zooniverse-glacier / notesFromNature

https://www.notesfromnature.org/
Apache License 2.0
13 stars 11 forks source link

User reports record jump and problem with images #335

Closed joanball closed 9 years ago

joanball commented 10 years ago

Another user reported that they had sudden increases in number of records, as well as a problem with multiple images loading in sequence.

http://talk.notesfromnature.org/#/boards/BNN0000002/discussions/DNN0000270?page=1&comment_id=534ecc9565a77d3c5e0006b7

"I should have credit for about 160 classifications. I now have over 250. Why? When clicking "OK" on the last stage of classifying, or "Next....", I am getting up to 10, sometimes more, files flashed past me almost faster than I can see, let alone click on. This will seriously muck with the finished database, since these images are being counted as completed classifications (I/we haven't even had a chance to see them!)..."

robgur commented 10 years ago

No question this is a problem.

On Wed, Apr 16, 2014 at 9:41 PM, joanball notifications@github.com wrote:

Another user reported that they had sudden increases in number of records, as well as a problem with multiple images loading in sequence.

http://talk.notesfromnature.org/#/boards/BNN0000002/discussions/DNN0000270?page=1&comment_id=534ecc9565a77d3c5e0006b7

"I should have credit for about 160 classifications. I now have over 250. Why? When clicking "OK" on the last stage of classifying, or "Next....", I am getting up to 10, sometimes more, files flashed past me almost faster than I can see, let alone click on. This will seriously muck with the finished database, since these images are being counted as completed classifications (I/we haven't even had a chance to see them!)..."

— Reply to this email directly or view it on GitHubhttps://github.com/zooniverse/notesFromNature/issues/335 .

joanball commented 10 years ago

Here is more information from the user today:

by Noharrr OK - I logged about 30 files today but . . . when I left the site yesterday, I was still under 900 (only approx 160 of which were actually processed by me). Yet when I logged in today, the counter was at 945. When I finished, my CalBug counter was at 975. Yet when I reconnected via the ZOONIVERSE home page, my counter was a 1075.

That's a lot of images (ie over 800) that show as transcribed, when I never even saw the files.

The images I transcribed today were stable; I was not aware of images flashing past, and the counter on my transcribing page seemed to be working normally -- except for the additional 100 tossed in somewhere between CalBug classifying and re-entering via the ZOO home page.

joanball commented 10 years ago

Looking more closely again at another thread, I think this image loading issue has happened for others as well...

http://talk.notesfromnature.org/#/boards/BNN0000002/discussions/DNN00001ze?page=2&comment_id=5345e0c8699fda0582000663

by HelenBennett57 I'm getting this problem too. I've transcribed maybe five herbarium sheets today and got credit for about forty.

I haven't skipped the in-between ones, it just happens

It looks as if some other records appear and disappear after I have submitted one, and before the next one settles down on the page

Using Chrome 33 on Windows 7 Home Premium

Noticed for the first time this morning (14th March); could be because I'd gone straight into the saved URL http://www.notesfromnature.org/#/archives/herbarium/transcribe rather than through the "front page" http://www.notesfromnature.org/

Posted a month ago by majortim Just clicked OK "record done" and then a new specimen was loaded, appeared for a few seconds, then another 1 was loaded and so on. Over 50 bugs were skipped.

Using FF 27.0.1

JoyceGross commented 10 years ago

Pasting from an email I sent to some folks earlier today:

I would like to request that skips stop counting towards a user's total transcriptions (we've been talking about this among the CalBug team). Skips are related to some but not all of the strange user counts so why not clean up all related counting messes at once. Reasons not to count skips:

1) The UI right now adds 'skip' counts to a user's total transcriptions ONLY after the user leaves the transcription screen (at least that is what is happening for me). So if I start with 5 transcriptions done, do 2 more transcriptions, and skip 3 images, at first my total is 7 (which is correct). But as soon as I leave the transcription page (ie, click on the power button icon in the upper right corner of the screen to go back to the CalBug splash page), then click "Start Transcribing" again, my total number transcribed jumps to 10. Clearly it is adding the skipped records, but it didn't add them in at the time that I skipped them so the numbers appear to jump around. THIS IS REALLY CONFUSING TO USERS AND NEEDS TO BE FIXED.

