Looked like frog eggs, right size and colour but not in a pond and no pond nearby in the woodland.
No interactions present.
...this was already linked because the panel was complete - no need to move all our agreements!!
Unfortunately just because the panel is 'complete' with links to eol etc does not mean that it is fully linked to the species dictionary and if it is not linked then many of the other functions on ispot do not work with that observation and neither do the statistics I am currently compiling about the ispot data.
Note some of the functions may appear to work with just the 'common frog' but behind the scenes others don't. There are a few thousand observations that are not linked (out of over half a million) so it is not a big issue but annoying as these often get missed out and they may be important observations.
OK then, I've completed it, by making it Likely - such power! Such an old post!
"There are a few thousand observations that are not linked" - is there a way to tell which, when the panel seems complete?
One way, which may not be fully reliable, is to look at the format of the common and scientific name. Common name in normal text and scientific name in italic.
If you are very keen then I could probably get a listing of the unlinked ones. I have being doing some links myself over the years but it takes ages so only when waiting for other analysis to complete.
Here are a bunch of examples if you want to have a look, in some cases the proper linked ID is there but the unlinked one has the likely ID
Some parts of the site work even with the unlinked name but not everything and it is more difficult to deal with if we want to export the data to schemes and socs and work out statistics.
I think I need to know more. I've looked at those.
Take this one http://www.ispotnature.org/node/7103. It has the correct configuration "Common name in normal text and scientific name in italic". There is nothing wrong with the first ID, except it does not carry the Scientific name in the Poster's ID (Header) (say from the Drop-down).
Another added ID is hardly going to make the iSpot Computer happy because it will never become Likely.
In those cases where there are agreements but only the common name in the Header, adding another ID will surely put the post in the doldrums - yes?
I think I need a small explanation, if you can bear it, of what constitutes "... linked(ing) to the species dictionary".
If one clicks on Red Admiral in the first ID (of my example) it leads to the UK Species Dictionary and lists more than 20 pages of correctly identified and designated Vanessa atalanta posts - each (though I haven't checked them all) has other observations and a complete panel AND refers to the UKSI at a click.
I understand what Tony Rebelo is doing - getting the Post to refer to SANBI and NOT CoL for all ZA posts. But I do not understand the problem you refer to.
I am coming to believe that effective Curation is needed here, that every post is scrutinised by one of a big team of lay-curators; a note left for the poster about protocol, a correction made (by the lay curator) and move to the next.
If you have time, go here http://www.geograph.org.uk/profile/60128. Even my spelling mistakes raise a query from a moderator and a failed hyperlink gets nailed quite quickly. There are thousands of daily posts and a good number of moderators (Lay curators). It is a very smooth site.
No hurry to respond here Mike - there's only the two of us!
Well I am a bit hazy on this too but I do know what effect it has on the output data (and it is why I was able to give you that list). The link via common name is not as good as linking via scientific name as there are various problems with this method, it is more approximate and it can cause species to link incorrectly in some cases. This can be even more of an issue when there are multiple species dictionaries not just the UK one.
In some cases there is little point in trying to shift over to making the correct link due to the weight of agreements on the unlinked ID but in other cases it can be done fairly easily.
By the way, off topic, have you seen this http://www.open.edu/openlearnworks/ it allows anyone to potentially create their own courses e.g. on identification. I have not used it but it was recommended to me recently by someone who said a number of people were trying it. I was going to post it in one of the forums where people were discussing identification resources and thinking about producing some themselves but I can't find exactly where and I think you were one of the contributors so perhaps your memory is better than mine!
Lat/Lng: 51.820857, -0.586661
OS grid ref: SP975145