ID Feedback Loop

[Upupa epops]

Hi everyone,

Recently I came across a problem with bad research grade observations that I think has a pattern that could possibly be solved. Sometimes there are many research grades observations of a species which I don't know if there's anyone on iNaturalist who knows how to competently separate it from similar species. I think this is caused by a feedback loop with several different aspects.

It starts by people identifying based on what looks similar, without investigating potential similar species. For example, people will look on Google Images or an identifying website, find a species that resembles their photo, and identify it to that species. At this stage, when they're only incautiously identifying their own observations, it's not a big deal because someone will usually correct them. But if not, eventually they develop the confidence to identify that species on other peoples' observations, and that's when the next step comes in.

We're all aware of a tendency on iNaturalist to agree with an identifier without double checking. If I go through and identify a number of observation to a species that is hard to ID (several similar species, obscure distinguishing marks, etc.), a good portion of the observers will agree with me without comment. Then the observation will become research grade. I highly doubt most of them double checked the ID using external sources. This is fine if my ID is correct, but I'm liable to make mistakes, and people are unlikely to distinguish between an authoritative identifier and someone who isn't. So if someone is determining their ID's based on the previously mentioned method, there will now be unreliable research grade observations.

After a certain number of identifications on others' observations, other people will start to think that this is the correct ID and use it themselves on other' observations. If this is a common species and these things start happening, pretty soon there will be enough RG observations to show up in the AI suggestions. Evidently a lot of people will base initial IDs solely on AI suggestions, which encourages the pattern even more.

Here are a couple examples:

Lucilia blowflies, with currently 168 RG observations of a couple species combined. There are several similar genera, and within the genus the species are difficult or impossible to separate without very good photos from several angles. Some of the RG observations were identified based on good photos and careful guesswork, but most are just people agreeing with each other based on the AI, as far as I can tell. There were even a number of RG observations of a European species not found in North America, although I think those are gone now.

Eupeodes americanus, which has (or had) almost 100 RG observations. There are 2-3 nearly identical species in North American, and 3-5 similar genera. Most of the observations have the genus and subgenus correct (whether by skill or coincidence, I don't know), but some definitely have the wrong genus, and I'm not sure if any of us on iNat have the experience to ID them to species level. When I ask on an observation how the species was determined, the answer is usually "Because there are lots of other RG observations of it in the area, and it looks similar."

These examples can be solved on an individual level; I can go through all the E. americanus observations and either add a higher level ID or mark them "Community ID can be improved" (A few have already been done, so the numbers above are smaller than they were), but the cause of the problem will remain and I'm sure there are other examples. 

One possible problem keeping the pattern going is that it's often with difficult species that many identifiers are reluctant to engage with. I've suspected the issue with Lucilia for a while, but I'm reluctant to go through all the observations and do anything because I don't know the group very well. There's a good chance many of the observations are identified correctly, because they are common species, but I don't have the expertise to say for sure whether or not it's identifiable based on the photos available. So it's either that I leave the issue and there will be many invalid RG observations, or I go and get myself involved with it, possibly rejecting many valid RG observations (which will hopefully be re-identified at some point).

Now, it's possible I'm just crazy and I'm seeing a way bigger problem than there is. Please let me know if that's the case. :)  And maybe the only solution is to retroactively chug through them all and correct them. 

A couple ideas I had for improvement:

When you press "Agree", a message would pop up saying "Are you sure you have eliminated similar species?"

I'm guessing people automatically agree because they assume the identifier is an expert, but that's not always true. This would encourage them to confirm independently or ask questions, I hope.

Also, the taxon page does have a great similar species tab, but it is fairly hidden. I don't know how it would be done, but maybe it would be more useful if we could manually add similar species and genera and if it was more visible in situations where you would be likely to mistake similar species while identifying?

Thoughts?

Thanks.

[Charlie Hohn]

Yeah, these things are real problems, and believe it or not similar issues exist in 'non citizen' science as well. There are so many species, they are so hard to learn, and there's no real way to get anything done if you key out each species each time you see it. I personally don't think a nag screen would help. I know for me it would just make me ID stuff less often, and hopefully my IDs are usually correct! 

I think a lot of things COULD help... first of all continuing to build the community... and making it an environment where people respond to questions and discuss ID with each other without feeling threatened. Also better ways of dealing with when things are out of range. The student issue plays in here as a lot of the time students are adding IDs to each others' observations. Potentially eventually adding the merit based reputation system that has been discussed on here before. Better options for sorting through one's old IDs. i  know in the past i have made an ID but later reconsidered and not been able to easily search for and find that ID. For instance if i learn more about how to identify a difficult sedge, I may wish to review all my past IDs of that sedge to make sure I still think they were correct. Continued improvement of the algorithm with more good data. I like your idea of manually adding similar species data. I even think it could be neat to have a page where users can add notes about how they identify a species - perhaps a wiki like page for each species with notes and maybe even photos, similar species, etc. And even notes that could pop up when identitfying. I've been sorting through a lot of ponderosa pines identified in LA which are actually Canary Island pine for instance. Maybe if there could be a non obtrusive note that pops up when you ID a ponderosa pine that has a note like 'Note that at low elevations, this species is very rarely present'. or whatever. Also, encourage people not to agree with an ID if they have no idea... now with the new way community ID works they don't have to to get that labeled.

C

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

============================
Charlie Hohn
Montpelier, Vermont

[jesse rorabaugh]

Yes, this is a problem. To name two more species I know which are in the same boat: Aphis fabae (black bean aphid), Uroleucon rudbeckiae (brown ambrosia aphid) and Tamalia coweni (manzanita leafgall aphid). All three are technically very challenging to identify. I am sure half or more are correct, but I doubt more than a couple have been conclusively identified.

This is something that once discovered a small number of experts can take on in just a few minutes. I have done so on occasion. It is something which does result in some angry people though when you push back their observations. It therefore takes me quite a while to build up confidence to take on one of these species.

[paul_t...@ymail.com]

It would be very helpful if the curators of a taxon where this occurs could flag particular species where identification to a species or genus level is not possible or is difficult from a photo alone. Happens in birds also where a call is needed to confirm the correct species. In those cases more stringent criteria for upgrading to research grade would be required.  It would also help non-expert users know that the genus or higher level is the best level they can identify to.

[Robert Taylor]

These issues may be partially solved by weighting the IDs of people who are recognized in the community as experts in a specific group more than those who have some or no knowlage.

These experts will have the responsibility of IDing the observation to the most suitable level. These same experts are likely to be the ones to suffer if they are reckless with their IDs as they are most likely going to be the ones using the compromised data.