In addition to the above problem, there is the issue of huge jumps in transcriptions that do not correlate exactly with numbers of actual skipped records. I haven't personally seen this yet but it is clearly happening.

2) 11% of our transcriptions to date are skips -- makes no sense to be giving people badges for all those.

Joyce

ps. I'm using Firefox 28.0 on a PC (Win 7). I don't believe the counting issues I've seen are browser-related though.

chrissnyder commented 10 years ago

First problem is easy to fix.

Only problem with making skips not count towards a users transcription count is that they will eventually get the same record again. Maybe we should change the verbiage to something like "records seen" rather than "transcriptions done"?

joanball commented 10 years ago

According to what people have said in the forums, they are getting the same images again anyway, both skipped and non-skipped (although maybe not as frequently as they might?). I don't think that a "records seen" field would work well with the badges or any competitions that we might do in the future. I can also say that people do not seem to like getting credit for skipped fields.

On Tue, Apr 22, 2014 at 6:53 AM, Chris Snyder notifications@github.comwrote:

First problem is easy to fix.

Only problem with making skips not count towards a users transcription count is that they will eventually get the same record again. Maybe we should change the verbiage to something like "records seen" rather than "transcriptions done"?

Reply to this email directly or view it on GitHubhttps://github.com/zooniverse/notesFromNature/issues/335#issuecomment-41041779 .

Joan E. Ball-Damerow PhD Candidate | UC Berkeley Environmental Science, Policy & Management http://joanballdamerow.wordpress.com/

robgur commented 10 years ago

I agree with Joan.

On Tue, Apr 22, 2014 at 8:09 AM, joanball notifications@github.com wrote:

According to what people have said in the forums, they are getting the same images again anyway, both skipped and non-skipped (although maybe not as frequently as they might?). I don't think that a "records seen" field would work well with the badges or any competitions that we might do in the future. I can also say that people do not seem to like getting credit for skipped fields.

On Tue, Apr 22, 2014 at 6:53 AM, Chris Snyder <notifications@github.com

wrote:

First problem is easy to fix.

Only problem with making skips not count towards a users transcription count is that they will eventually get the same record again. Maybe we should change the verbiage to something like "records seen" rather than "transcriptions done"?

Reply to this email directly or view it on GitHub< https://github.com/zooniverse/notesFromNature/issues/335#issuecomment-41041779

.

Joan E. Ball-Damerow PhD Candidate | UC Berkeley Environmental Science, Policy & Management http://joanballdamerow.wordpress.com/

— Reply to this email directly or view it on GitHubhttps://github.com/zooniverse/notesFromNature/issues/335#issuecomment-41043730 .

JoyceGross commented 10 years ago

The data confirms users are getting repeat images, even though skipped images are supposed to prevent this. See #334 which has links to examples.

joanball commented 10 years ago

A couple more recent comments on the record count issue...

http://talk.notesfromnature.org/#/boards/BNN0000002/discussions/DNN00001ze?page=3&comment_id=53723ef38a5a6d60660003bd

by HelenBennett57 Just hit this problem again, 9.40 AM UK time, herbarium, browser/OS as before. Haven't been doing any funny cache-messing things this time (unless NFN doesn't play nice with something on iNaturalist, where I was looking at the first screen of the UI), but I'm about to try clearing out the cache as before and reopening the browser.

Edit - clearing the cache and restarting the computer stopped the problem happening, but this isn't a long-term solution as it takes me a while to get all my work stuff set up, so there's a big hit when I shut/reopen it all. Next time I'll try just clearing the cache (also quite a hassle) and restarting the browser. Not a good fix but maybe it gives a better idea of why counts are going wierd.

Posted 3 days ago by LanceSnider I'm seeing the opposite issue. I haven't been hugely active on this site, but all of the sudden it says I've only ever done 10. I have no idea what the real count is, but I can tell you it's more than 10. :)