A Few Answers to John Ascher's Questions

As listed at https://www.inaturalist.org/people/johnascher @johnascher

"1) Why is no effort made to encourage proper citation of localities in the proper sequence country: state: county: location?"
What is John referring to here? Possibly he is referring to the Establishment Means tab on taxon pages, where a county may be listed out of context, e.g. just "Colusa" when it should say "Colusa County, CA, USA." I agree that's frustrating/annoying. Is there somewhere else that this occurs? We should submit a feature request to fix this and other instances.

"2) Why is it acceptable for contributors to submit uncropped images?"
Cropping an image removes the context of the observation. I greatly enjoy being able to see the bigger picture of an ecosystem and not just the individual organism. Helpfully, we can zoom in to view the creature up close without need for cropping. It can definitely be a bit of a game of "Where's Waldo" sometimes, but that's all part of the fun.

I would also urge people to take more context photos to show the organism in relation to its associate species as well as more "ugly" photos of entire life cycles rather than just glamour shots of flowers in peak bloom, for example. Of course taxon page default photos are ideally close up and/or highlight characteristic features but there's no need to frass middling or low quality photos. iNaturalist serves as a nature journal for most people. To support removal of low-quality photos would be like going into someone's house and ripping pages out of their personal journal. One thing that is helpful, John, is to "favorite" observations with high quality photos. That way they appear higher in the list of photos when viewing all photos from the taxon page.

Not everyone can afford high quality camera equipment with the ability to take macro shots of moving insects. Not everyone has photo-editing software or skills. Be mindful of the variety of people on iNaturalist: we come from different socio-economic and educational backgrounds with varying levels of experience and skill. Encourage people with helpful tips and tricks. Don't berate them.

"3) Why are records not validated by an expert considered Research Grade?"
Because semantics. Neither the wording nor system are perfect. But having a photo or audio file, >2/3 majority vote on an identification, date, GPS location, etc. seems like a decent start to me to begin whittling down to what one might consider a "quality" observation. What are John's recommendations? Who is an expert?

"4) Why no taxon pages for some of the most obvious, widely-known, and well-accepted monophyletic groups generally recognized by the public such as bees (=clade Anthophila)?"
Technically it looks like Anthophila is rankless. iNaturalist doesn't currently have a system to position rankless groups together -- you have to select an established rank. I cheated and created Anthophila, but it must be categorized as an "epifamily" in order to place it correctly within the taxonomy. Calling it "epifamily," to my limited knowledge of bees (see #7 below) and after browsing this paper quickly, may be wrong, but the only way to do it in the current system on iNat.

Is there another group John is referencing in this question?

"5) Why is it acceptable for contributors to submit images with no organism or no detectable organism as an identification request? This practice sure does waste our time."
What is John's solution? It's certainly not an encouraged activity given that it violates the basic tenants of what an "observation" is, and there are plenty of data quality tools to deal with these swiftly. Feel free to use some of the canned responses to these types of observations in order to encourage new users to use iNat correctly: https://www.inaturalist.org/pages/responses

"6) Why must all images be retained even if they add no value to the site as determined by a qualified expert?"
See #2 and #3.

"7) Why can't non-experts take the time to identify the most obvious species such as the Western Honey Bee?"
There it is, the dreaded "it's obvious." Speaking for myself, as primarily a plant person, I know there are several species of Apis, that there are tens of thousands of other species of bees, and that many hymenopterans and other insects are lookalikes/mimics. Get it wrong once and if insects aren't your current focus, you're probably likely to be much more tentative about your IDs in the future.

"8) Why is it acceptable to cite subgenera with no indication of the genus in question? This unprofessional practice is particularly unfortunate if the goal is reliable communication with non-experts."
A documented annoyance. Please contribute thoughts here: https://groups.google.com/d/msg/inaturalist/oZKEzDlIWd4/zhFpcs8ZAQAJ

"9) I understand that not all contributors wish to share their precise location, but can't they at very least note the country in question? If that must be secret why post anything to the internet?"
A documented annoyance. Please contribute thoughts here: https://groups.google.com/d/msg/inaturalist/s-xSk35ljUI/W2AZpu4fAQAJ

Yes, iNaturalist is a public website, but again, it's also a nature journal. Which easy-to-use software geared toward both beginners and experts does John recommend for people to download to their computers that has the same features available on iNaturalist?

由使用者 bouteloua bouteloua2018年01月25日 23:20 所貼文


Is John asking these questions in the Google forum? Hard to get a discussion going otherwise.

發佈由 pfau_tarleton 超過 6 年 前

Most grievances have come in via the flagging system, some of which you can see here. No use of the Google Group that I have seen.

發佈由 bouteloua 超過 6 年 前

I have encouraged John to post his concerns to the Google Group, but I have not discussed it with him in a while. Perhaps he did not find the Google Group? I would suggest tagging him on this post so your answers aren't wasted, but the Google Group would be a MUCH better place as I would anticipate it turning into full blown discussion. I personally don't use the Google Group because I'm quite happy in my world of oblivion. Could you extract his questions and post them yourself (or would that open a can of Annelids)?

發佈由 kimberlietx 超過 6 年 前

I thought about posting it to the Google Group (https://groups.google.com/forum/#!forum/inaturalist), but decided that I preferred to have it as a journal post. Partly because I know that John has not used the Google Group, partly because of the tone, partly because it's addressing multiple different topics, and partly because while it's clearly not a private conversation, it's somewhat intended for a smaller audience than would be blasted to the GG.

John has already been tagged in the post, but I'll do it in a comment too because I know journal tags have been buggy lately.