[Chris Cheatle]

A couple of comments on things above:

• I would stay away from any new "are you sure" dialog box. After about 3 instances of seeing it, it will simply become another box no one reads or considers before just clicking yes to make it go away
• one possible route to help here is increased discussion and awareness of the acceptability of genus or higher level records at research grade. I think there is a desire (or in some cases "need" in class assignments) to see records get this label. Scientifically, at least to be a solid confirmed genus level record showing occurrence at a place and time is more valuable that a "guess" among different species levels. I see this behaviour even among friends going to species level on their submissions when in areas that supposedly require microscopic evaluation to differentiate at species level strictly because they want the RG status. Either the record gets dubiously confirmed, or sits pending forever, when had it been done at genus or family level, it could have progressed.

Thanks.

[Charlie Hohn]

Another approach, though not always popular, is to expand groupings at a subgenus type level. For instance if a genus of flies has 200 species, but you can certainly narrow it down to 2 species, is there a way to track it as such. That's a tricky one from a database standpoint and proliferation standpoint (the two honeysuckles I can't tell apart in Vermont might not co-exist elsewhere, or perhaps there is a third species in Texas that also would be in the group there, etc). 

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Upupa epops]

Perhaps part of the reason people are so eager to get to species level for RG is that it's really difficult to get RG on a genus level observation. 

We can mark them "Community ID cannot be improved", but people are pretty reluctant to do that. It's possible we'll discover a new field mark that can be used to separate these species, and it's possible an expert with way more experience will come along and be able to tell, so therefore technically it's possible that the Community ID can be improved. Somehow we need to come up with point at which we can be reasonably sure these observations aren't further identifiable. Ultimately if they are identifiable, a genus ID isn't incorrect and the expert can go back through the RG genus observations and identify them.

Do you think it's safe to make an observation RG if there are 3+ consecutive genus IDs? Actually, is there anything wrong with making a community agreed genus ID RG right away? Or maybe using the newfangled "Community Grade" from the cultivated observations thread?

Charlie, I think species groups like that are great and would be quite helpful in situations like this. BugGuide puts nearly all of these confusing hoverflies into species pairs: https://bugguide.net/node/view/739078/bgpage?from=5 And adding a species pair ID would make it pretty clear to the observer that this is a difficult to separate species. EBird supports "slashes" (e.g. Sharp-shinned/Cooper's Hawk) which I use quite regularly. However, iNat does not support them: https://www.inaturalist.org/observations/8993064 I can understand that if the iNat taxonomy is meant to represent actual taxonomy.

Fortunately for the hoverflies there is a subgenus (Metasyrphus) that we can put them in.

On Friday, 30 March 2018 09:09:27 UTC-4, Charlie Hohn wrote:

Another approach, though not always popular, is to expand groupings at a subgenus type level. For instance if a genus of flies has 200 species, but you can certainly narrow it down to 2 species, is there a way to track it as such. That's a tricky one from a database standpoint and proliferation standpoint (the two honeysuckles I can't tell apart in Vermont might not co-exist elsewhere, or perhaps there is a third species in Texas that also would be in the group there, etc). 

On Fri, Mar 30, 2018 at 8:35 AM, Chris Cheatle <cmch...@gmail.com> wrote:

A couple of comments on things above:

• I would stay away from any new "are you sure" dialog box. After about 3 instances of seeing it, it will simply become another box no one reads or considers before just clicking yes to make it go away
• one possible route to help here is increased discussion and awareness of the acceptability of genus or higher level records at research grade. I think there is a desire (or in some cases "need" in class assignments) to see records get this label. Scientifically, at least to be a solid confirmed genus level record showing occurrence at a place and time is more valuable that a "guess" among different species levels. I see this behaviour even among friends going to species level on their submissions when in areas that supposedly require microscopic evaluation to differentiate at species level strictly because they want the RG status. Either the record gets dubiously confirmed, or sits pending forever, when had it been done at genus or family level, it could have progressed.

Thanks.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Charlie Hohn]

Yeah, you can see my comment I just made in that observation... we made some 'holding bin' fields that help someone but don't factor into research grade.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Millie Basden]

I use the "No it's as good as it can be" button on genus level IDs of my own observations, if the person(s) who identify to genus are experts--e.g., for plant observations, if the ID is by the curator of botany at the local natural history museum or the person who wrote the treatment of the genus in the Jepson Manual (I'm in California). I hesitate to mark that on other's observations due to the sentiment that there may be someone someday who could improve the ID.  I know Charlie hates the idea of "expert identifiers" based on outside credentials, but I do wish their IDs had more weight than the IDs of those of us who are well-intentioned, have made lots of identifications, but are not experts. I may have been identifying the same plant incorrectly for a long time because I don't know what I don't know. Others in the same boat agree with my ID and the result is a research grade observation that shouldn't be. 

Millie

[Charlie Hohn]

It’s not that I hate it per se it’s that there was a long discussion about it and a lot of people didn’t want to do it and had some major concerns. I know it’s not a democracy and I know only a small number of inat users see the google group... but I saw the vast majority not being into the idea both on the board and those who contacted me or discussed it elsewhere. I won’t reiterate the issues with it because I’ve done so too many times already. 

A more timely question is how do we reduce people agreeing with ids when they shouldn’t ? 

Sent from Gmail Mobile

[jesse rorabaugh]

One idea that I have always liked is the ability to tag taxon pages with how difficult the ID is, and maybe a short description about what characteristics can be used to identify it. Perhaps using a system like:

Difficulty of Identification:
Easy: Often can be identified even in low quality photos.
Medium: Typically can be identified from good quality field photos.
Difficult: Can be identified from very high quality field photos of the proper characteristics.
Some word which means really really difficult: Can only be identified by examining microscopic details, genetic tests, or other lab work.

There would be some ambiguity but most species clearly fall in one of two of those categories.

[Charlie Hohn]

That would be amazing. 

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Jeremy Hussell]

I think I would prefer making extra community-derived taxon-specific knowledge visible on the site instead of a reputation or status system. I think the latter would inevitably lead to people gaming the system, while the former would help people on the site to learn to ID more carefully.

