User:Jmabel/Final draft of talk for WikiConference North America

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

An extended, written version of my talk at Wikiconference North America, Toronto, Friday 10 November 2023; co-presented along with a talk by Dominic Byrd-McDevitt (User:Dominic) who has been bringing in the DPLA images.

I've been going through tens of thousands of images from Seattle Public Library (SPL) uploaded by Dominic as part of his DPLA work. I occasionally find something useful for illustrating a Wikipedia article, or a little more often for a Wikidata item—there were a surprising number of photos of prominent musicians, actors, and the like who toured through the region, and a particularly surprising number of the were people we didn't have pictures of previously, or didn't have good pictures; but I'm going to take a more Commons-focused approach today. By categorizing these photos and fleshing out descriptions, I've been able to really improve Commons coverage of late 19th and early 20th century Seattle in particular and, more generally, the Pacific Northwest and Alaska.

Above all: I want to thank SPL for consistently making high-resolution scans available online for these public-domain items in their collection. So many institutions provide only low-resolution scans, useless for maps, marginally useful for text documents, and often rather disappointing for photos. SPL has consistently provided genuinely useful images. The only ones that fall at all short are scans of things like an old newspaper whose own half-assed low-res half-toning did not make for very useful images in the first place. Even then, SPL has generally done as well as possible.

That said: it actually can be pretty remarkable how many errors there are in library metadata. I've dealt with enough institutions to say that SPL is probably better than par for the course: in general, images there were captioned by someone who had a clue and looked at the individual photos. For some other institutions that I won't name, I've encountered collections where most photos had little identification beyond a broad geographic area and a date range of a decade or two. I've also encountered collections where it is clear that the metadata was written by a student intern whose knowledge of the subject matter was roughly what you'd expect from a random undergrad. Probably about 95% of the images in the SPL collection as uploaded by the bot were described accurately and in relevant terms. I'm guessing that I identified the majority of significant errors; I'm also absolutely certain that with so many errors, some have gone right by me, especially for things outside of the Seattle area.

Besides outright errors, there are also some cases of "missing the lede": for example, although there is no outright error in a description, a famous person or specific identifiable event might not be identified. Also, there's stuff like attributing two copies of the same photo to different photographers, or completely mis-transcribing what someone wrote on a photo (in one case "trestle" became "fressee"), and so forth. Also, SPL is a bit sloppy with attributions on photos old enough to be public domain. In particular, Asahel Curtis, who some of you may have heard of—he took a bunch of pictures in Alaska and Yukon during the Klondike Gold Rush, and his brother Edward Curtis did some of the most famous photographs of Native Americans ever taken—had a tendency to buy up rights to older Seattle photographs and put his own mark on them, and SPL tended to take that at face value, so they would have him credited for a photo taken before he was old enough to walk and talk.

Whenever I encounter errors, I add the correction to the Commons file page; I also make a note in a file intended for SPL, and roughly once a month I send them my corrections, which seem usually to find their way to the site. In my edits on Commons, I try to separate my corrections from the original content, usually just by using a horizontal line. If a title is blatantly wrong, I change it, recording the old title as "original description". About the only thing where I just correct it. put my remarks in the edit summary, and leave it at that is fixing the many wrong attributions to Asahel Curtis. In my email to SPL, I use their URLs rather than ours, so they can see the issues within their own site (though every so often, if things are really complicated, I give them an overview and refer them to the Commons page where I've explained in detail, especially if I needed to use ImageNotes to explain the issue).

These are some examples of the sort of issues I've found.

For the first two of images, SPL said, "Exact location along Sixtymile River is unknown." I immediately know I'd seen photos of the place before, and quickly nailed it down as the same location as the third picture shown here, a much lower-resolution scan from the University of Washington Libraries collection, who identified it as "the terminus of portage tramway company at Miles Canyon Landing, Yukon Territory" and added "Probably the Canyon and White Horse Rapids Tramway Company used for portaging around the Whitehorse Rapids on the Yukon River."

SPL had titled this "J.L. Luckenbach in Elliott Bay, ca. 1910" with the description "Seattle. Ships. Luckenbach (steamship) in Elliott Bay. Name changed to Evergreen State when purchased by the States Marine Lines in 1959." One of my other ongoing projects—one that might be evidence of Obsessive Compulsive Disorder—is en:List of structures on Elliott Bay, so the name immediately rang a bell. For many years (though before I lived in Seattle) there was a Luckenbach Steamship Company with a dock on Elliott Bay on Seattle's Central Waterfront. The statement about the name being changed to Evergreen State immediately struck me as unlikely: States Marine Lines purchasing a 50+ year old ship? A quick online search turned up a ship list for Luckenbach Steamship Co., New York, 1850-1974, where I quickly established that there have been several ships named J.L. Luckenbach and that the J.L. Luckenbach that became the "Evergreen State" was built in 1943, so it certainly cannot have been photographed circa 1910. Mission almost accomplished. Looking at the rest of the ship list, I was able to add, "Assuming the circa 1910 date is roughly accurate, this must be the J.L. Luckenbach built in 1886, although the date could imaginably be about a decade off and be the one built in 1919 and named J.L. Luckenbach in 1922."