發佈由 bouteloua 超過 6 年 前

Buggy. Ha! I see what you did there! ;)

發佈由 kimberlietx 超過 6 年 前

I think the first question is referring to the "location" on observations (e.g. on this one: https://www.inaturalist.org/observations/9593106 it's Nipissing District, ON, Canada). This is auto-generated from Google Maps, and anyways seems pretty redundant to me since observations are already attached to coordinates and an accuracy value.

發佈由 reuvenm 超過 6 年 前

@reuvenm I thought that might be the case too. It's definitely not clear when this information is manually entered (e.g. https://www.inaturalist.org/observations/9395352) and when it's auto-generated from Google Maps, so I can see how it might be confusing to users who aren't experienced adding observations themselves.

發佈由 bouteloua 超過 6 年 前

Yeah, I definitely agree with you on the 'it's obvious' point. As a novice, it's the times that I've went 'it's obvious', I've found I was most likely off, and at worst, overlooking valid instances of just how varied wildlife can be.

發佈由 tanyuu 超過 6 年 前

"more "ugly" photos of entire life cycles rather than just glamour shots of flowers in peak bloom"

I am alllll about this.

發佈由 vvoelker 超過 6 年 前

Ditto on the "ugly" photos. The problem is so many don't get ID'ed because they aren't "obvious" in the other stages. I know a few people that have really gone the extra to take photos of different plant parts and different stages. When I see those pics, I try to update the taxon photos to include them. Very important for proper IDs!

發佈由 kimberlietx 超過 6 年 前

"If these questions are answered my enthusiasm for the site will increase."

發佈由 bouteloua 超過 6 年 前

I'm sorry, but I have a hard time keeping up with inaturalist itself and do not wish to engage in a further discussion on another site. I suppose that can be another question, Why must forum discussions take place via a Google group rather than through the site itself?

It seems like people are apologizing for shortcomings of the site rather than trying to address them.

Regarding localities, the issue is obvious. People place a dot on the map but very often do not note the country: state: county: and locality name unabbreviated and with correct capitalization. Thus, we must zoom into the map to determine where a sighting is found. All citizen science records should have at least the most basic locality information cited clearly as text with proper capitalization of name and in the correct order. Relying on auto generated locality names is inadequate and correct citation of such names is not redundant! It is a matter of names not being cited properly, not a matter of my failure to understand how localities are auto generated. I know what is happening and do not like the result.

The problem of who is an expert who can validate records seems to have been solved at Bugguide so why not inaturalist?

As for "obvious" identifications, there is only one Apis species in the New World. It is not clear to me why very active contributors from California, Texas, etc., cannot identify the one species in their hemisphere? I would understand if they were cautious and refrained from proposing species IDs on more difficult taxa, but since many such contributors routinely propose (mistaken) species IDs for Xylocopa (difficult!), why so much caution when it comes to identifying the single New World Apis species?

Regarding cropping, we cannot identify an insect that we cannot even find. No expensive equipment is needed to crop an image.

As for observations with no images, I am not objecting to their inclusion in the site in general. I am, instead, objecting to their being submitted as ID requests, e.g., for Community ID. This is inappropriate and wastes everyone's time. No need for ID verification when there is nothing to verify, right?

發佈由 johnascher 超過 6 年 前

--Yes, I believe there is a discussion happening about possibly moving the community forum from Google Group into iNaturalist. My understanding is that the website was created out of a Master's thesis project. At that time, it probably made sense to use an existing forum platform like Google Groups rather than spend a lot of effort building it directly into the website.

--I'm confused, so either I'm an idiot or the issue is not obvious. Why would I spend my time adding county, state, and country to all of my thousands of observations when that information can be extremely easily discerned from the coordinates / auto-generated Standard and Community-Curated Places ? Is this just an American and European bias? The auto-generated county/region/state/country information is quite good for areas that I have used iNaturalist.

--I'm not sure what your last paragraph is referring to. Observations with no images (or sound) are automatically classified as "Not Verifiable" and are never included in "Identify" by default.

發佈由 bouteloua 超過 6 年 前

I found "It seems like people are apologizing for shortcomings of the site rather than trying to address them" to be almost laughable considering you posted a list of grievances on your user profile rather than raising them on the established community forum and providing recommended solutions.

Many of us spend hours volunteering each day improving the website. You included. Thanks! I look forward to a productive discussion.

發佈由 bouteloua 超過 6 年 前

I think it's important to say that iNaturalist is not BugGuide rebooted. It's a completely different concept that has different pros and cons. There is no need to remake BG if it works so well. Unfortunately, a lot of people complain about the unwelcoming comments on their submissions at BG, and I for one would like to prevent iNat from going in that direction. I want to see more people of all backgrounds working together, not just hobbyists submitting photos and waiting for an expert to tell them what it is. We learn differently at iNat. We (hobbyists) learn by trying and taking positive corrections when things are wrong. If you don't consider something Research Grade, you are under no obligation to bless it as such even if your ID is outweighed by a group of others. If you are actually using the data beyond identifications, you have complete control over what specifications you want to include/exclude.

Apparently people have tried to give you answers so you can see what iNaturalist is all about, but you don't like them. Perhaps iNat is not for you. (Although I sincerely hope that is NOT the case.) We need you as an expert just as much as we need the hobbyist who is photographing their first wildflower to find out what it is. I suggested this before, but if the Google Group is not the forum you want to utilize for answers, you can always email help@inaturalist.com. That account is answered by Ken-ichi Ueda who is the co-founder of iNat. I can't think of a better way for you to get your voice heard than that.