A couple of additional examples of species that are hard to distinguish:
 - Narrow-headed Sun Fly (Helophilus fasciatus) vs. H. latifrons (supposedly equally common, see e.g. http://www.canacoll.org/Diptera/Staff/Skevington/Syrphidae/Helophilus/Helophilus%201.jpg, but there are 222 observations of the former and 5 of the latter)
 - Greater Scaup vs. Lesser Scaup (part of a large genus, easily distinguishable from the other species, but no way to ID them as Scaup spp., only as Diving Ducks)
 - Grey Cross Spider (Larinioides sclopetarius) vs. L. patagiatus (an example where the community has done a good job of letting people know that they aren't easily distinguishable)

[John Bestevaar]

On Saturday, March 31, 2018 at 2:14:29 PM UTC+10, John Bestevaar wrote:

My first post here, so no idea what i am doing.
The web has made possible Inaturalist . But it also means that Information and knowledge must be restructured to bring forth the benefits for us.
In the time of books training experts and making books is the most efficient way to make knowledge available. That is no longer the case in the age of the web.
The proof to my mind is Wikipedia. Knowledge is constantly iterated and referenced to sources more so that relying on the pronouncements of experts.
So that is the way forward, Difficulty of ID by bad photo alone etc.Knowledge is no longer a body with an absolute truth value attached. Useful within the relevant context is the new description of knowledge and truth.
Cheers JohnB

[Robert Taylor]

Could we not have both the current RG and a reputation system at the same time? We could maintain the RG which clicks in when the second person adds their ID. But next to the first of each ID where it says 'improving' or 'maverick' could be a score. An identification from a novice will score 1, a score of 2 will be allocated to someone who has IDed more than 10 individuals incorporating 10% of the species in the genus/family (ie. not just identifying the same species multiple times), through to a score of 4 if the identifier is has IDed more than 300 individuals in the order/family covering more than 30% of the species or 40% of the genera. A taxonomist/museum curator of that order who may be a generalist in his field but has a fair knowlage of the taxonomy could also score a 4; whilst a score of 5 will be assigned if the person's life body of work has been in that family and he/she has an intermit knowlage of the taxonomy. This score need not be used to determine RG it just needs to be there so that researchers can get some idea of the value of the identification. Please note that the scores, percentages and taxonomic levels I have assigned to this example have been made up on the fly. The actual scores will have to vary between orders and families should be debated and decided upon after careful consideration. Point rewards might also be given to persons who consistently provide detailed motivation to their IDs showing that they have considered all the other options. One way to facilitate this would be to have a way in which the community can approve of the notes/comments given with the ID. When that person get 600 approvals they get a a +1 to their ID score. This will allow a knowledgeable hardworking layman to eventually be able to score the same an expert!

An observation scoring 2 might suggest that the observer is just agreeing with the first identifier. Whilst an observation scoring 5 is (ie 5 novice users or 1 expert) is less likely to be the result of the blind leading the blind and a score of 10 will be a solid score with little doubt in the ID.

The score should be searchable - as one does for observations that 'Needs ID' - one should be able to pull up all observations which score less than an imputed value.

On Thursday, 29 March 2018 22:54:28 UTC+2, Upupa epops wrote:

[Jeremy Hussell]

After thinking about this for a bit, one step that might help break this feedback loop would be to remove the "Agree" button. It's there as a shortcut for the most common case, when the existing ID is correct and your independent ID comes to the same conclusion. But new users to the site seem to be interpreting it as "I accept your ID" (i.e. "I don't disagree with you, therefore I must agree with you."), even when they don't know enough to make the same ID.

Brainstorming options:
 - Don't change anything.
 - Remove the button.
 - Reword the button, e.g. "Correct" or "Concur" or "I have independently reached the same ID" instead of "Agree".
 - Add a popup for users who have less than X IDs on the site, e.g. "I have independently reached the same ID" and "I don't disagree with your ID. Thanks for the help!", where the second button does nothing, or adds the "Thank you!" text as a comment, or redirects to the FAQ explaining how the "Agree" button is supposed to be used.

In deciding between these options we should consider:
 - convenience for people making IDs (How much difficulty will this add for people making IDs, and how much will that impact the site?)
 - too-quick agreements (What's the rate, and how much of a problem is that, really?)
 - amount of developer effort required (They should spend their time on the most useful thing they can do first.)
 - other constraints (For example, lengthening the button text might be impossible because of space constraints on the interface.)

My first instinct is to remove the button, and thus force people to spend a bit more time to agree with an ID. This would give people a few more seconds to consider the evidence, and thus a better chance to double-check the ID, at the cost of slightly slowing down the rate at which identifiers can blow through the obvious, easy-to-identify observations. Much of the value in iNaturalist is in the quality of the IDs. Naturally we want the best of all possible worlds, where we can get a correct ID instantly, but given the choice between an incorrect or misleading ID now and a correct ID later, I'd prefer correctness over promptness. (Up to a point. If observations are going to go unidentified for years at a time, I would be working to shift the balance the other direction.)

[Chris Cheatle]

If there is going to be a reputation system (and I personally believe there should not be one), I see no point in a "progression" system. It should be all or nothing, experts/professionals weighted strongly, amateurs not weighted at all. To use your example:

• I see no reason why my ability to correctly identify something is at one level when I have 100 done, and then is suddenly "better" because I do #101, and then again suddenly better again at #201 etc. Realistically I will have learned nothing about more accurately recognizing the species in the interim
• it infers that "amateur" knowledge begins at the time they join iNaturalist. As of writing, I have identified 42 Blue-winged Teals. But I've known how to identify a Blue-winged Teal for 25+ years, iNaturalist simply was not around to demonstrate that
• Forcing people into a system where they need to "level up" will simply lead folks to needlessly seek out records to agree with. If I or anyone need to find 200 records of a species and add agreements to them in order to be perceived as having a level of knowledge, all that will happen is run a search on records for that species and start hitting agree over an over. That time is of no value in my improving my skills, time I could spend on other areas or learning, and massively annoy user as they get all kinds of confirmation notes as the 22nd person agrees that my submission actually is a Blue-winged Teal in order for them to level up.
• if you make it only "first" agreements, then few will ever reach these levels as the mass of new records for almost all species is too low
• I dont know how the genus restriction would work. If I only ever want to id records from my area, and I somehow add identifications for 3,000 Belted Kingfisher records, I should get no "points" awarded on my contributions because I don't do ID's on the other 3 members of the genus ?

[Robert Taylor]

Obviously I am of the opinion that knowledgeable/expert contributions are valuable and should be able to sway identifications on iNat. But perhaps, as a compromise, no-one need ever know their score that they contribute. This could be hidden (again I would rather have it discreetly placed on the page in small writing) and the score could be shown when a researcher downloads the data - adding a lot more value than an arbitrary RG. Obviously if someone tried hard enough they could figure out their score by regularly downloading data, but so what? 

There is no-need for anyone to level-up as the scoring could be continuous, going up with every observation and every ID and I imagine that many people do learn more with every observation or ID that they post.

Unfortunately, I can't think of anything to compensate for your example of years of IDing Blue-winged teal before joining iNat, asides from people who are recognized by their peers (through the acceptance of scientific contributions) as knowledgeable or expert, as per my discussion above. Any suggestions?

