Improve Flow of Identify Mode

#1

Right now the way the identify page works is not very inuitive. This is how it seems like it should work:

  1. I see 30 observations to identify, and go through them, identifying as I can
  2. I click on “page 2” when I get to the bottom, or click “skip to next page” when that popup is shown…
  3. I see the next 30 observations to identify.

But there are some issues with this. This page is not showing me the next 30 most recent observations to ID. Rather, it is recalculating the number of observations that meet the filter criteria, and showing #31-60 of those

As a result:

  • If new observations are added while I was identifying, I may now see observations I passed over on page 1 showing up again on page 2.
  • Depending on how many observations I reviewed on page 1, I will be skipping that number of observations when I go to page 2. If I hit “review all observations”, going to page 2 will actually show the 30 observations that were originally on page 3

It is not very obvious to the user that this is what is happening. There are some workarounds (refreshing the page after clicking “mark all as reviewed”, or working backwards from the last page), but these are both unintuitive and have some other annoying problems.

I realise that improving this probably isn’t simple on the technical side.

0 Likes

Add "mark all as reviewed" button to bottom of ID page or page numbers to top of ID page
#2

Related topic from GG: https://groups.google.com/forum/#!topic/inaturalist/QwlGHcIBnJ8

0 Likes

#3

I have a workaround for the ‘jumped pages’ issue.

When I’ve filtered for what I want on the Identify page, the results of the filters, after I’ve reviewed the first page, will mean that I either have the same number of obs filtered for, or fewer (the original number less the number I’ve reviewed).

If the former, then I click on Skip to next page. If the latter, then I click the cross (close the box), and refresh the page I’m on. This means that the next 30 obs appear (‘page 2’ which is now page 1) and not page 3 (which is now page 2).

{In case an explanation of my 2nd paragraph is needed, here goes:
If my filter includes, say, “Reviewed: any” the number of obs will remain the same after I have reviewed obs on the first page.
If my filter includes, say, “Reviewed: no” the number of obs will change after I have reviewed obs on the first page.}

0 Likes

#4

I’m not sure there’s a good solution here. IMO, if you are actually trying to identify all records within a set of search parameters, the most logical workflow is to go through every observation on the page and identify and/or mark as reviewed, then click “View more” to load more unreviewed records and never use the pagination links. Personally I find that annoying because I almost never want to identify all records within a set of search parameters, hence the pagination links, which allow me to skip ahead, even if pagination is sort of an insane way to navigate a set of records that’s constantly being added to.

To me, a more sensible approach to navigating dynamically generated records like this is to just have a “More” button that loads records that occur before the last record on the current page (in whatever sort order is chosen), but that comes with its own annoyances, b/c I actually like skipping 5 pages in or whatever just to find something to ID.

So, I agree the current state of affairs is confusing, but all the remedies I can think of also make Identify less useful in certain contexts.

Pet peeve: the word “intuitive” drives me nuts. Intuition is almost completely subjective and based on your own background and expectations, so saying an interface is not “intuitive” is no more helpful than saying “I don’t like it.” The initial post here was well-written in that it details exactly why pagination in Identify is confusing, but saying something is “intuitive” or “not intuitive” does not communicate anything to me other than a desire to use the word “intuitive.”

4 Likes

#5

The main problem with the page numbers as they exist right now is that while you’re viewing a static list, records often get added and removed from the list, and the page links send you to pages re-computed at the moment you click the link even though the page numbers of the links were computed at the moment the page was loaded. I think you can get the advantages of both approaches with keyset paging, which makes paging less jumpy when the list you’re paging through can change while you view it.

The “View More” button would be a lot more useful if it refreshed the current page, instead of sending you back to page 1. I’ve accidentally clicked that button at least a dozen times trying to go to the next page while 100 or more pages into the list, and then had to navigate back manually because the back-button doesn’t work on that page. :frowning_face:

1 Like

#6

Thanks for the response.

The “view more” button is actually sort of what i want - somehow I’m not sure I’ve ever hit it before!

I should probably get in the habit of hitting “R” on observations I don’t want to identify. It’s a little unnecessary though given the “mark all as reviewed button”. The issue:

  • The “mark all reviewed” button is only at the top of the screen, and the “view more” button is only at the bottom of the screen, meaning I have to do a bunch of extra scrolling.
  • If I hit “mark all reviewed”, and then hit F5 to refresh, this is almost exactly what I want. The problem is that it seems to often take a while to actually mark everything as reviewed. I’m not sure if this is related to computer speed or internet speed, but if I don’t wait ~5-8 seconds between hitting “mark all reviewed” and refreshing, some observations won’t be marked.
  • “View more” seems to work similarly. If I hit the button too fast, I will see repeat observations. The difference being that the repeats will be marked as reviewed, despite showing up on the new page

I guess what I would like is something along the lines of a button that combines “mark all reviewed” and “view more”, and that it would wait until all observations have actually been reviewed before displaying the new set. Ideally this button would be in the box that pops up with “Skip to next page”.

Not suggesting that this is a huge problem or anything - but I find every process I’ve tried in identify mode slightly but continually annoying, so thought I should post and see what other people thought. Doesn’t look like it’s an especially common complaint judging by the number of votes.

Intuitive: I guess when I use the word unintuitive I mean something very specific (this solution is not one of the first few things I would think to try), or more broadly that this solution is not one the first few things most people would think to try. I could be factually wrong of course, but that’s the view I’m trying to express.