發佈由 kimberlietx 超過 6 年 前

John, the iNat development team is extremely active and responsive to the iNat user community (in the Google forum) and improvements/add ons are being made continuously. Here's one example: https://groups.google.com/forum/#!topic/inaturalist/oRGMkskv8yY

And from that improvement, observations can now be exported into a spreadsheet (with the option to have a column for county, state, country). I'm a scientist that uses locality data from museum specimens on a regular basis. I've been continually frustrated with the locality data on specimens tags (e.g. where, exactly, is 4 mi W St. Louis, Missouri, USA?). When I map specimens, I'm constantly having to guesstimate where to plot them based on the given locality data. I spent an entire week dealing with some specimens not long ago because the collector put W instead of E of a certain town. I eventually tracked down the student's thesis adviser who found her notes and figured out the mistake. So, I'm not really seeing the advantages to having everyone enter this information manually. I think the problem is probably compounded when dealing with observations from other countries (e.g. how could I read locality data in Japanese?). With the research I do, the EXACT locality is critical--a vague locality description just won't do in many cases.

Experts. I'm not so sure this has, or ever will be, solved (anywhere). Correct me if I'm wrong, but "Experts" in BG are folks that can view high-resolution images and make IDs (I put it in quotes because I'm referring to the particular kind of expert as defined in BG, not just the general term). Editors can do these things also, plus some other things: https://bugguide.net/node/view/280353. I don't see it formally stated anywhere on BG what criteria have to be met to become an Editor or Expert--somehow, someone identifies these folks. I've had about 10% of my submissions to BG initially IDd by an Editor/Expert, only to be later IDd as something else by a different Editor/Expert (e.g. https://bugguide.net/node/view/1386092). I have no way of knowing which one is correct and don't know what the process is of them duking it out or coming to a consensus. The last person making the ID gets the final say, and one of them, I suppose, has to back down. I suppose there's some kind of informal hierarchy that develops amongst the Editors/Experts.

Some iNat observations are made by high school students, and their observations usually truly suck! It would be nice if they didn't suck so badly. But unless someone wants to play policeman for a global community of observers, I don't see anyone here improving that state of affairs. I just ignore them. The ignorance of the general public cannot be overstated--most don't know the difference between a honeybee and a housefly. Hopefully, at least some of those folks will become less ignorant over their course of engagement of nature via iNaturalist (or whatever tool they use). I don't think that's anything we can fix either (other than helping to educate them with the same kind of patient and polite and encouraging engagement that we ourselves would like to receive for things we are ignorant of). Remember, just because you help one person learn how to ID a honeybee doesn't mean you've taught everyone how to do so (it's one person at a time). It's like teaching students in a college entomology lab, you can't just tell them the name of the bug that they show you, you have to guide them to the relevant literature, help them learn the identifying characters, etc--all while not going insane because you're answering the same questions over and over. Speaking of which, I'm off to teach my next class...but first, let me help a person on iNat learn the distinguishing characters between Zelus renardii and Zelus luridus.....again (I've developed a guide for that very purpose).

發佈由 pfau_tarleton 超過 6 年 前

Updated Q #5) Why is it acceptable for contributors to submit images with no organism or no detectable organism as an identification request? This practice sure does waste our time. [Update: note that I am objecting to such images being submitted as ID Requests, not their inclusion on the site as a whole]"

Again, it's not acceptable, and I'm not sure what your recommendation is here. Do you propose a method of stopping them somehow before being uploaded? Just mark as "no evidence of organism" in the Data Quality Assessment section and it's marked as "Casual Grade." Casual grade observations are not included in Identify by default. Users who refuse to follow the iNaturalist guidelines despite being warned are suspended. Flag the observations if the behavior is persistent and a curator/admin will take a look.

發佈由 bouteloua 超過 6 年 前

There's no such thing as an "ID request" on iNat. Everyone uploads observations, and if anyone wants to identify them, they can. It's impossible to upload observations for inclusion on the site and not have them available for anyone to ID if they want to.

What exactly IS BugGuide's criteria for selecting people who are good enough to be allowed to ID submissions?

發佈由 pfau_tarleton 超過 6 年 前

No time to respond to all of the above but I would like to ask one question: Why is it so hard to understand and validate my desire for people to cite the hierarchy of occurrence County: State: County: general locality: details of locality for all records. Yes, I realize that one can zoom into the Google Map and figure this out, but this places a burden on the identifier to do the zooming. I understand if Google Map technology makes it difficult or impossible for contributors to cite localities as preferred for biodiversity records. However, if so, why not admit this technical limitation rather than being so defensive and trying to excuse failure to cite the most basic locality information information in association with all records. In general, I understand that there may be good reasons why records with no obsservations persist for years on the site or why people who can't or don't both to identify Apis mellifera insist on identifying Bombus (Pyrobombus) or Amegilla (Zonamegilla) to species level. But why argue with me that such practices are optimal or best for the site? As for unwelcoming comments, if the site were set up to encourage best practices by contributors from the outset then there would be less occasion for these. For example, there should be an easier way for identifiers to unlink unrelated images when these are submitted.

發佈由 johnascher 超過 6 年 前


發佈由 calebcam 超過 6 年 前

I don't think anyone is being defensive about the County/State/Country info. The information is all there, though. Automatically. e.g.

I have proposed (https://groups.google.com/d/msg/inaturalist/h21UNGFxx9Y/K2K3UvuEAwAJ) that this same pop-out or at least the county/state/country information be somehow displayed on the Identify page as I think that would be quite useful (right now you must click through to see it on the individual observation page).