Does it really matter when an observation reaches RG? as this is a arbitrary point. The only downfall I can see is that some people only ID observations on the "Needs ID" page and they would miss out on any observations reaching RG. But perhaps you could enlighten me as to other purposes of RG? I hope that all data goes to GBIF whether it is GR or not - it should be up to the end-user as to whether or not they use the data - and not the fault of an obscure genus with only one expert who is not recognized by iNat. Sorry I am new to these forums but I see some of this thread has been continued at: https://groups.google.com/forum/#!topic/inaturalist/PHr7VerDhUo. Following on from a suggestion in that thread but more relevant to this thread - could the user not input what level he/she would like to review observations to, eg. you could go to "Needs ID" and then select observations with none or 1 ID to view the current "Needs ID" page, but you could also look for all the observations with 4 or less identifications, basically allowing the user to decide what is RG or not. Perhaps doing away with RG completely. 

Doing away with RG will solve the problem mentioned in both this and the other thread.

[Patrick Alexander]

If I were using iNaturalist observations for research I would not pay much attention to the "research grade" marker. I would follow the same approach I do with herbarium specimens: I check the specimens myself; if I can't check the specimens myself, I rely on the IDs of people I know have the expertise to do it accurately; if I can't check it myself and there isn't a third-party ID I know is reliable, I don't use the specimen data. Just trusting the name on the sheet is probably fine in cases where there is very little chance of misID, but in general a bad idea that's going to give you error-filled data.

I would expect careful researchers to do the same. I think it would be difficult to automate this kind of evaluation in iNaturalist.

I do think acknowledgement of expertise in the iNaturalist ID system would be helpful for getting folks with expertise to engage on iNaturalist and stay here. It would still be difficult to figure out how to do so effectively, though, because it would need to take into account expertise gained and demonstrated outside iNaturalist. The kind of problem we'd be trying to address is along the lines of "world-renowned expert on genus X pops into iNaturalist on a lark and all her IDs get overriden by undergrads who only heard of the genus yesterday".

[Charlie Hohn]

Are there examples of that happening? It would be good to track and post them so we knew the extent of the issue. I haven’t seen anything other than the well-discussed student account issues. 

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Patrick Alexander]

Hello Charlie,

That exact scenario? Not that I know of. However, having observations with IDs from people who don't really know what they're doing mixed in with IDs from people who've worked with a genus for a long time, and having both get equal value in the iNaturalist community ID / research grade calculations, with the less knowledgeable IDs outweighing those of experts some of the time, is just the normal, routine functioning of the iNaturalist system.

Returning to my discussion of using observations for research--I guess where I'm going with that is that I don't really see the community ID and research grade markers as being all that relevant to, at least, the kind of research use of observations I imagine. I guess the caveat here is that I'm thinking of things primarily from a plant taxonomy point of view, so approaches to data may well vary among other groups. In any case, to me those aspects of iNaturalist are important internally to iNaturalist and not so much to external use for research. For instance, I wouldn't expect the ID feedback loop to be as relevant in this context; I'm assuming, for instance, a researcher working with observations on the genus Lucilia would know that it's a difficult genus where reliable ID from photos is probably not possible for most users, and would know not to trust those IDs whether "research grade" or not. So the feedback loop would not mislead such a researcher. However, other folks who are working with observations in iNaturalist and don't necessarily know the genus or work on it for research probably would be misled by a bunch of research grade observations.

For meta-analysis kind of stuff where researchers aren't in a position to evaluate individual observations (say someone is just wanting to map plant diversity by county or something), the community ID and research grade markers would become more relevant, but I think the basic take-home message for researchers in that position is, yeah, the data's going to be messy and there's going to be an substantial error rate, at least in some taxa. If you can't address that, you just accept that it exists and hope it doesn't lead you too far astray... probably the most useful thing iNaturalist could do in that context is be transparent about what the data means and what sources of error we know are going to have an effect.

If I wanted to pull observations related to distribution / habitat modelling of Agave, as one example that's been relevant to me recently, if the community ID is way off it's going to be hard for me to pull out the relevant observations. Assuming it's at least right to genus and I can pull the pile of them out, I'd be checking species-level ID myself and would not be very concerned about the community ID and research grade. At that point, the variables that would be of interest to me could be split up a few different ways in terms of specific fields, but might be along the lines of: yes/no, is the ID verifiable by me or through IDs of people I know are trustworthy; ID for each observation; source of ID for each observation (me or a particular other identifier). Right now, I would find it difficult to record and track those variables within iNaturalist. I think all of the needed functionality to do that exists, but most of it is available through "back door" interfaces that aren't really publicized or documented, and it wouldn't be difficult to figure out how to use effectively. The easiest thing for me to do would probably be to export observation information and work with it outside of iNaturalist in Excel or the like, but there would be some hiccups in doing that--I would probably be switching back and forth between looking at the media associated with the observation in iNaturalist, and annotating my external data separately. So, I guess part of the question when it comes to research use is whether we're envisioning iNaturalist as a source of data that will be used primarily externally or whether we're envisioning iNaturalist as a place where a lot of that work can be done directly. In the first case, I think the weak link is being able to export and work with the media externally. In the second case, I think the weak link is being able to work effectively with a pile of observations according to an identification workflow that does not correspond with the community ID. I think there would be benefits both to researchers and to iNaturalist in taking the second option. It's good for the researcher to have an efficient workflow, and in many ways the iNaturalist interface is much better than anything one could easily cobble together outside of iNaturalist; it's good for iNaturalist to have researchers directly engaged with and working with the platform, and to have some of their expertise captured on iNaturalist. I think we could get a lot of the way there by having a "project ID" or "my ID" kind of function... where, e.g., I could set it to display a set of observations according to my ID, or the ID provided by members of a list of trusted users, and set iNaturalist to use this "project ID" in the same way it uses "community ID" and otherwise work the same. The sticking point would be cases where my use of data requires taxonomy other than the iNaturalist taxonomy. This is part of the reason I like to harp on having iNaturalist support taxonomic flexibility and translation between alternative taxonomies so often.

Regards,
Patrick

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Charlie Hohn]

i do think you can track curator ID in a project.  So if you had an agave project, you could make yourself and other agave experts curators and go through and add IDs. I *think* you can then download them and sort by that ID instead of 'research grade' ID.

I do strongly believe iNat should be an internal meritocracy, meaning any reputation or ranking should come from what people do on the site. That being said there are probably creative solutions. I think in the case of amateurs a lot of time the issue is when their ID is challenged they don't get involved in the discussion. Fled users are a big issue in this case I think. Otherwise... most people are ecited about experts and don't seem to argue with them about ID that I've seen. For instance i was trying to help with some Ceanothus IDs then one of the people who wrote the keys for that species appeared and corrected a couple of my IDs.. I have to say it was great, but also I am now really  confused about Ceanothus crassifolius. But I removed my ID because, why not listen to someone who has so much expertise? 

