Commons talk:British Library/Mechanical Curator collection/map tag status

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


Please add your name to the top of the list below, to give an idea of who is helping, where you are editing from, and which sections you are working on.

Flickr keyboard shortcut[edit]

@Daniel Mietchen, Malchugan, Ccrisp03, Hsarrazin, and Mauricen53: @Jamesmcmahon0, ClemRutter, Andrew Davidson, LastGrape, and Thryduulf: @Andrew Gray, WereSpielChequers, Aa77zz, PamD, and Nealmcb: @Rodw, Technolalia, and Thryduulf: Thanks to everybody for a fantastic effort so far.

This is just to make sure that everybody knows there is a keyboard shortcut on the Flickr page to start tagging an image -- keying "t" automatically puts you into the "add tag" box. Probably you all knew about that all already, but I only discovered it yesterday, and found it made a huge difference to how quickly I could tag some of the big books with lots of images.

I've added a line to the instructions summary on the status page about it, but I thought I should ping everybody with this post to make sure we all knew. Cheers, Jheald (talk) 12:56, 7 November 2014 (UTC)Reply[reply]

I hadn't discovered that, thanks! Thryduulf (talk) 14:48, 7 November 2014 (UTC)Reply[reply]
Thanks - I only discovered it a few days ago too, after googling for keyboard shortcuts. Agree it's quicker this way. -- Daniel Mietchen (talk) 21:03, 7 November 2014 (UTC)Reply[reply]
I am new to Flickr (had to create an account), but I was happy enough to find it quite quickly - it sure is very useful for books with many images - thanks :) --Hsarrazin (talk) 21:05, 7 November 2014 (UTC)Reply[reply]

Big Galleries[edit]

Just a quick question: Does any one experience problems with big Flick galleries? When working with galleries containing more than 900 images the flickr starts eating memory like crazy and the page renders very slowly. If this is a common problem is it possible to split the galleries in parts so they could be traversed easily?--Malchugan (talk) 09:26, 8 December 2014 (UTC)Reply[reply]
For me, these problems start much earlier than at 900 files. Don't see an easy way to split the gallery, other than by using dedicated tags. -- Daniel Mietchen (talk) 09:32, 8 December 2014 (UTC)Reply[reply]
This 4000-odd file example (Az Osztrák-Magyar ...) had Vol: tags, so (after I had already reviewed them once myself) I added each of the volume sets for others to review. Mark Hurd (talk) 04:19, 16 December 2014 (UTC)Reply[reply]
My old PC regularly crashes when I try larger sets, sometimes as small as 200 images. For the largest ones I've learned to dip in do a few maps in the first few screens and move on, I can only mark things as complete if they have perhaps a 100 non map images. This one with 5,830 images is particularly time consuming because a high proportion are maps.  Untagged maps?   WereSpielChequers (talk) 20:07, 17 December 2014 (UTC)Reply[reply]
Your description of "to dip in" made me realise an option for these large galleries with lots of maps is to not scroll down unless there are no maps in the first screen. Just middle click on each of the maps, tag them as such and close them. Then refresh the page.
From what I've seen previously, with larger galleries with fewer maps, the recently tagged maps may still appear in the gallery for a while, but normally sorted later in the display. As such you should be able to attack these map-full galleries without actually retrieving many pages of images, until lots of non-maps remain. (I've been attacking your example this way for a few minutes already.) Mark Hurd (talk) 00:26, 18 December 2014 (UTC)Reply[reply]
The only problem with this approach is you get no idea of how far you've got to go! (BTW the non-maps displayed seem to "page" forward as well when you refresh this way.)
I've now spent about 40 minutes on just this gallery and so probably added about 40 maps, but I should be working, so I'll stop now. Mark Hurd (talk) 00:53, 18 December 2014 (UTC)Reply[reply]
Yup that's what I mean by dip in. I've done that with many of the atlases and it should be possible to clear them eventually. But this one is particularly big, though with many nice maps. When I started dipping into it there were relatively few non maps, now it is about fifty fifty and I fear there will come a point where this method no longer suffices and it needs someone whose PC can handle rather more images than mine. WereSpielChequers (talk) 05:57, 18 December 2014 (UTC)Reply[reply]


