Commons:British Library/Mechanical Curator collection

The Mechanical Curator collection is a set of approximately one million image files uploaded by the British Library to Flickr Commons in December 2013. (BL blog post; news reports). It contains illustrations and other imagery from out-of-copyright books and other sources drawn from the areas of travel, geography, ethnography, geology and literature.

A curated set of highlight images from the collection has been put together by the BL team, along with other sets, including 'Portraits', 'Maps', 'Christmas' and 'Science Fiction', found by the community; in addition to which the eponymous Mechanical Curator posts a new random set of 100 images each Friday.

Images from the collection that have been uploaded to Commons can be found in

Category:Images from the British Library Mechanical Curator collection.

This page is to help co-ordinate the sorting and reuse of this rich collection.

Away from Commons, the British Library has set up this wiki for the collection, and tag #bl1million is active on Twitter.

Georeferencing campaign. (Started April 2015). Currently underway is a campaign to locate 50,000 maps and plans found in the collection images, by placing them accurately on present-day maps. See latest progress, and instructions. This in turn should lead to the ability to upload them to Commons with automatic accurate categorisation.

Finding images

Exploring by book subject

A project to create a title-level thematic and geographical index to the books that were scanned to create the collection (or, more specifically, to those books which have a significant number of useful images in them) is underway at

Commons:British Library/Mechanical Curator collection/Synoptic index.

Currently (as of January 2014), the focus is so far mostly geographical, with little development as yet of thematic classification. The titles in the index so far account for about 40% of the images in the whole collection. A further 35% of images can be found through a more loosely organised set of "to do" pages, with another 20% in books classed as fiction, plays, verse or works.

A full list of the titles of the books that were scanned for the collection is also available, at

Commons:British Library/Mechanical Curator collection/Full list of books

and succeeding pages, which may also help in browsing what is available.

To help build up the synoptic index, new titles to add to it can be found either by browsing the "to do" pages, or the full list of books, or simply by going to the British Library's Flickr account, or one of the sets derived from it, and exploring. If you find an interesting image, you can go from the image to the tag page for the book by clicking the "Click here to see all the illustrations in the book" link in the last paragraph of the description on Flickr. If the book fits a category in the index, just add it to the list.

Flickr search and tags

Flickr's standard search can be specialised to search the BL photostream by adding w=12403504@N02& to the URL. For a Flickr tag search, the most convenient form is probably

  https://www.flickr.com/search/?w=12403504@N02&q=africa+map&m=tags

where multiple separate tags can be specified (in a URL of this form) by separating them with a '+'. Tags can also be excluded, by preceding them with a '-',

  https://www.flickr.com/search/?w=12403504@N02&q=africa+-map&m=tags

A list of tags added by users up to December 2014 gives a snapshot of what sort of things people may have added tags for.

The tagging can already produce some quite nice collections (e.g.: bridge), in addition to the selected sets posted by the Library; and it can also be a good way to keep track of particular specific groups of images of interest. However the extent of tagging is still quite limited (with the exception of the map tag, which was comprehensively added in a systematic tagging drive from October to December 2014), so in most cases most of the images to which the tag could (or should) be applied will not have been so tagged, and therefore will not be returned by a tag search.

One thing which should be comprehensively tagged is the author of a book, which can be especially helpful in searching for illustrations of works of writers of fiction, e.g. the works of Sir Arthur Conan Doyle.

Beyond what has been tagged, the BL photostream can also be searched using Flickr's more general search function, for example https://www.flickr.com/search/?w=12403504@N02&q=westminster%20abbey, without the extra &m=tags. This appears to typically find images with the search words in the book title -- which sometimes will reflect the actual content of the image, but sometimes not.

The number of images with a particular tag combination can be counted using the API interface to Flickr search,

https://www.flickr.com/services/api/explore/flickr.photos.search

The user_id parameter should be set to 12403504@N02 to specify the British Library's images, and tags should be separated with commas, e.g. world, map, -hemispheres, with tag_mode set to all if all the tags should be present. Further options can be found discussed on the search API documentation page. The API search is limited to returning details of at most 4000 images -- however the count for the total number of images matching the criteria appears to be reliable, even if it is greater than this.

Uploading to Commons

Images can also be downloaded manually to your own computer and then uploaded in the usual way. If proceeding manually, use the download icon at the bottom right of the Flickr image to acquire the highest resolution possible.

Alternatively multiple images from a single book can be automatically transferred from Flickr to Commons using the Flickr2commons tool. If using this tool, the British Library's Flickr user id is 12403504@N02, and you will need to supply the sysnum........ Flickr tag to identify the book.

For admins and some established users, it is also possible to upload a single image from the collection using the "Share image from Flickr" button on the first page of the usual Upload Wizard. If using either of the bots, they should be able to find the original size automatically, from just the main Flickr page URL for the image. Single images should be added to Category:Images with no book category from the British Library Mechanical Curator collection.

For details on how to then generate a standard information page for the image, see the #Image descriptions section below.

If an image needs to be rotated, this can signalled by adding the {{Rotate}} template, using {{rotate|90}} if an image needs to be rotated clockwise, and {{rotate|-90}} for an image to be rotated anti-clockwise. Tagged images should be automatically rotated by Rotatebot within about an hour.

Cropping of images can be achieved interactively in place on Commons, using croptool.

Attempts to clean up or 'restore' images -- e.g. sharpening, colour correction, contrast enhancement, background lightening, dirt and stain removal, etc. -- are welcome. Leading Commons restorationists Durova and Adam Cuerden have contributed some notes on basic image restoration in this tutorial on Wikiversity; alternatively, here is a video tutorial with some suggestions. Some of the images may also lend themselves to conversion to vector format -- for example File:Dragon-woodcut2.svg has been contributed by Salvor.