I guess if we can do so without picking too much on specific users, it would be good to document actual issues. I feel like there is a lot of 'experts want this' or 'experts don't want that' or 'amateurs want this' and some of it doesn't seem to be actually happening on the site. And most of us are in little silos, i only look at plants in parts of North America for ID and me trying to do moth ID in Singapore or herp ID in Texas is absurd, I know nothing. So I don't know how some parts of the community act. 

Also I am concerned about trying to set up iNat as some sort of curator or ranker of external naturalist/taxonomy credit. How do we choose who is an expert? Sometimes it's obvious. But sometimes someone with a bunch of papers knows less than an accountant who is obsessed with the taxa. Sometimes there are two or more leading scientists in taxonomic 'feuds'. Some genera are massive, and even though i can key out most Vermont sedges there is no way I am a global sedge expert. Science is like any other human endeavor in that interpersonal interactions are complicated, drama can happen, and systemic bias against different groups of people or organisms can exist. By using only internal info, we avoid having to wade into all of that. Most experts I know are excited about their taxa and I'd imagine would soon have a high reputation in any system that worked at all. Perhaps some 'experts' are offended by having mear mortals and youngsters and people with less than 12 publications question them, but to be honest those are the sorts of people we can do without on iNat... and in science in general. Haven't we seen enough examples by now of horrific behavior among a few people who are let slide because they are 'experts'? Anyone with strong expertise, ability to know their actual strengths and weaknesses and the desire to be a member of the community will rise up as such in no time.

Anyhow... I think this is an important conversation but I also know I've posted an awful lot about it and don't want to dominate the conversation. So yeah... sorry.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

Sent from Gmail Mobile

--

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

[Upupa epops]

There are some groups of organisms for which I think the research grade observations will be nearly entirely accurately identified. For example with birds, most species aren't too difficult to separate with a good photo, and there are a ton of competent birders on iNaturalist (at least in North America, most of this probably doesn't apply elsewhere). There are definitely going to be a few misidentified observations in some of the harder to ID species pairs, but overall I assume the quality of RG observations is pretty good. And if an observation is Needs ID, chances are the bird is difficult or impossible to ID from the photos given (this is why I rarely ID birds even though I can; the few that others haven't already gotten to are really difficult).

I'm not a researcher but I feel like I could probably trust RG observations of birds on iNat without actually looking at them even more than eBird observations because most eBird observations don't have photos to be corrected on. And even the ones that do have photos aren't actively curated.

On the other hand, as far as I can tell, the quality of RG Lucilia observations is no different from the quality of Needs ID observations, which is basically useless. The species is likely wrong, the genus could very well be wrong, and some observations might not even be of blowflies.

I guess my hope for this discussion was to find a way to balance this discrepancy out. As far as I can tell, based on BugGuide, these flies can be identified to genus and sometimes species based on good photos. So the iNat community should theoretically be able to curate these observations to a point where there will eventually be only a few RG observations but all of them should have IDs that are reliable to a reasonable extent. But that still doesn't stop the problem from coming up again with a different obscure but common genus.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

Sent from Gmail Mobile

--

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Jeremy Hussell]

It seems to me that most of the worries expressed in this conversation are about ID quality. From one side of things, researchers who want to use the data generated by iNaturalist wish that more of the ID work was done to a higher standard of quality. From the other, the community of people posting photos to iNaturalist and learning from the site wish for pretty much the same thing.

So what's the cause of the low quality of IDs? A lot of it seems to be due to information about how to ID species in an area not being easy enough to find. A lot of it is locked up in obscure papers, keys aimed at researchers who have specimens right in front of them instead of photographs, and yes, in the heads of experts. All the suggestions about reputation systems seem to be aimed at simply identifying the few existing experts and highlighting their contributions. I think a much more effective approach in the long term would be to make iNaturalist an easy place to make and/or find a guide to identifying any particular taxa from photographs. For example, by having a wiki-page associated with each taxon, where anyone can see the collected wisdom about what the known differentiating field marks are, how reliable they are, the areas where the ranges of similar species overlap, etc., learn from it, and add to it.

The birder community is a good example of what we should be aiming for in all taxa: a large, active, knowledgeable community. The only way to get there is to educate more people who are interested. Simply identifying experts won't be enough, because there aren't enough of them. We need to create them.

Bugguide.net is also a good example: they have community-made and maintained guides for all taxa, which shows it can be done. Right now a lot of people in N. America get better results by looking at bugguide.net to identify insects than by looking at iNaturalist. But it's limited to only insects, and only in North America, and it lacks range maps (or has them and I just haven't found them yet; the site's not as user-friendly as iNaturalist). iNaturalist should be aiming to be as useful as bugguide.net, but in all taxa, global in scope, and easier to use.

[Robert Taylor]

I think Jeremy has a good point. Does iNat currently have a place to host keys and the like which can be linked to on the order, family and genus pages?

[Chris Cheatle]

I know the admin team seems to have mixed feelings about them, but is the "wiki" functionality not really available via guides ?

Guides allow you to create any section or content. Good ones like this example http://inaturalist.ca/guides/2669 already are being set up. 

If a good quality guide page exists, or can be created, can you not just add a link to it under the "More Info" area of the About tab on a taxa page ?

On Thursday, March 29, 2018 at 4:54:28 PM UTC-4, Upupa epops wrote:

Hi everyone,

Recently I came across a problem with bad research grade observations that I think has a pattern that could possibly be solved. Sometimes there are many research grades observations of a species which I don't know if there's anyone on iNaturalist who knows how to competently separate it from similar species. I think this is caused by a feedback loop with several different aspects.

It starts by people identifying based on what looks similar, without investigating potential similar species. For example, people will look on Google Images or an identifying website, find a species that resembles their photo, and identify it to that species. At this stage, when they're only incautiously identifying their own observations, it's not a big deal because someone will usually correct them. But if not, eventually they develop the confidence to identify that species on other peoples' observations, and that's when the next step comes in.

We're all aware of a tendency on iNaturalist to agree with an identifier without double checking. If I go through and identify a number of observation to a species that is hard to ID (several similar species, obscure distinguishing marks, etc.), a good portion of the observers will agree with me without comment. Then the observation will become research grade. I highly doubt most of them double checked the ID using external sources. This is fine if my ID is correct, but I'm liable to make mistakes, and people are unlikely to distinguish between an authoritative identifier and someone who isn't. So if someone is determining their ID's based on the previously mentioned method, there will now be unreliable research grade observations.

After a certain number of identifications on others' observations, other people will start to think that this is the correct ID and use it themselves on other' observations. If this is a common species and these things start happening, pretty soon there will be enough RG observations to show up in the AI suggestions. Evidently a lot of people will base initial IDs solely on AI suggestions, which encourages the pattern even more.

Here are a couple examples:

[Charlie Hohn]

guides are only editable by one person as far as I know. They are also not being maintained any more, they are a depreciated feature in my understanding. And they don't connect well to individual taxa.