I also don't think anyone is arguing that people who identify things incorrectly or with no basis of knowledge is either optimal or best for the site. I have often pointed out to people not to "trigger-happy identify" things. People get quite defensive about this though.

"For example, there should be an easier way for identifiers to unlink unrelated images when these are submitted."
Yes, it stems from the level of control the user should have over their own content (much higher control than sites like BugGuide). Worthy of discussion is further abilities for site curators to have for content curation/data quality control.

發佈由 bouteloua 超過 6 年 前


Yeah, what cassi said. This is a site where people share their observaitons of nature. We want to discourage outright wrong data such as things mapped in the wrong place or spam photos. Beyond that, any observation of a living thing is acceptable with a few exceptions. Wild things are best.

If iNat required I type in location data manually (i use the app), required me to crop images, or any other such nonsense I'd leave the site. I use thsi site to record as much biodiversity info as I can as quickly as possible. Sometimes the photo is blurry, sometimes there is no photo (those don't come up on ID please though). Sometimes the photos are aesthetically pleasing and sometimes they are ugly. I don't really care. When i take a particularly beautiful photo I put it on instagram. When i find a naturally occurring organism I want to record I put it on iNaturalist.

If you have super specific needs for how you want your data to be, you need to design your own methods for tracking it. That's what I do for work data. Then when appropriate I share to iNat. A lot of it I don't for lots of reasons and that's ok. But I share what I can so everyone can see it.

The site is mostly awesome, though I have ideas on how it could be better, like anyone else. However if you profoundly dislike the way the site works, I recommend you not use it.

發佈由 charlie 超過 6 年 前

As it goes with all too many discussions, the participants find themselves talking past one another.

Having observers manually typing in locality information for each observation might be good practice to verify the locality data (i.e. if it doesn't agree with coordinates, then obviously one or the other is incorrect), but that doesn't appear to be your argument for doing such. The locality information you seek (as an identifier) is readily available (as Cassi pointed out). Having it more visible would indeed be an improvement and has been discussed in Google Groups.

Poor practices are obviously not optimal or best--surely none of us is suggesting that to be true. Is the solution to have a bunch of us designated as the deleters of suboptimal observations? I'll pass on that job, thanks. At iNaturalist, each person owns their observations and has complete control over them--even if they're crap. The purpose of iNat is not to collect great observations, but to allow people to make their observations public.

I don't understand why you want to have ultimate control over other peoples observations. If unrelated images need to be unlinked, point that out politely and the observer will almost certainly respond. If not, there are ways to flag the observation. Perhaps there needs to be a better way to flag observations like this. No one is suggesting that there isn't room for improvement.

I don't think any of us are developers, however. So I'm not sure how valuable our discussion is here. It would probably be more productive in Google Groups in the presence of the developers.

發佈由 pfau_tarleton 超過 6 年 前

Doesn't do much good to have location information available if not readily viewed on the identify page.
If not "readily visible" then how can we expect to make an efficient identification?

Rather selfish isn't it to be unwilling to make minimal effort to crop photos before sharing so as to facilitate identification?

Prioritizing the needs of those who submit data as opposed to those who review or view it creates structural problems with the site that should be addressed.

Observers do not routinely respond to requests to unlike their images.

Surely inaturalist can be an efficient archive for biodiversity data, so why be so dismissive of this function?

發佈由 johnascher 超過 6 年 前

my goal is to gather as much data as possible, as such I have over 27,000 observations. No, i don't feel 'selfish' putting in the thousands of hours to share this biodiversity data with other scientists as well as everyone else on the planet with an internet connection. I could spend time screwing around with my photos and re-uploading them, but if i did that I'd have a fraction of what I have now. I'm more concerned with gathering biodiversity data than making pretty pictures. If you want pretty pictures instead, you could try Project Noah.

FYI am a scientist and policymaker who processes large amounts of data, iNat is something mostly that I do on the side or when i find data that I can share publicly. I would much rather have verifiable but blurry data than not have it, and indeed I use such data on iNat all the time. You don't speak for all, or even most of us, and honestly, calling people selfish because they don't care about shiny pictures that meet your criteria is pretty poor taste.

The magical thing about the Internet is it isn't compulsory. if you dislike our community of tens of thousands of active members and many millions of data points, no one is making you use it. Actual feedback is great, but henpicking and complaining and calling other people names is not. To be honest I couldn't care any less if you like my blurry photos of trees taken from the road, or not. We have so little data on things that are so important, and frankly it's absurd to request i stop collecting data because you don't like the photos.

For what it's worth I'd support a cropping or tag+zoom feature being built inside the website, i think it would be awesome, if not high priority.

(to be clear I speak only for myself, not the admins, who are much more tolerant of rude behavior than I am)

發佈由 charlie 超過 6 年 前

Wow good replay @charlie my thought exactly

發佈由 ck2az 超過 6 年 前

John, I'm going to quote what you told me in response to my thoughts regarding BG improvements: "How about contributing more yourself to this and similar sites and reading through the archives before telling us what we should be doing? Most of your suggestions were proposed long ago and we already have taken them into consideration." By the way, why didn't you scold Edward L. Ruden (contributing editor at BG) for his comment (following yours) suggesting that BG halt frassing now to prevent data loss as it transitions to the new system? Do you not scold BG insiders?

Here's a link to a discussion regarding one of your complaints--since you can't be bothered to "read through the archives":

I think our patience are starting to run thin. Some of your suggestions for improvement are good (and we agree with some of them). But they're already being discussed in a location where the developers can see and respond. BG hasn't been improved much since it's inception--iNat improves regularly. I'm not criticizing BG--just pointing out your hypocrisy. I think it's great that you're suggesting improvements (I've done so myself), but puuuleeez read the forums first and then discuss your concerns with the developers in those forums.