It is not a bad idea to make sure a copy of the original image pre-modification is uploaded here first, so that Commons viewers can directly compare the enhanced version with the source material. If the derived version still presents essentially the same image, for example the cropping of File:Common seal of the Mayor, Aldermen + Burgesses of the Borough of Birmingham.jpg, the new version can just be uploaded over the original, and both will be visible in the list of revisions. For more major changes, such as File:Statue of George I and Hogarth's House, 1790 (detail).jpg extracted from File:Statue of George I and Hogarth's House, 1790.jpg, or File:Royal Birmingham Society of Artists circa 1829 - interior.jpg derived from File:Royal Birmingham Society of Artists circa 1829.jpg, make a new upload and leave the original, linking them each to the other via the |other_versions= parameter of the artwork template.

Image descriptions

Please replace the default description template with the pre-formed upload template {{Commons:British Library/Mechanical Curator collection/script}}, which will produce a formatted image description using {{Artwork}}, populated with the standard metadata provided. It works best if you subst: it!

You may copy:

{{Subst:Commons:British Library/Mechanical Curator collection/script
| PAGE = 
| VOLUME =
| TITLE = 
| AUTHOR = 
| YEAR = 
| SHELFMARK = British Library HMNTS 
| SYSTEM = 
| FLICKR = 
}}

A title and/or description for the image should then be added to the artwork template by hand.

The {{Mechanical Curator image}} that the script adds will automatically register the image in Category:Images from the British Library Mechanical Curator collection.

Manual method

Alternatively, tag images with {{Mechanical Curator image}} directly. To ensure a link back to the catalogue, add |system=xxxxxxxxx, where xxxxxxxx is the nine-digit identifier number given for the book, and |flickr=http://... to give a link back to the image on flickr. The further parameter |hasLicense=yes can be set provided the page includes an appropriate license-tag, for example {{PD-old-70-1923}}

You may copy:

{{Mechanical Curator image|system= xxxxxxxxx |flickr= http://... |hasLicense=yes}}

This template implicitly adds the image to Category:Images from the British Library Mechanical Curator collection.

Derived images should also be added to Category:Images derived from images from the British Library Mechanical Curator collection.

Categories for books

When uploading these images from Flickr, find or create a Commons category for the book, like those in Category:Books with images from the British Library Mechanical Curator collection, and include them in it.

In the synoptic index the template {{HasBookCat}} should be used at the start of an entry for a book title. This will show a link to the category on a green background, for example

 (Book cat (100%)) 

which is the result of the code

{{HasBookCat|Das Deutsche Volk in seinen Mundarten, Sitten, Gebräuchen, Festen und Trachten (1847) von Eduard Duller|complete}}

The name of the category (without the prefix "Category:") should be given as the first parameter. If all of the useful images from the book have been added to Commons (i.e. usually everything apart from decorative fillers and ornate capital letters), please add the word 'complete' as a second parameter.

On the category page for the book itself, please include template {{BL1million bookcat}} to add a link back to the image set for the book on Flickr, using the sysnum parameter from the Flickr URL.

Tagging on Flickr

Wild Turkey - Sportsman and Naturalist in Canada.jpeg

When you have uploaded an image from the collection, please add a "machine tag" to the corresponding Flickr page, with the Commons file name. Do this in the format "WikimediaCommons:filename=Wild Turkey - Sportsman and Naturalist in Canada.jpeg" (include the quotes to escape any spaces or commas; omit the "File:" prefix). See for example, the machine tags on http://www.flickr.com/photos/britishlibrary/11072237684/in/set-72157638544764936 which has been uploaded as File:Wild Turkey - Sportsman and Naturalist in Canada.jpeg

You may copy:

"WikimediaCommons:filename=XXXXXX"

Wikipedia

If the image depicts the subject of a specific Wikipedia article, for example en:St Paul's Church, Birmingham, tag thus: "Wikipedia:en=St Paul's Church, Birmingham" (adjust two-letter language code as appropriate; again, use quotes to escape spaces and punctuation).

You may copy:

"Wikipedia:en=XXXXXX"

Wikidata

Alternatively, or additionally you may wish to tag with a Wikidata ID. For https://www.wikidata.org/wiki/Q915614, tag with "Wikidata:item=Q915614"

You may copy:

Wikidata:item=Qnnnnn

Standard tags

The British Library has put up on its tagging guidelines page a list of some of the standard tags it hopes will be used to tag the collection - but don't forget to add the information, as categories, here, too!.

Some useful suggested Flickr tags include map, architecture, decoration, letterA etc., music, portrait and fashion.

Other useful tags might include titlepage, cover, advertisement, signature, handwriting, heraldry, shield, coat of arms, landscape, nature, montage, stereopair or colour.

Images which need rotating should be tagged rotatec to be rotated clockwise, and rotatecc to be rotated counter- or anti-clockwise. (Note -- the pages that list tagged images may be slow to update).

See also a list of all the tags added to the collection in the first 12 months.

Issues

Some of the imagery is presented without adequate context. Research is required to place some of the images. The British Library online viewer/downloadable PDFs may be able to help, as may Google Books or the Internet Archive.

Some images may depict places that have changed names or represent it in such a way as to confirm the prejudices of the time and represent now outdated systemic biases. Some scientific imagery may be outdated or misleading. Caution and critical thought should be exercised before unthinkingly using the imagery in Wikipedia.

See also

• Commons:Internet Archive/Book Images collection -- a similar collection of 2.4 million images released by the Internet Archive in August 2014