I think adding a wiki to species pages is at least worth trying. No more open to abuse than anything else and if it's a problem we can shut them off. But maybe there is some other dev reason not to do it.

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Jeremy Hussell]

I'm not very familiar with the guides, not having made one myself, but I watched the introductory tutorial videos this morning, and there are some substantial differences compared to a wiki:
 - guides are edited by one user, while a wiki page can be edited by anyone (guide editors can manually give permission to other users to edit their guide, while a wiki can be limited to logged-in users or some subset of users)
 - guides don't have an associated 'Talk' page where discussion about changes to the guide can happen and be recorded
 - guides don't keep a record of all previous versions of the guide (if the someone deletes something, it's gone)
 - guides seem to have a lot of specialized sorting, searching, categorization, and map widgits which most wikis don't have (but could be kept)
 - guides seem to have a lot of specialized editing tools which most wikis don't have (but could be kept)

The 'About' tab of a taxon page currently seems to load information from other sites (Wikipedia, mostly), and link to a few other sources (how do sources get added?). I'd like to see the Wikipedia articles turned into another link in the 'More Info' section, and replaced by a wiki or guide that's editable by any iNaturalist user. The simplest possible step forward would probably just set up a wiki page for every taxa (so, a wiki page with a Talk tab, and read/edit/history modes for both tabs). Maybe using MediaWiki (the wiki software behind Wikipedia), maybe using some other wiki software. That wouldn't have any of the tools available to the guides, but it would require a lot less developer effort to create. The initial page could be blank, or could use a default template. Maybe just create the same list of links from the 'More Info' section by default, and get rid of the 'More Info' section so that there's more room on the interface. If the developers get enthusiastic, they could see about recreating or integrating some of the tools available for guides.

[Jeremy Hussell]

It does seem like taxon pages could easily have links to all published guides which cover the taxon, maybe filtered by location.

[Chris Cheatle]

Jeremy,

The Wikipedia link is automatically generated on all taxa pages. It actually works by looking for a corresponding Wikipedia article using the latin name. There is no process needed to add it.

I did not realize the 1 person default editor limitation on guides, although as you note, multiple editors can be configured. 

I guess the point was the technology to implement much of this is already in place, and funded via an EOL grant. Leveraging it might be faster, cheaper and easier than a net new approach.

[Charlie Hohn]

the 'best case' scenario I see is a mix of field notes and ID guide, curated by the community, that integrates with the ID and range mapping process. Maybe it's too much a diversion from the main site to be a dev priority, and certainly there is a LOT of work for the small team to deal with. It certainly follows the iNat directive to connect people with nature, though. 

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Patrick Alexander]

It's not clear to me why iNaturalist would want to recreate Wikipedia.

On 4/3/18, 8:06 AM, Jeremy Hussell wrote:

The 'About' tab of a taxon page currently seems to load information from other sites (Wikipedia, mostly), and link to a few other sources (how do sources get added?). I'd like to see the Wikipedia articles turned into another link in the 'More Info' section, and replaced by a wiki or guide that's editable by any iNaturalist user. The simplest possible step forward would probably just set up a wiki page for every taxa (so, a wiki page with a Talk tab, and read/edit/history modes for both tabs). Maybe using MediaWiki (the wiki software behind Wikipedia), maybe using some other wiki software. That wouldn't have any of the tools available to the guides, but it would require a lot less developer effort to create. The initial page could be blank, or could use a default template. Maybe just create the same list of links from the 'More Info' section by default, and get rid of the 'More Info' section so that there's more room on the interface. If the developers get enthusiastic, they could see about recreating or integrating some of the tools available for guides.

--

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Charlie Hohn]

Wikipedia isn't a place to create field guides. it's an encyclopedia. The file format is called a 'wiki' but otherwise it would be very different. Wikipedia doesn't have ID tips, pictures of diagnostic features, or links directly to the iNat community and observations. I'm sure not going to start using Wikipedia for any of this either, nor do they want me to.

On Tue, Apr 3, 2018 at 10:51 AM, Patrick Alexander <paal...@polyploid.net> wrote:

It's not clear to me why iNaturalist would want to recreate Wikipedia.

On 4/3/18, 8:06 AM, Jeremy Hussell wrote:

The 'About' tab of a taxon page currently seems to load information from other sites (Wikipedia, mostly), and link to a few other sources (how do sources get added?). I'd like to see the Wikipedia articles turned into another link in the 'More Info' section, and replaced by a wiki or guide that's editable by any iNaturalist user. The simplest possible step forward would probably just set up a wiki page for every taxa (so, a wiki page with a Talk tab, and read/edit/history modes for both tabs). Maybe using MediaWiki (the wiki software behind Wikipedia), maybe using some other wiki software. That wouldn't have any of the tools available to the guides, but it would require a lot less developer effort to create. The initial page could be blank, or could use a default template. Maybe just create the same list of links from the 'More Info' section by default, and get rid of the 'More Info' section so that there's more room on the interface. If the developers get enthusiastic, they could see about recreating or integrating some of the tools available for guides.

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[espam...@gmail.com]

The advantage of the BugGuide info pages in part is that the software allows photographs submitted as observations to be easily reproduced and linked; finding non-restricted images is the hardest part of making a guide

[Colin Purrington]

Removing the Agree button would probably reduce many problems by an order of magnitude. It's simply just too easy to agree (1 click) compared to doing anything else. But I suspect that's a non-starter.

Would it be possible to display the images on "Other Species Commonly Misidentified As This Species" upon first identification? That might prevent the observer from leaping to a species ID initially. Similarly, it would be great to display the "Other Species Commonly Misidentified As This Species" information under the Agree button. Having these warnings in your face at these critical moments would likely reduce some of these feedback loops. Currently, I think, one needs to navigate to the "Other Species Commonly Misidentified As This Species" ... and I suspect that is not always happening. 

[Charlie Hohn]

it would also reduce the number of identifications by an order of magnitude unless there's some way around it. It's going to cut down my IDs by at least 50%, probably much more, if i have to type in the species every time I agree.

I'm not perfect and i do make mistakes but I usually have a good idea of how confident I am and if i have an idea, but am not sure, i add a comment instead of an ID. I think this is true for most established identifiers. Maybe you could have it so that newbies have such popups and they go away later.. though I'm not sure how best to deal with that.

On Tue, Apr 3, 2018 at 11:58 AM, Colin Purrington <colin.pu...@gmail.com> wrote:

Removing the Agree button would probably reduce many problems by an order of magnitude. It's simply just too easy to agree (1 click) compared to doing anything else. But I suspect that's a non-starter.