End of discussion.

發佈由 pfau_tarleton 超過 6 年 前

I was sharing my first impressions of the site with whoever cared to read my thoughts, not attempting to make a fully informed global critique.

My objections to the extent that these exist are not with the site itself so much but rather with the attitude of those unwilling to acknowledge the desirability of citing and displaying locality data correctly and of posting photos where the subject is detectable among the background, etc.

發佈由 johnascher 超過 6 年 前

John, you had posted a bunch of questions and I made an attempt to answer them with some historical and philosophical context that I know that you, as a new user, did not have. I'm glad some of the questions have been answered and issues resolved. I'm confident that several others will be resolved soon, based on ongoing conversations on the Google Group, which I again would recommend you read and contribute to. I would also recommend that you add some observations so that you can see what it's like on "the other end" of the website. Doing so, I think, would answer several of your questions.

philosophically (from the site's about pages)
"iNaturalist is a place where you can record what you see in nature, meet other nature lovers, and learn about the natural world."
"iNaturalist is an online social network of people sharing biodiversity information to help each other learn about nature."
"our primary goal in operating iNaturalist is to connect people to nature"

iNaturalist just also happens to be the best place out there for crowdsourcing biological occurrence data, but per the above, that's not its primary goal. Telling people their photos suck, without kind, constructive recommendations, or implying that they are stupid, selfish, or slow...or valueless because they are not an expert (whatever that is) isn't very productive toward building a community that encourages people to connect to nature.

Thanks for your thoughts -- everyone.

發佈由 bouteloua 超過 6 年 前

What I would really like to see is some support for my efforts to oppose the tragedy of the commons in general, even though my particular recommendations may be naive. No doubt I have not been as diplomatic as I could have been, but I'm sincerely trying to raise the level of discourse. Can those accusing me of being rude while insisting on their right to post blurry, uncropped, improperly labeled, and carelessly misidentified photos say the same?

發佈由 johnascher 超過 6 年 前

you are trying to 'raise the level of discourse' by calling people selfish because you don't like their pictures. Sorry, but lol. Seriously. Go back and read your own posts.

We don't have a 'right' to post anything. it isn't the constitution of a country guaranteeing rights. We fillow the terms of service, and when we do that, we can post photos of whatever sort we want. There is no requirement for cropping, lovingly focusing every image, using photo editing software, anything like that. And to be honest, if you, like most people here, care about gathering biodiversity data, you won't care about the aesthetics of a photo. if it can't be identified skip it. If it can' identify it so that we have the data point. Period. If you don;t like the photo, ignore it. It isn't yours to criticize or whine or toss insults about.

Anything that is mapped in the wrong place or misidentified needs to be corrected. Good thing there are mechanisms within the site to do that. If it's mapped in the wrong place, use the data quality assessment to note that it's mapped in the wrong place. If it's misidentified add the correct ID. If it's misidentified by several people and they refuse to fix the ID, consider sending an @ to some other users who also know the species, to get more correct IDs to override the wrong ones. Pro tip: it will be easier to get people to help you if you don't call them 'selfish' or otherwise insult them. But you know, if you want to run around pissing everyone off, that's your right to as long as you don't violate the TOS.

發佈由 charlie 超過 6 年 前

Let's all tone it down a little please. : )

"What I would really like to see is some support for my efforts to oppose the tragedy of the commons in general"
John, did you get a chance to check out https://www.inaturalist.org/pages/responses ?
Those copy/pasteable phrases include several ways of encouraging people to post better observations. Open to other suggestions to add to the list.

發佈由 bouteloua 超過 6 年 前

I am interested in discussing best practices not arguing about what is minimally acceptable.

It is remarkable to me that people who comment in forums here or at BG tend to be concerned about their own photos (Don't frass it! Keep it on the front page of ID Request so everyone will see it...) and so rarely concerned about the implications of their choices for the site and its users.

I would prefer that a well-conceived site and a community culture emphasizing data quality will facilitate best practices. I do not think it efficient to address such issues routinely at the level of individual observations.

發佈由 johnascher 超過 6 年 前

I don't think you're actually really reading or understanding other people's posts. It isn't about caring more about photos than data. I haven't seen anyone say they want their photo on the 'front page of id please' and there isn't even such a thing anyway. and while I know that frass is insect poop I have no idea what 'frassing' is. I guess you want to be able to delete my photos and thus remove research grade from my observations if you don't like them? What a horrible idea. If that's how bugguide works I guess i won't ever bother with it. The site and users are better served by having more data than having pretty photos taken by a few self obsessed users. you seem to think you speak for the site or users when no one here is even agreeing with you. Seems you speak for yourself and your desire to control what others do. That might be how it works on bugguide but that isn't how we do things here. Again, if you don't like it, you can go back to bugguide. I'm glad that you are helping with insect ID or whatever but if you hate the site so much, I imagine you must be in agony the whole time.

We do have a well conceived site and a community emphasizing data quality. What we don't have is your ability to slow down, listen, and see how the site works. you think your ideas are new, unique, and important when in reality this has already been discussed to death. There is a proposal to allow people to rank photos which would inform what shows up on the front page. In the mean time as someone already said you can fave good photos. If you are sure something can not be identified by the photo you can mark it as such, but don't do this lightly because others might be able to ID what you cant. If the photo is sufficient to identify the organism, then it is sufficient to be on the site. period. This isn't instagram, bugguide, or 'rate my photo'. While the primary purpose of the site is to connect people with nature the primary purposes of the site for most 'power users' is to collect biodiversity data to be used in conservation, and attempts to editorialize, control their content, and block their ability to do that will just make you unwelcome.

i don't think i'll be doing more responding here unless it appears that you are listening, since pointless back and forth will just fill up cassi's notifications. But if you read nothing else, read this: it is what it is. Whining, complaining, yelling, insulting, and other such behavior will not earn you the response you want.

