I like it because it is simple and quick. For instance, it allows to find all files uploaded by the Open Access Media Importer that are
missing categories or lacking descriptions, so that these gaps can then be filled systematically.
BaGLAMa and GLAMorous
These two provide an overview on how files from a given category on Wikimedia Commons are being used across Wikimedia projects, along with aggregated pageview stats (cf. Tool testing reports for January and February 2012). Both use the Wikipedia pageview stats tool to gather pageview data, which sits on yet another server and has been only intermittently accessible in recent days. Having its database exposed via Wikimedia Labs would help the GLAM tools run more smoothly (details). Hourly view stats are available here and have been used to generate weekly stats from 2008 onwards (example).
Applying BaGLAMa and GLAMorous to the Commons category Open access (publishing) yields the following: BaGLAMa for May 2013 (WebCite copy) indicates 39 million visits across 207 Wikimedia projects to a total of 45k article pages that use images or media from this category, whereas GLAMorous in early July (WebCite copy) indicates a total of 19218 files in that category, 5111 of which have been used a total of 93839 times across 297 Wikimedia projects.
In theory, both tools should yield the same number of Wikimedia projects (perhaps give or take a few to account for the different times of scanning), but in practice, BaGLAMa scans fewer Wikimedia projects (for example, Wikivoyage is not yet included). I am also not entirely sure about what those 94k image usages from GLAMorous mean in the context of BaGLAMa's 45k article pages, since a mean of two images from that category per page that has any image from that category seems rather high for wikis other than outreach.
Finally, it seems that some datasets have gone missing in BaGLAMa. For instance, the stats for December 2011 till June 2012 are much lower (currently) than they were reported in February 2013.
I would just like to note that the new gallery stuff was never intended to fix issues with linking to description pages of images. That's a broader issue with the TimedMediaHandler, and applies to all image displays, gallery or otherwise. It is however an issue we do want to address.
I would also like to note that if one wants to link to the file descirption pages as part of the default caption for image galleries, that is an option -
showfilename="true" (And has been since mediawiki 1.17). For example:
Bawolff (talk) 00:14, 22 July 2013 (UTC)
- Saw that comment only now - I had not been aware of showfilename. Thanks! --Daniel Mietchen (talk) 22:46, 23 August 2013 (UTC)