Would it be possible to display the images on "Other Species Commonly Misidentified As This Species" upon first identification? That might prevent the observer from leaping to a species ID initially. Similarly, it would be great to display the "Other Species Commonly Misidentified As This Species" information under the Agree button. Having these warnings in your face at these critical moments would likely reduce some of these feedback loops. Currently, I think, one needs to navigate to the "Other Species Commonly Misidentified As This Species" ... and I suspect that is not always happening. 

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Russell P]

I think Jesse's idea of a rating system for difficulty of identification should be explored. The central objective underlying iNat is identification. If folks had a way of rating ID difficulty for a species, and if that rating appeared VISIBLY next to the species name, that might be of huge value. An icon like stars or chili peppers or puzzle pieces or little microscopes could be used with the number of icons (1-4) reflecting difficulty. One could click on these icons to see how many people had voted on each of the four categories (like Amazon). Following the system used at Amazon, this icon would also link to comments about identification (analogous to reviews) and ability to post questions about the species (to which people could respond). 

On Friday, March 30, 2018 at 1:06:08 PM UTC-5, jesse rorabaugh wrote:

One idea that I have always liked is the ability to tag taxon pages with how difficult the ID is, and maybe a short description about what characteristics can be used to identify it. Perhaps using a system like:

Difficulty of Identification:
Easy: Often can be identified even in low quality photos.
Medium: Typically can be identified from good quality field photos.
Difficult: Can be identified from very high quality field photos of the proper characteristics.
Some word which means really really difficult: Can only be identified by examining microscopic details, genetic tests, or other lab work.

There would be some ambiguity but most species clearly fall in one of two of those categories.

[paloma]

I agree that iNat needs something to help people learn as they use it. I'm an amateur who has used iNaturalist regularly for about four years, and to me the biggest shortcoming of iNaturalist is the difficulty in finding out how to distinguish among similar species that occur in a given area, or even whether they can be distinguished by photographs. This information may be scattered in numerous discussions in multiple observations, but it is not easily found, and I think the taxon pages would be the logical gateway to this information,

[Patrick Alexander]

One other thought about workflow on providing IDs: "Agree" might be too easy and removing it might cut down on knee-jerk agreement but, yeah, it would also cut down on the number of useful IDs as Charlie points out, perhaps dramatically so. If we were to ditch the "agree" button, I think we'd need a good way of substantially reducing or eliminating the of time it takes for iNaturalist to look up names in the taxonomic database. As it is, I typically find myself typing in a name and then waiting a few seconds for the look-up to run. Sometimes the look-up stalls for whatever reason and I have to re-enter the name or reload the observation. Personally, I'd be fine with having to type in the name for all identifications... but not if it means more time just staring at the screen and waiting for the look-up to run.

On 4/3/18 9:58 AM, Colin Purrington wrote:

Removing the Agree button would probably reduce many problems by an order of magnitude. It's simply just too easy to agree (1 click) compared to doing anything else. But I suspect that's a non-starter.

Would it be possible to display the images on "Other Species Commonly Misidentified As This Species" upon first identification? That might prevent the observer from leaping to a species ID initially. Similarly, it would be great to display the "Other Species Commonly Misidentified As This Species" information under the Agree button. Having these warnings in your face at these critical moments would likely reduce some of these feedback loops. Currently, I think, one needs to navigate to the "Other Species Commonly Misidentified As This Species" ... and I suspect that is not always happening. 

--

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Colin Purrington]

Maybe there should be a "Disagree..." button. The "..." would be needed to invoke the options therein. You'd need to make at least one of the disagreement options super friendly, something a novice or child could use (i.e., current options when disagreeing are way too confusing). Most people will still be drawn to Agree just to be polite and because they assume previous IDs correct, but perhaps the error rate could be reduced somewhat.

[Patrick Alexander]

Understood, but my feeling is that the "place to stick information about species online" niche is already well-filled. For instance, if I google "Larrea tridentata", within the first 30 results I'm getting 13 different websites that provide descriptions / ID information / ecology or distribution information / photographs for plants. They are listed below. Each one of these has its own particular goals and limitations, of course, and none align entirely with the goals of iNaturalist. When I'm seeing a list of 13 different websites that are all doing more or less the same thing in this case, though, my first thought is not that we really need a 14th.

Wikipedia
The USDA PLANTS Database
Lady Bird Johnson Wildflower center
The USFS Fire Effects Information System
SEINet
Texas Native Plants Database
Flora of North America
PFAF Plant Database
desertusa.com
Vascular Plants of the Gila Wilderness
CalFlora
USA National Phenology Network
Encyclopedia of Life

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Charlie Hohn]

i don't know of a site where an interactive field guide type system is being created and maintained by a naturalist community at all. Do you? Tell me, I want to go join it. What i see is existing field guides and floras and such, static online (still useful), Wikipedia, which is totally different, and iNaturalist.  I guess you are envisioning something different than I am... and maybe I'm explaining wrong... but it's drifting pretty off topic so i should just let it be i guess.

[Upupa epops]

There are certain groups for which there is a lot of information available to naturalists and similar people who want to ID organisms in those groups. I think plants (except maybe mosses and grasses and similar things) are one of those groups. Also mammals, birds, dragonflies, butterflies, moths, reptiles... probably more I can't think of. For all of these there are plenty of field guides available and a lot of people are interested in them as a hobby.

But there are also a lot of groups for which there is not much information, mostly small obscure things that people don't usually notice I think.

I have been trying to figure out how to identify blow flies from photos a bit. I've tried searching in various ways on Google for help, but all I get are websites about eliminating pest blow flies, how interesting specific species are in the medical field, and how to key them with a microscope, none of which is particularly helpful. However I know it's possible because they identify them with photos on BugGuide.

I'm having similar difficulty with certain groups of hoverflies. There is knowledge out there but I cannot find it.

BugGuide does sort of have a wiki system like what is being described here, but from my perspective it is not very successful. There is lots of valuable information on BugGuide but you have to search for it (literally looking through the comments on every single observation), it's not usually consolidated into the info pages as it should be. 

I'm not sure entirely why it hasn't worked, it may be that only curators are able to edit it or that there aren't enough active and engaged curators. It could also be that they have the mentality described by Jeremy of taking advantage of the experts that already exist but not trying to make new experts.

(also BugGuide does more than just insects, I think it does all arthropods, and you can get maps by going to the "Data" tab)

I don't know if a wiki type thing is something iNat can do, but I think it could be an amazing achievement and super helpful if there were enough people interested in actively contributing to it.

Although I don't know if it would actually help with the topic of this thread since a wiki page would probably be just as easy to ignore as the "Similar Species" tab.

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

============================
Charlie Hohn
Montpelier, Vermont

--
You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.

To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist...@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Russell P]

Back to Jesse's rating system for level of difficulty making an ID, here's a great icon to use!

http://robpizzolato.com/images/editstock/icons/difficulty_meter.jpg

The more I think about it, the more I think this rating system would be a useful feature. Especially if combined with ability to have discussions on the species page.