發佈由 charlie 超過 6 年 前

@johnascher, I agree that more could be done to encourage the right kind of observations from the outset. But it also does not feel futile to me to address them individually because at the same time, I get to personally welcome and speak to another current or potential nature nerd.

As far as well-conceived and data quality, what do you see working on other websites that you don't see on iNat? Did you have some recommendations besides giving curators the ability to unlink photos? iNat has more crowd sourced data quality measures than any other website I've seen before. Like Charlie says, if a photo is unidentifiable to species, you can mark it that way. And yes, be careful with that checkbox and only mark it once the community ID is the lowest possible taxon. It immediately removes it from Identify. Same with everything else in the Data Quality section -- a thumbs down will immediately mark it as "Casual Grade" and remove it from the Identify pool.

Cropping, for example, cannot be a sitewide encouraged activity. I often take photos of ecosystems showing their vegetative components living together. Cropping is irrelevant there. What would be helpful is nicely nudging folks to crop photos with the encouragement that yes it's easier to identify them if cropped. How would you do that other than individually? That's certainly how it worked on BugGuide when I uploaded my first observation - someone was quick to point out it wasn't cropped closely enough.

Citing localities is irrelevant because of how it can be displayed programmatically, e.g. the screenshot above and see how they are now displaying country on Identify. It's not perfect -- I would want municipality (when relevant), county, state/region to always be displayed, but it's a good start. This occurred because of ongoing, constructive, community discussion on the iNat forum. Join in. :)

發佈由 bouteloua 超過 6 年 前

Cassi, is there a page somewhere that documents each of the changes made to iNat as they're made? Something like a "versions" history?

發佈由 pfau_tarleton 超過 6 年 前

Yes, on Github (though I don't 100% understand how this stuff works). It's really great that this info is so transparent though!

發佈由 bouteloua 超過 6 年 前

Sure has been a lot of vitriol directed at me and impugning of motives. I am simply requesting some rather basic improvements to an already excellent site and community that I have enjoyed a great deal so far since reengaging. Am I such a terrible person for wishing that the genus name was cited along with a subgenus, that photos be cropped at least to the point where a subject is detectable (if the photos is submitted for community ID), that reviewers be empowered to unlink photos, etc? Some of your critiques may be valid, but I refuse to accept that it is irrelevant to cite state and county prominently on any sort of biodiversity record submitted for identification. A few of my comments will only make sense if you also follow the BG forum, where users routinely complain about the fate and position in the guide of their photos, so I apologize if this was off topic..On the topic of BG, those who are too thin skinned to post there are missing out on a wonderful resource and community.

發佈由 johnascher 超過 6 年 前

I feel the need to speak a bit more loudly, so pardon me. WE DO INDEED WANT LOCATION TO BE MORE PROMINENT...NO ONE IS ARGUING OTHERWISE (there may have been some confusion earlier as to what you were referring to or asking for). WE ALSO WANT THE GENUS NAME TO BE INCLUDED ALONG WITH THE SUBSPECIES NAME (there's much discussion about that in the forum). I think it's absurd for anyone to complain about fate and position of their photos in the "guide" of BG (the guide should include only the best of the best and curated by BG editors). I do, however, think it reasonable for some BG users to agree with some BG editors that images not be discarded because every specimen is a datapoint. I was scolded for doing that despite there being debate among BG editors regarding function of BG as a guide or as a data repository or both. An editor discarded one of my observations posted on BG because he thought it wasn't identifiable. I took it out of frass and soon thereafter Mike Quinn came along and confirmed my ID (my observation is currently the only living specimen of that species on BG).

I don't understand why we seem to be talking past one another...can we not read what the other is writing? I feel like I'm in an episode of the Twilight Zone.

發佈由 pfau_tarleton 超過 6 年 前
發佈由 pfau_tarleton 超過 6 年 前

Sorry for the "talking past one another" but in my defense I'm responding to several people, some of them quite uncharitable, about nearly a dozen topics, with others bought in incidentally to further confuse matters.

Maybe it is not the right venue to talk about BG at all, but I can say that discarding of credible submissions there i.e. "frassing" will cease in the future when site capability increases (version 2.0), at which time lower quality images can be retained without compromising the guides. There are technical solutions to some of the most contentious issues, and in the future we may be able to have the best of both worlds. Regarding inaturalist, my critique was not so much that people retain personal observations with low photo quality, missing organisms, etc. as that is their business. My objections was that unidentifiable images tend to remain in the "Identify" tab for years even if no organism is visible. I now realize that inaturalist has some mechanisms in place to remedy this, but it takes longer than I would prefer. I'm not sure if the necessary fix is technological or just better use of existing tools by a more active pool of identifiers or simply more patience on my part.

Glad to see progress regarding subgenera!

發佈由 johnascher 超過 6 年 前

The ICZN makes clear how to cite animal names, so why has correct citation of subgenera for these been contentious or perceived by the programmers or anyone else as optional or a matter of opinion?


發佈由 johnascher 超過 6 年 前

Correct citation has not been contentious. The developers are few and time-strapped. Prioritization of re-coding the website is the issue.