All now done!
The volume trick helps a lot with this one:
vol 01
vol 02
vol 03
vol 04
vol 05
vol 06
vol 07
vol 08
vol 09
vol 10
vol 11
vol 12
vol 13
vol 14
vol 15
vol 16
vol 17
vol 18
vol 19
My computer is a bit newer than Jonathan's (WereSpielChequers'), but at least for me these are now small enough that I can attack them in my Standard Operating Procedure -- viz scroll to the bottom, then work my way back up, right-clicking images into new tabs (right-click then "t" on Firefox), then tagging the images once I have about ten open. It becomes like shooting fish in a barrel, and should take only about 5 minutes for 40 tags. Jheald (talk) 08:48, 18 December 2014 (UTC)Reply[reply]
Though that said, when I reload the page, it seems to find new maps that weren't offered the first time. So I guess one needs to be careful not to be complacent with books this size. Jheald (talk) 09:04, 18 December 2014 (UTC)Reply[reply]
Thanks for doing this, I think that makes the project achievable. I've now dipped into all of these volumes at least once, so tht should make things easier for others. As for reloading, I did find a couple of maps in one of the three completed volumes. WereSpielChequers (talk) 06:07, 20 December 2014 (UTC)Reply[reply]
Thanks Jonathan. It seems like there was a lot less to do in vol 4 that I looked at just now than in the couple of vols I looked at last night, so on the basis of one data point, I think your sweep has made a real difference, definitely. Jheald (talk) 08:34, 20 December 2014 (UTC)Reply[reply]
It's actually a really nice book. I see we've got an article on the author, discussing how it was his life's work, was published volume-by-volume over almost 20 years, and in the end the Paris Geographical Society gave him a gold medal for it. It's great to think that by this time next year, people will be able to browse it eg at the Internet Archive, and every time they come to a map there should be a little link in the corner to let them see it overlaid on a map from today. But I am quite glad there's only one copy scanned for us to do.  :-) Jheald (talk) 14:41, 20 December 2014 (UTC)Reply[reply]

Image search getting stuck at the bottom of the page[edit]

Flickr's image search definitely seems to be having more problems than it used to. I find increasingly it's 'sticking' at the bottom of the page, with its rotating pink and blue circles going round, but no more images ever getting uploaded. I don't know if Flickr has recently changed its code for presenting lots of images, but I'm getting this a lot now -- often with books with only 150 or so images -- much more than I think I ever used to.

One thing that it seems can sometimes help is to add a few more search conditions to the url -- eg to add +-rotatec or something quite random like +-fish. It seems that this can sometimes make the search get all the way to the bottom of the page and complete. I don't know if it's anything about adding the extra conditions specifically, or whether it just gives another roll of the dice that sometimes succeeds. But I have managed to sign off a few pages this way. Jheald (talk) 13:15, 21 December 2014 (UTC)Reply[reply]

Yes, seeing that too. Mark Hurd (talk) 14:06, 21 December 2014 (UTC)Reply[reply]
Can we subdivide more into subvolumes? WereSpielChequers (talk) 22:44, 21 December 2014 (UTC)Reply[reply]
@WereSpielChequers: Absolutely. Just feel free to add relevant search links for the different volumes in the right place on the relevant index page, and then start crossing them out again :-) The URLs are very straightforward -- it's just an extra +vol03 or whatever to include the relevant volume tag. Jheald (talk) 00:15, 22 December 2014 (UTC)Reply[reply]
Again I should be working, but I've provided the volume links for the first section of Synoptic index, USA. Unfortunately the 2182 image book, King's Hand-book of the United States, is all one volume :-( Does anyone know if there is a definitive way of knowing whether there are non-contiguous volume numbers? Mark Hurd (talk) 01:00, 22 December 2014 (UTC)Reply[reply]
@Markhurd: With King's Handbook, I think we have a 100% upload to Commons courtesy of User:Metilsteiner (IIRC it took him over two weeks to upload & annotate); so one option might be to go through the Commons category, open any map images on Commons, then click through from each Commons file page to the relevant file page on Flickr.
With the spinning balls, according to a commenter in a Flickr help thread, if you click on a thumbnail (not right-click), then go "back" to the thumbnails page, that apparently sorts it. I have tried it once, and it worked. Jheald (talk) 01:11, 22 December 2014 (UTC)Reply[reply]
I found this was only somewhat better than just refreshing, but it's definitely failing for the "only 179 image" Eclectic Physical Geography – each attempt is actually displaying fewer images :-( Mark Hurd (talk) 00:49, 27 December 2014 (UTC)Reply[reply]


we've just lost a couple of hundred WereSpielChequers (talk) 22:39, 21 December 2014 (UTC)Reply[reply]

@WereSpielChequers: It happens. Flickr runs a federated system of servers. Tags that get added get registered first by the nearest server, then propagated to the rest of the network. Somewhere along the line, before the tagathon even started, one of the servers missed out on about 180 map tags. So every once in a while, the total number arbitrarily drops by about 180, then usually comes back the very next time the search is run, ten minutes later. The larger number is the correct number, as checked by polling each book individually. Jheald (talk) 00:27, 22 December 2014 (UTC)Reply[reply]

Well done all[edit]

We could now go back through all the maps and add the other tags they want, where not already done... :-) Mark Hurd (talk) 15:50, 28 December 2014 (UTC)Reply[reply]

As we got further in I found myself focussing more and more on the maps and plans, so there are lots of images that could do with having rotatec or rotatecc added. More prosaically, there are lots of ships and historical figures that would be interesting to migrate to commons, though hopefully in a less labour intensive way. WereSpielChequers (talk) 16:27, 28 December 2014 (UTC)Reply[reply]
Mince pies all round.--ClemRutter (talk) 19:59, 28 December 2014 (UTC)Reply[reply]

Quasimondo's map set[edit]

@Mark Hurd, WereSpielChequers, ClemRutter, and Mauricen53: Everything taken out by December 28th -- congratulations to everybody, and as proposed by Clem, mince pies all round!