Here's another good outcome of looking at far too many pictures of the Seattle waterfront: I was able to look at this circa 1905 postcard and say that it was based on a very old photo, before the 1889 Great Seattle Fire, because the building at right is a boathouse that burned in that fire.

Seattle circa 1876, two decades after its founding and 13 years before the Great Seattle Fire wiped away anything you see in the foreground here that made it through that long. This was originally identified as "View looking north on Commercial Street from Cherry Street" followed by a pretty accurate list of some of the buildings here, including Yesler's Mill. Immediately, that didn't make sense. If you know Downtown Seattle at all, you know that one of the big dividing lines is Yesler Way, the old Skid Road or Mill Street leading to Yesler's Mill. From about 1890 into the 1960s, it was the line between the relatively respectable Downtown to the north and the "Skid Road" vice district to the south. Cherry Street is north of Yesler Way, so you can't be looking north from Cherry Street and see the mill. But there are the original University of Washington buildings in the distance at left, before the university moved out of the heart of the city to its current campus, so we are looking north. And, oh, "Commercial Street": that was present-day First Avenue South, south of Yesler. North of that it was Front Street, and looking at the picture I pretty much immediately identified the messed-up pre-Fire T-junction at Mill Street (they smoothed out the connection from Commercial Street to Front Street after the Fire) and about ten minutes of research told me just where was the Cosmopolitan Hotel whose sign is just right of center, and I was able to say this is First Avenue South, looking just east of north, roughly from South Washington Street, with South King Street about halfway between the point of view and Yesler Way. Whee!

SPL had described this postcard as being from Columbia Street, two blocks south of what was almost certainly the actual point of view. At first I thought the image was fictional; in fact I thought that for quite a while, because given that it was supposed to be from Columbia Street, the buildings diddn't make sense (the Colman Building at far right is actually north of Columbia). Then I saw, from a different source, the photo by Arthur Churchill Warner below that, and realized it had to be real. That led to a bunch of sleuthing about what was tall enough, built at the right date, and stood at the correct angle to First Avenue South stretching into the distance. (It all gets tricky because, as you can see here, two street grids meet at an angle.) It turned out the be the American Savings Bank and Empire Building (demolished within the time I've lived in Seattle) at Second and Madison, two blocks north of Columbia Street.

Here's one where the library failed to trust good information. They titled this, accurately enough, "3rd Ave. and Jefferson St. after the fire, July 1889" but wrote that "[the c]aption on photograph 'Down Yesler Way' may provide the incorrect citation. Other photographs, including one taken by McManus, identify the location at 3rd Ave. S. and S. Jefferson St. [etc.]," but Third and Jefferson is only about 50 feet from Yesler Way. Someone didn't look at a map: this is in a spot where two grids meet. If you know Seattle, this is across Third from the King County Courthouse, and the Prefontaine Fountain sits in the triangle of Third, Jefferson, and Yesler.

Or there's stuff like this, a case of "missing the lede". This was just captioned, "14th Ave. homes on Capitol Hill, ca. 1910". That's one of the most famous private houses (well, originally private houses) in Seattle, at left, the Moore Mansion. Moore pretty much created Capitol Hill as a fashionable district; he also was responsible for the Moore Theatre, still going strong after 115 years, where I've seen everything from Laibach to Laurie Anderson to the U.S. premieres of The Stuntman and Soldier of Orange.

And sometimes it is as basic as noticing out of thousands of postcard scans that the two postcards here are the same image, and that it is a colorization of this photo by Arthur Churchill Warner that we already had a copy of from University of Washington Special Collections. Commons categories made it really easy to spot something like that. Also: check out the ImageNotes I added to that black-and-white version. I've spotted things like this literally hundreds of times, which besides everything else lets us identify the photographer responsible for the image on a postcard.

Here's another case where categories came in handy. The photo at left came from SPL and was dated "ca. 1884". It's John Leary, a prominent Seattle businessman and mayor; several homes in his family are still standing and serve as things like clubhouses, the seat of the Episcopal Diocese, things like that. In 1884, Leary would only have been 47. Further, the picture has to be later than the one on the right, which was published in an 1891 book that used mostly then-recent pictures. I'd guess the SPL picture is toward the end of Leary's life (he died in 1905) and for once might legitimately be by Asahel Curtis (who started working professionally around 1897) rather than Curtis reprinting someone else's work. Anyway, I re-dated it as circa 1900.

And one last example of where some combination of Commons categories and staring at way too many photos of the Seattle waterfront paid off. SPL dated the first photo here as circa 1905 the Seattle Mail & Herald published a different crop of it on December 20, 1902 (bottom photo in the middle). And the photo at right, another from Seattle Public Library described by them as circa 1907, may not be the same photo, but they were certainly shot within minutes of the other because the same ships are in the same place. (Also, that last one shows the Lincoln Hotel, so all three also can't possibly be before 1900 when that was built. It burned rather dramatically in 1920, by the way.)

Retrieved from "https://commons.wikimedia.org/w/index.php?title=User:Jmabel/Final_draft_of_talk_for_WikiConference_North_America&oldid=821822313"