User:Jim.henderson/Picture pipeline< User:Jim.henderson
A few people have asked how I get my pictures into Commons and Wikipedia, so I'm describing the process here. I use one camera but multiple Windows computers, the majority of them obsolete and about half of them belonging to someone else, so my method emphasizes portability and generic software. Other photographers may prefer to use Mac or only one computer with the custom software that came with their camera. Each picture goes through the same steps in rigid sequence: Shoot, select, retouch, categorize, describe, upload, link, and revisit.
I carry my little camera everywhere in daytime in case of opportunities, but more pictures are from purposeful photographic expeditions. Until summer 2012 I usually selected a number of targets near each other and in a line, folded my bicycle, took a train or boat or other transport to the beginning of the shoot line, unfolded and pedaled along. Some shots are so casual, I don't stop pedaling but just shoot from the saddle. Other targets call for dismounting, walking around, climbing a wall, whatever. Often I notice an unexpected sight off the path, go chasing that, and don't necessarily get back to my intended targets or route. A typical expedition gets a hundred or more pictures including usually some of the intended targets but rarely all.
Early in 2012 I learned to use the smartphone's Google Maps feature for hunting. In those cases I might not do much planning, but join a biker gang going quickly to a suburb, quit after lunch and go solo, following the W marks on the phone map, sometimes checking for an existing photo, sometimes just snapping and hoping. Often on the way between those targets I spot a marker that some local or state history org has installed, or simply see an interesting building, and snap it. Geotags and additional photos of street signs and the like help me to identify the subject later.
Photograph your hometown is partly mine, with further suggestions.
My usual upload connection is a slow DSL, so unlike some photographers I don't upload all my shots of an interesting target or send the pictures as large, uncompressed files. Instead I insert the camera's SD card into the computer or a USB card reader. Microsoft Windows Explorer automatically pops up, shows me the contents, and I go down to the folder containing the pictures.
Sometimes I don't know what I photographed, only its location and that it was an impressive sight. Google Maps and Web searches may identify it for me, or Wikipedia's list of en:National Register of Historic Places for the county. If reasonable efforts fail, I give up and go on to the next photo.
I select the best picture of each target and copy it into the root directory of the card with a new, descriptive, name. Selection takes considerable time when a target has many pictures suffering various defects requiring an opinion as to which is least bad. Many pix show things for which there is no Wikipedia article or the article already has as good a picture and thus mine would be clutter rather than a contribution. My main goal to illustrate Wikipedia articles; not offer pix that will only seen by people who use Commons. Thus I seldom upload a picture without soon using it an article.
I must straighten pix that are tilted because I was dodging potholes, crop people who stood too close to my architectural target, adjust color, etc. In the 2000s I used Adobe Photoshop 7 on someone else's computer. Losing that access, I learned Google Picasa, outgrew it in late 2012, and started using GIMP which can do far more sophisticated things than I am ever likely to study.
So, I finally bought a Vista computer and a book. Cool. From looking things up on the Web that the WinVista book mentions, I found and downloaded the free Microsoft Pro Photo Tools. Now I can mark and right-click a bunch of photos, select that program, and see coordinates.
This is just a bunch of numbers, but then click on the Map tab and up comes a Microsoft Live View map of the world with a thumbtack next to New York since this is one of my Newark photos. Pan and zoom into downtown Newark, and see a cluster of thumbtacks representing the various pictures. Zoom a little closer, and the map shows that the camera three weeks ago thought it was still at St John's Roman Catholic Church even though I had pedalled half a mile north to New Jersey Performing Arts Center.
Now comes the cool part. Drag the thumbtack to the correct spot on Center Street, and drop it. The numbers change, and when I quit the Tools program, it asks whether to save the info and yes, the new location goes into the metadata section of the image file or files. Then any program that understands geotags can find it correctly. Not just cool. Way cool.
Correcting my coordinates without a computer that can read them was a huge hassle before. I had to upload the pictures to Wikimedia Commons, wait hours or days for the Wiki geotagging bot to convert from EXIF metadata to Wiki file description template, click a few clicks on an XP computer modern enough to read online maps, roughly measure the error in feet (sometimes nautical miles) and convert mentally to degrees and minutes. Then edit the change into the Wikimedia template. Wait a day or two for the change to percolate through the system before seeing the result, then correct the correction. The delays caused endless errors on my part, especially when I didn't know I was seeing locations that corresponded to last week's numbers and making the corrections on this week's numbers, causing overshoot that would only be discovered next week. Or next couple days, anyway.
No more. Way cool. I had intended to use Google Earth and Picasa for my coordinate corrections, but Picasa is an image cataloging and editing program, and only incidentally a geographic metadata program, and its coordination with the online mapping program is clumsy. Microsoft's online maps are mostly not as good as Google's (lacking overt street address numbers at extreme scale, for example) but the operation is far slicker. Now my pix will arrive at Wikipedia with the correct coordinates, not ones that are five or ten minutes out of date and a mile off. Wow; what a concept. At last I'm getting what I paid extra for a fancy geocoding camera and the newer Windows Vista to get.
Now comes the hard part. Online, I can study where the picture will fit. That generally breaks down to two questions: Commons category, and Wikipedia article.
To find a good category, I point my browser (usually Opera, sometimes MSIE or Firefox) at Commons. Having thought of what the picture is about, I do a few keyword searches, and pick a search result, for example an existing picture that seems close. It is seldom in exactly the right category, but it's probably close in the tree of categories. I look at its subcats and if those seem unpromising, at its parent cats which sometimes have other subcats that are more relevant. Working my way up and down the trees, I select the best single uncrowded category; rarely two cats at this stage. For example, for a bronze statue of an American Civil War hero in a park in Gutenberg, New Jersey, my first category will be a strictly geographical one. If there's no category for the park, I look for the town, and if there's none for the town, then its county. Were I dealing with some US State where not all the counties have a category, I'd look in the State category.
Having found my first Commons category, I open a second browser tab in Wikipedia and find the best single article for this picture and leave it open. Very rarely, two articles. Returning to the Commons tab, I drag the mouse over the name of the category, hit Contr-C for Copy, and click the Upload File button of the Commons menu.
The Upload Form comes on the screen, and I click to select the filename in my SD card. Down near the bottom of the Upload Form, I use Contr-V to Paste the category name into the appropriate spot. I select the license, which for me is always Public Domain. I fill in the date, and describe the picture in the description field. I go to the Wikipedia tab, drag the mouse over the name of the article, Contr-C, return to the Commons tab, put in the the square [[:en: brackets and ":en" mark indicating a link from Commons to Wikipedia, Contr-V to Paste, and the final ]] brackets to finish the link.
When the description, category and licence are satisfactory, I finally click the Upload button at the bottom of the form and sit back for half a minute while my slow DSL connection sends my picture to Commons. When the picture arrives, it creates a Wiki media file with a title same, similar to an article. I drag the mouse through the name, Contr-C, and go to the Wikipedia article tab. There I click Edit, insert the square brackets at the top of the article or other appropriate location, Contr-V to insert the picture file name, then Thumbnail indicator, caption, create an edit comment so other editors will understand my actions, and click Show Preview. If all is well, click Save Page.
If something is wrong with the file description, category, article insertion or other aspect of the operation, it must be fixed immediately. If not, the job is complete until revisit time. After a few days, it trickles down to the bottom of my watchlist, so I look again and at this time usually add a few more words of description and a supplementary category or two. Since my pictures are almost all about places, not people or things, the initial cat is usually strictly geographical, but the later category may say it's a church or river in New York, or part of the Brooklyn waterfront or whatever kind of target it is.