As for community managed ID guides, I think it would be challenging to implement a system where the community would create and curate ID guides. The best guides show annotated images. I'm not seeing how a community could create this sort of thing. The community could, however, easily create and curate a list of resources for species, genera, etc. Literature cited for obscure literature (too bad much of it is copyrighted; but at least we'd know it exists) and links to guides that others create--such as this one that I'm experimenting with: https://www.inaturalist.org/journal/pfau_tarleton/15022

-RP

[John Bestevaar]

I assumed that uploading my photos to Inaturalist automatically meant
that my pictures could be used by whomever for whatever so long as it
was not for commercial gain in the market? Is that not the case? If not
then policy needs to change pronto. Maybe a small name recognition in
the bottom right corner would help convince people?
Regards JohnB

[John Bestevaar]

Yes..agree , as a non expert that makes sense to me.
Cheers JohnB

On 04/04/18 01:58, Colin Purrington wrote:

Removing the Agree button would probably reduce many problems by an order of magnitude. It's simply just too easy to agree (1 click) compared to doing anything else. But I suspect that's a non-starter.

Would it be possible to display the images on "Other Species Commonly Misidentified As This Species" upon first identification? That might prevent the observer from leaping to a species ID initially. Similarly, it would be great to display the "Other Species Commonly Misidentified As This Species" information under the Agree button. Having these warnings in your face at these critical moments would likely reduce some of these feedback loops. Currently, I think, one needs to navigate to the "Other Species Commonly Misidentified As This Species" ... and I suspect that is not always happening. 

[Charlie Hohn]

people can set their own terms of use.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[John Bestevaar]

Maybe have an "agree" button on each of the candidates that make it to the list of confusing species. Logic would be exclusive or?
Cheers JohnB

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[John Bestevaar]

Yep agree. I am layperson too interested very much mostly in my own backyard of Tamborine mountain region.
Happy to create a local group of observers and provide the data for INaturalist purposes but we want in return a good social user interface like the way FaceBook groups does it. 
Regards JohnB

--

[John Bestevaar]

Yep agree.... I suggest its a case of humans just love to be in control of of their own little patch and get the recognition as well.
So if INaturalist can create the user interface that works with that.....? Nobody likes to be treated as just a cog or a second class person.
Cheers JohnB

[John Bestevaar]

Charlie. I think the need is there but the solution not well implemented
I am member of several FB groups that deal with interests in common in a defined geographical area. Posting a picture/pictures  of an unknown plant or animal observed in that area usually gets you an ID within 24 hrs. Works well and social  but the data is not useful other than to the person posting. This is why i stopped developing my www.Earthtracker.net any further. And also because AI like Plantsnap will soon leave the static websites behind. And also cause i am no computer nerd.
Cheers JohnB

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[John Bestevaar]

So that tells me the resources contributed by people has myriad  of restrictions on the uses to which Inaturalist might like to put them to. No wonder you are having problems achieving any of your objectives.
Cheers JohnB

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Charlie Hohn]

do you really think PlantSnap will go anywhere? i haven't tried it because it costs money and they seem to be spending it all on annoying ads. But i find it hard to believe it will be better than the free inat/seek algorithm. Then again you never know. The Internet isn't always merit based.

On Wed, Apr 4, 2018 at 2:29 AM, John Bestevaar <phantom...@gmail.com> wrote:

So that tells me the resources contributed by people has myriad  of restrictions on the uses to which Inaturalist might like to put them to. No wonder you are having problems achieving any of your objectives.
Cheers JohnB

On 04/04/18 12:53, Charlie Hohn wrote:

people can set their own terms of use.

On Tue, Apr 3, 2018 at 10:45 PM, John Bestevaar <phantom...@gmail.com> wrote:

I assumed that uploading my photos to Inaturalist automatically meant that my pictures could be used by whomever for whatever so long as it was not for commercial gain in the market? Is that not the case? If not then policy needs to change pronto. Maybe a small name recognition in the bottom right corner would help convince people?
Regards JohnB

On 04/04/18 01:50, espam...@gmail.com wrote:

The advantage of the BugGuide info pages in part is that the software allows photographs submitted as observations to be easily reproduced and linked; finding non-restricted images is the hardest part of making a guide

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--

============================
Charlie Hohn
Montpelier, Vermont

--

You received this message because you are subscribed to a topic in the Google Groups "iNaturalist" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/inaturalist/WnjvyvoZQCs/unsubscribe.

To unsubscribe from this group and all its topics, send an email to inaturalist+unsubscribe@googlegroups.com.

To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "iNaturalist" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inaturalist+unsubscribe@googlegroups.com.
To post to this group, send email to inatu...@googlegroups.com.
Visit this group at https://groups.google.com/group/inaturalist.
For more options, visit https://groups.google.com/d/optout.

[Charlie Hohn]

also i don't think INat is having problems reaching its objectives at all. Its grown exponentially and has an amazing community and keeps adding new features. Sometimes issues come up with licensing but it seems rare and who wants to join a community where you lose all rights over your data? if it isn't meeting your objectives you could always try a different site.

[John Bestevaar]

Yes i have Plantsnap and paid the 4 bucks it costs. Like you say its not merit that wins the internet, its MARKETING.
My OS is Ubuntu cause i hate Apple trying to fence me in and also hate Windows cause the system is crap and bloated.
But i do use Facebook because the interface is so good. I hate twitter cause your train of thought can be no longer than 140 characters. I think i am not alone in my profile but the mainstream will be looking for Plantsnap i am certain.
Cheers JohnB

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Charlie Hohn]

Interesting, i don't know anyone who uses it. How well does it work? It's not that the price is too steep compared to what it offers, if it works, it's just that it seems to be a clone of some iNat features without the actual data collection. To each their own I suppose. 

[John Bestevaar]

Yes the American deal" take it or leave". Your rights over your data are entirely a figment of your imagination. As soon as you put anything out on the web it belongs to whoever has the most power and money. But it is soothing to the users to say that I naturalist would like to use the data posted to improve the website for all but not actually own it. So the originator can still do whatever with their data except stop Inaturalist from using it.
Cheers JohnB

To unsubscribe from this group and all its topics, send an email to inaturalist...@googlegroups.com.

[Charlie Hohn]

I think most people here are well aware of that.  But I don't think it's an 'american deal' for it to be rude to come into a new community where you don't need to be and complain and criticize in a condescending manner. I am guessing that is considered rude in most cultures. If you have constructive feedback it is welcome but if you don't like iNaturalist and think it is a failing website, why would you participate in it? Anyhow I am seeking to avoid arguing on this forum so I will leave it at that. I hope you find the functionality you are looking for either here or elsewhere because we all need to collect as much biodiversity data as we can.