發佈由 bouteloua 超過 6 年 前

I’m beginning to understand where you are coming from. and some of your
Concerns as @bouteloua has pointed out this site is continuously evolving
Writing code and putting it to work is something I’m not familiar with
But as she said those who are doing it are few,
I’ve seen a lot of changes here since I joined 15-16 months ago
Ideas seen to be embraced here, the resources are limited do you have a idea to possibly bridge the gap ???

發佈由 ck2az 超過 6 年 前

I am not so presumptuous as to think that the first impressions of the site that I shared in my profile should be considered an adequate basis to actually improve coding of the site. My main intent was to highlight seeming problems in usability and data quality that were readily apparent upon casual reengagement, not to claim novel or comprehensive insight into these issues or how they might be fixed.

Also note that a lot of my concerns were not technical but were instead with a site culture that seems to prioritize being nice to its least conscientious contributors (if not to me! See above) over improving data quality.

發佈由 johnascher 超過 6 年 前

John, most of us on this thread spend a lot of time engaging new users politely and simultaneously improving data quality. They're not mutually exclusive activities. :)

發佈由 bouteloua 超過 6 年 前

Anyway, this has sort of been beat to death.
Happy to stop going around in circles and hone in on those novel and technical resolutions for fixing actual issues.

發佈由 bouteloua 超過 6 年 前

John, I really don't think I was any ruder to you than you were to me. Really. Go back and read the posts. Now that doesn't mean I should be acting rude just because someone else does it, and probably i could have been 'nicer'. But i mean, you called me 'selfish' immediately after my first post so that kind of set the tone there. Or maybe you were talking about someone else? Anyhow, i am up for resetting the conversation in a less adversarial way if others are too.

I think there is a real solution to the cropping annoyance, which is that there should be cropping or 'tagging' ability on iNat (tagging would be like the facebook thing that lets you tag faces, only it also lets you zoom in). A lot of people use the app and there's no real way to crop from within the app.

發佈由 charlie 超過 6 年 前

It seems like the app has been slower to see improvements, but a cropping feature should be top priority. We gotta remember here that most users are rank amateurs (if that). Quite a few students also. If we tell them all that their pics suck or just go around deleting them left and right, that's the fastest way to turn people away from nature that I can think of. It's always good when folks that point out weaknesses also provide potential solutions. That allows everyone to better understand the problem and come to a concensus on how to improve.

發佈由 pfau_tarleton 超過 6 年 前

i actually think a cropping feature on the website itself would be more useful than the app. The app is great for gathering data but it isn't so great for manipulating and editing data

發佈由 charlie 超過 6 年 前

Oh, I agree to that also! Vast majority of people don't know how to use image editing software to edit a photo. Even younger folks--especially younger folks. I'm constantly amazed at how unskilled college-aged folks are when it comes to technology (beyond video games and snapchat).

發佈由 pfau_tarleton 超過 6 年 前

well, i don't know about others... but i don't really have a lot of time to mess around with that sort of software especially since my computer is super slow. I'm sure I could learn to use it if there was a need but I don't want to do it just to make the photos pretty. If it would help get an ID of a particularly difficult observation, sure, if my computer could do it.

I don't think the 'kids these days' use image editing stuff much. But from what I've seen with college kids using this, especially for classes, there are a lot of other issues that rise to the top above poor/no photo editing. but there are whole other threads about that

發佈由 charlie 超過 6 年 前

There's the issue of time, not being sure what to crop, it having a bad quality photo. I have an old copy of Photoshop, so it's easier for me, plus that's my background. My video gaming past time has not eaten into my skills. ( I kid, I know we're just keep kidding around. ;) )

Though, I'm kind of amazed how many people younger than I am (I'm admittedly on the early cusp of even being considered a millennial) don't have desktop or laptop computers. But I'm getting off topic.

發佈由 tanyuu 超過 6 年 前

Thanks for the interesting comments.

Yes, I should have adopted a friendlier tone and will aspire to do so in the future.

發佈由 johnascher 超過 6 年 前

I would like to revisit the query, "What is John's solution?" in response to my question: Why is it acceptable for contributors to submit images with no organism or no detectable organism as an identification request? This practice sure does waste our time."

My solution would be to employ AI to flag and attend to ID Requests with a removed or otherwise non-existent images and to flag exact duplicate images submitted under the same taxon name and to otherwise clean up such things.

Why employ the robots only to attempt species IDs, which is demonstrably not working very well, as opposed to having them clean up myriad examples of duplicated or unhelpful content?

AI should also be used to address some of my other points such as point 1 about correct citation and mapping of localities. For example, why can't AI flag cases where reported coordinates do not match the locality (at least when the mismatch is at continental scale such as a US report mapping in China)? Also, why can't AI "fill in" crucial structured locality data using the reported coordinates even if that is not directly input by the observer?

In general, it seems that AI is used to try to replace or enhance taxonomic expertise but instead it wastes the time of expert identifiers by empowering myriad misidentifications.

Meanwhile we are expected to live with or address manually all sorts of problems that could easily be attended to by improved robots.

發佈由 johnascher 超過 1 年 前

Hi John @johnascher - here is an existing feature request you may be interested in voting for or discussing further:

Duplicate prevention: Notify observers if their image checksums match others on the site

For your other suggestions, you could perhaps propose your own feature requests like: "train computer vision on observations that have been marked as containing no evidence of an organism in the Data Quality Assessment section".

For the place name vs. GPS locations, I think an algorithm would struggle with localities such as Peru, Illinois vs Peru the country, or Ottawa, Illinois vs. Ottawa, Canada. Do you know of a comprehensive dataset in multiple languages that would address that?

why can't AI "fill in" crucial structured locality data using the reported coordinates even if that is not directly input by the observer?