If anybody is having withdrawal symptoms for maps, there is one more thing we could usefully do before the georeferencing, and that would be to go over the maps tagged by Quasimondo, and make sure they are all turned the right way round (ie add the tags rotatec and rotatecc where necessary, as well as tags like split and world.

The good news is that we won't have to worry about issues with Flickr's infinite scroll -- each of the pages below (other than the last) contains exactly 72 images. The less good news is that there are quite a lot of pages, and tagged images won't immediately disappear -- though they will in time get turned the right way round, usually within a couple of days. It's another lot of work, but if we want to take it on and properly finish the job, I think it would make a lot of difference for the georeferencers. Jheald (talk) 23:25, 30 December 2014 (UTC)Reply[reply]

All done 13/01/2015 — Preceding unsigned comment added by Mauricen53 (talk • contribs) 20:13, 13 January 2015 (UTC)Reply[reply]
@Mauricen53: That's fantastic. Thank you so much!
But, as the saying goes, the reward for good work is more good work. I met Kimberly and Ben on Wednesday, and the plan is to aim to try to get the final list of files ready to Klokan in Switzerland by next Friday (23rd), so that georeferencing can be got started as soon as possible.

New albums[edit]

Ben has just uploaded the results of the tagathon into seven new albums, which show up at the end of the BL's page of albums: Map Check #1, Map Check #2, Map Check #3, Map Check #4, Map Check #5, Map Check #6, Map Check #7
Map Check #1 contains the maps tagged before October 31st; #2 to #7 contain the maps that were tagged during the tagathon.
So the aim is to work through as much of that as we can in the next week, rotating any maps that need turning the right way up by adding rotatec, rotatecc and rotate180 tags, tagging any definitely not a map with not_map, and any world maps using world, conical and/or hemispheres. Maps that already have the latter tags have been put to the back of albums #1 and #7.
I'll put up a cross-off list here as before. (From past experience, sometimes it's quickest to do a whole row of pages at a time, rather than break the rhythm by stopping to cross off each one individually).
It should be possible to get through a surprisingly large amount (eg I did the 45 pages of the BL's original album last night); and with luck for the tagathon ones, more should already have been correctly tagged and/or rotated already. So let's see what we can do! Jheald (talk) 15:22, 16 January 2015 (UTC)Reply[reply]
I should also mention that Kimberly has an assistant working on maps that have been tagged 'split', cutting out maps using Photoshop where there's more than one map to a page; and also the machine vision group in Oxford have an imaged-based content search, and code for recognising similar/identical images, which I am hopeful we might get some results from in time to identify some duplicates. Jheald (talk) 16:41, 16 January 2015 (UTC)Reply[reply]
I've now moved the previous set to
The cross-off pages for the new albums are at:
-- Jheald (talk) 15:57, 16 January 2015 (UTC)Reply[reply]

Map Check #1 album[edit]

(Note: these links are based on 72 images per page. However, Flickr is not always consistent, and users using some browsers might not see this.)

pages 1 to 30[edit]

001 002 003 004 005

006 007 008 009 010

011 012 013 014 015

016 017 018 019 020

021 022 023 024 025

026 027 028 029 030

pages 31 to 60[edit]

031 032 033 034 035

036 037 038 039 040

041 042 043 044 045

046 047 048 049 050

051 052 053 054 055

056 057 058 059 060

pages 61 to 82[edit]

061 062 063 064 065

066 067 068 069 070

071 072 073 074 075

076 077 078 079 080

081 082

Georeferencing now underway[edit]

MUSE awards 2015, about to be presented

@Daniel Mietchen, Malchugan, Ccrisp03, Hsarrazin, and Mauricen53: @Jamesmcmahon0, ClemRutter, Andrew Davidson, LastGrape, and Thryduulf: @Andrew Gray, WereSpielChequers, Aa77zz, PamD, and Nealmcb: @Rodw, Technolalia, Thryduulf, Ra uk, and Markhurd: @Tbhgeo and Paul the Archivist:

Hi, and thank you all so much for all your efforts last year, as a result of which we now have 50,000 maps and plans in all found in the collection. You may be interested to know that the next stage of the project is now well underway -- the georeferencing of those maps, allowing the past to be compared with the present, and making possible some pretty precise searching, which is the pay-off for all the previous tagging. For more information, see the new georeferencing campaign page, and the up-to-date georeferencing status page. Enjoy!

Oh, and you might also know we won an award! Yay! -- Silver in the category 'Open' at the American Alliance of Museums' MUSE Awards 2015 for Media & Technology. The award was for the tagging + the georeferencing + the categorisation and upload, though, so we'd better make sure we earn the rest of it! All best, Jheald (talk) 18:35, 30 April 2015 (UTC)Reply[reply]