2 Likes

#7

Having to scroll up to hit “mark all as reviewed” is not, to me, a big concern. In some ways it is better, because my next step is going to be hitting refresh which is at the top of the browser page anyway, so I will have the mouse up there. Otherwise I am moving my mouse further around the screen, which is equally (albeit minorly so) annoying! I am more and more getting into using “page up”, click the mouse (which is still over the “mark all” from last time), and then F5 for refresh. Everything else is also off the keyboard, once I got used to the keyboard shortcuts!

It just occured to me… you aren’t going to be identifying EVERY observation in Needs ID in the one session anyway, so if it skips a couple each page turn, no worries, you can catch them on the next session! And perhaps they would have been IDd by someone else in the meantime.

1 Like

#8

you aren’t going to be identifying EVERY observation in Needs ID in the one session anyway

There are a bunch of taxa-place combinations for which I have reviewed or identified every observation that wasn’t already research grade. Periodically I go in and I do go through all the new observations (a relatively small set) in one session.

0 Likes

#9

Maybe a little less jumpy, but you still need to deal with the fact that the act of reviewing and identifying itself changes the result set, unless you’ve set reviewed=any, i.e. you review everything on a single page, now page 2 of the initial result set is basically page 1 of the new result set. Keyset paging would make perfect sense on a page like https://www.inaturalist.org/observations, where there are similar problems, but not so much on Identify.

0 Likes

#10

What happens when you get to the end of that result set? Does it just give you a page 4 (for example) that has nothing in it, or will it work back to the last page in the set? If the latter, then surely you get to them all regardless?

0 Likes

#11

I’m not sure I understand what you’re getting at. (On a page like what?) In both cases (marking everything as reviewed or only marking some fraction of observations reviewed) the first observation on the next page should be the observation immediately after the last observation on the current page. The way things work now, if I ID 15 out of 30 observations on a page, then hit the next page arrow, the first observation on the next page is 15 observations after the last observation on the current page (assuming I’m working with a narrow enough filter that no new observations have been added while I was making new IDs). If I review all 30 observations on a page, then hitting the next page arrow effectively skips a whole page.

0 Likes

#12

Yes, it is very easy to end up in a situation where you click on page 2, but there’s nothing there because all the observations are on page 1.

0 Likes

#13

Sorry, meant “Keyset paging would make perfect sense on a page like https://www.inaturalist.org/observations”.

Yes, but once you get to page 2 in this scenario, when happens when you click back to page 1? Do you see your 15 reviewed and 15 unreviewed observations, or do you see the 15 unreviewed and 15 more unreviewed that used to be on page 2? Or worse when you go from page 10 to page 3 after having reviewed just a few on each page in between? What you’re proposing would work just fine for a
“More” or a “Next” button as I described above, but not so great for pagination. It would work fine if we ignored whether or not you’d reviewed stuff, but that’s sort of critical to the experience.

0 Likes

#15

I experimented with this, and it does give you a page beyond the end of the resultset, and it contains no results. For example, from p5 when p6 is the last page, I mark as reviewed enough obs so that p6 obs should now fit on p5 if I refreshed, but instead I click next (or p6) and get the empty page. Looking at the navigation, there is p1-p5, but looking at the url I am on p6. So I either click p1 to start over again, or p5 to go to the last page.

Either way it is not losing anything! There are no observations dropped unknowingly from the review pool. There is also no requirement that they be identified/reviewed in order. I can imagine a situation where someone might make a comment like “this is the same as in the previous observation”, which it could possibly be seperated from as far as Needs ID page goes, but then if someone identifies one and not the other, then that would have the same outcome. You could find it regardless by viewing that obs and clicking previous in the observation view.

If it is a problem that some drop back to previous pages, or move onto subsequent pages when large numbers of observations get uploaded (the reverse side of this issue!), then a custom filter can fix it, but I don’t think it needs to be “fixed” for everyone when it is working perfectly fine

1 Like

#16

for what it’s worth, i usually just use the ‘view more unreviewed’ button and when i get to the end, it seems that i’ve seen everything i wanted to see. I don’t fully understand what these new proposed changes would give us over that but i am not opposed to them so long as the ‘unreviewed’ option remains.

1 Like

#17

Some people seem to think that the way pages flow works OK, while a number of people are incredibly frustrated with the way things work now. I fall into the latter category. I think that the fundamental disconnect between the way I use the site and the way it was designed is that I would never click “Mark all as reviewed”. When I am identifying observations from a group of organisms, some of them I know and will identify. Others I don’t know and skip over. But I may decide to study up on some of the ones I don’t know and then work on identifying them too. If I ever marked an observation as “reviewed” just because I didn’t know what it was, then I’m losing an opportunity to learn something new or ever go back and identify that later. So the only reason I mark something “reviewed” is when the photo is so bad that I consider it unidentifiable. For someone taking this approach, it is extremely difficult to go through a group and know you have looked at every observation.
-Mark

2 Likes

#18

ah i get what you are saying. For me there are so many thousands more plant observations in the areas i look at, that i end up just skipping anything i don’t know pretty quickly (unless it’s something really cool) so i figure it’s ok to mark stuff as reviewed… but it makes sense you wouldn’t want to for those reasons. I wonder if there could be a way created to unreview speciic taxa only. That could be useful but maybe is off the scope of this.

Anyway, even though it works for me changes that make it work for more people are good too as long as it doesn’t beak it for those who it works for now :)

3 Likes

#19

I think there are filters for that… :wink:

0 Likes

#20

No, I mean If I mark sedges as reviewed when I see them but later want to un mark them and review them again because I learned more sedges. Or something. I’m probably not explaining well.

1 Like

#21

Doesn’t it work if you filter in Identify for “sedges” and “reviewed” and then hit “mark all as reviewed” not once but twice? (The first hit leaves them at reviewed and the second hit unreviews them.)

1 Like