What do you mean by this one? The iNat system does already fill in locality notes automatically. I almost never edit the locality notes myself. For example all I did was let it pull the GPS coordinates and it automatically generated the neighborhood I was in:
Canal Park, Duluth, MN 55802, USA https://www.inaturalist.org/observations/134761894

I'm just an iNat user like you, so here's more about how the Feature Request category of the forum works in case you want to propose your features to the iNat community and staff: https://forum.inaturalist.org/t/about-the-feature-requests-category-please-read-before-posting/69/

發佈由 bouteloua 超過 1 年 前

"all I did was let it pull the GPS coordinates and it automatically generated the neighborhood I was in:
Canal Park, Duluth, MN 55802, USA"

That is what I am hoping can happen, but in many cases an observation is mapped but there seems to be no automated filling in of the state, county, and other fields.

In some cases one can "view" such political area data but when in Identify mode it is not accessible. Not clear to me why crucial information about political area of the image is thus inaccessible to those using that mode (where accessing locality data is crucial)? In some cases my comments were disparaged because there is some way to address the issue within the site, if you know where to click (in this case on the not especially obvious "view" button) followed by another click on "details" under the map (also not obvious). However, when in Identify mode where immediate/efficient access to locality is crucial needed details may be or may have been in the recent past lacking (the site may now better autofill and display locality info in which case more records even in identify mode may have state, county, etc. cited).

I tried to propose various features to the group that might improve data quality but was rebuffed by various curators, often quite uncharitably, who preferred to defend the status quo even though there are indisputably many inefficiencies. The gratuitously negative and in my view largely unfair reception to what I thought were quite reasonable requests to improve the site (esp. data quality and standardization) have led me to question whether iNaturalist really is the nicest place where one can find reliably civil discourse (see NYT article https://forum.inaturalist.org/t/inaturalist-in-the-new-york-times/37790).

Our community would become nicer if there were more respect for taxonomy and for taxonomic experts who volunteer our time. I would also like to see more efficient tools to ensure data quality (e.g., fixes for duplicate and wrongly linked images; better ways to flag out of range species as such without the need to write an essay each time on the subject). It would become more civil if safeguards were in place to ensure that data once entered could not be mass deleted. it would be very nice if mechanisms were also put in place to discourage large-scale data mining by academics who rightly see iNaturalist as a rich source of data for their publications but who often do not see fit to include or credit identifiers in their studies. It would also be nice if full det. histories could be exported to GBIF so that 1) the work of identifiers can be properly credited and, most importantly, 2) so that fitness for use of records can be properly validated. Without knowing who identified images how can it be possible to use any of them for science with any confidence?

發佈由 johnascher 超過 1 年 前

@bouteloua now that I understand the site better I can see that some of the locality data I was hoping to see can be found "filled in" under "details" in "view" mode (and when reviewing records in Explore mode), but at the time I was commenting I did not see state and county data being routinely and correctly reported when reviewing records in Identify mode (likely things have improved nowadays).

So it is possible to find "details" about localities but this may entail two extra clicks and to me (and perhaps others) it was not immediately clear that the needed sequence was to click on "view" [wasn't I already "viewing" the record in question?] followed by "details" [country and state of the record are mere details?]. Am I wrong to think that this is a rather unwieldy way to locate what to me is essential information to inform identifications?

發佈由 johnascher 超過 1 年 前

I am not an expert in AI by any means so am reluctant to comment but I do wonder why it is applied to seemingly very difficult problems (species identifications) and not to seemingly simpler problems such as 1) removing records lacking images from being displayed in identify mode and 2) detecting exact duplicate images posted under the same taxon (it is fine to post the same image under different taxa if there are multiple present)?

I also don't understand why the AI cannot be upgraded to, e.g., display only Apis mellifera and not Apis cerana (Asian Honey Bee) when suggesting identifications for honey bees from the Americas? Is it that hard to add a negative geographic constraint by hemisphere or continent at least for the few very famous species that are most often misidentified?

發佈由 johnascher 超過 1 年 前

@johnascher: I tried to propose various features to the group that might improve data quality but was rebuffed by various curators

See my note above:

@bouteloua: I'm just an iNat user like you, so here's more about how the Feature Request category of the forum works in case you want to propose your features to the iNat community and staff: https://forum.inaturalist.org/t/about-the-feature-requests-category-please-read-before-posting/69/

It looks like you still don't have a forum account, so there are no proposed feature requests under your name. You can log into the forum using your iNaturalist credentials at https://forum.inaturalist.org/ and create feature requests, such as the above, there.

發佈由 bouteloua 超過 1 年 前

I believe there was a grant from Google or some other group to develop the identification AI algorithm. No one gives grants to create AI for menial data management. Perhaps they should, but they don't.

Also i find it odd to see you say iNat needs to respect taxonomists more, to me the site seems to weigh taxonomy heavily over other uses, like having the site usable to non-taxonomists. It is heavily oriented towards constant taxonomic change, sometimes before papers are even peer reviewed. I suppose it all depends where you are standing. The site has always been slow to develop and adopt new things, except when it was new. They always cite the tiny staff, and it is true they are understaffed, which is familiar to anyone working in the ecology field.

I'm not sure iNat is really as nice as the article describes, i've experienced and seen some pretty not nice behavior and the persons continue to do so sometimes for years. But 'nicest place on the internet' could still be true in terms of websites with public access to comment, the Internet isn't very nice overall.

And yeah, if you don't use the forum you aren't going to be able to propose new features, for better or for worse, that is where proposals are collected or tracked.

發佈由 charlie 超過 1 年 前


登入註冊 添加評論