Dictionary & Browser Issues

we need a firmer record of Dictionary and Browser issues here
Here is a link to the old thread Review of plant taxonomy (Britain and Ireland dictionary)
Dictionary anomalies
Browser - 2 entries for P.vulgare, two different sets of Other Observations

1 Like

Thanks for posting this, dejay.

False trail https://www.ispotnature.org/communities/uk-and-ireland/species-dictionary/NHMSYS0020536309/tubiferaceae
Lycoperdon is a FUNGI Genus but there is a Slime mould called Reticularia lycoperdon
One mistaken Observation here https://www.ispotnature.org/communities/uk-and-ireland/view/observation/852997/lycoperdon-species-perhaps

I put Pinus cembra in the Search box…. & got a genus of SA Proteaceae.
That was a surprise. I repeated several times in case I had made a mistake . Same result (which is what you get with computers).

That is interesting, the search is ‘powered by Google’ or whatever the relevant words are and it is hopeless as everyone says. It is difficult to find why it is so bad as in theory it should work very well but Googles search in general now turns up things based on a whole variety of unknown factors instead of just what is there (this is a general comment about Google search not just how it is implemented on ispot).

Yes, interesting

If we take “Pinus cembra iSpot” to Google there appears to be NO occurrences (even though it’s in the browser)
Unlike say “Larch Angels iSpot”
For years we have been ‘complaining’ about the way iSpot responds to searches.
It generally does not do very well, preferring to err towards ZA community and often finding occurrences in impossible lists (Filters or Collections)
It should be easy to find all occurrences of a Species name or unique occurrence (like albino squirrel) - it isn’t
I always use the Species browser for such things and so found this https://www.ispotnature.org/communities/global/view/observation/752362/arolla-pine-cone
There is only one, with only one agree-er and commenter
If you search iSpot (from the UK Community) for Arolla Pine you get two Global results. Neither of them have the words Arolla Pine in the text or titles.
There is no trouble searching Google for the same, so we ought to assume that iSpot has a faulty indexing system.
And yet try searching iSpot for Code Runner from any community, it does incredibly well.
But try finding my Abies koreana

Neither DuckDuckGo, nor Google, nor Bing, nor Yahoo, find any occurrences of Pinus cembra on iSpot. So they look for occurrences of less that all of Pinus cembra site:ispotnature.org. Google finds some Serruria images, while the others find occurrences of other pines. If you use “cembra” or “Pinus cembra”, all come up blank, though Yahoo still serves a few ads.

I’ve noticed that putting in infrequent terms to narrow searches works less well with search engines that it used to - they double guess you, and serve frequently accessed pages that match less well than infrequently accessed pages that match better.

But I do wonder whether the search engines aren’t properly trawling iSpot. I’d check out robots.txt, or if bits of the site are disconnect by nofollow attributes.

The other question is whether everything on the site is trawlable. Observations scroll of the main carousels fairly quickly, and might not be registered by the search engines before they do so. Presumably they could also be found via other observations, but the search engines have to know about one observation of the taxon beforehand, and that still doesn’t give access to all observations (or frequently observed taxa). The search engines can’t get in via Your iSpot, but I would have thought they could have followed observer/identifier/commenter names in to Activity Trackers. And presumably they can also get in via the dictionary.

A different approach would be to look at the server logs and see what the search engines are actually reading.

we have been trying to get Google search to work properly since it was put into the new site. The best it did was to get (rough guess) about 70% of the things it should do when there was a huge effort to pursuade it to do the right thing - programmers doing everything that was asked repeatedly and constantly submitting the right things to the right places in google. But as soon as you reduce the constant fiddling then it looses most of the things it should have but still has some spam that it should have cleared out years ago and which we have repeatedly asked google to clear out.
Basically I suspect the only way to have proper search on site is to produce your own. We don’t have time to do this at the moment unless someone wants to help with this.

I cannot pretend to understand how a Search mechanism works. But I do know that iSpot HAS a search mechanism that works very well. It locates and orders EVERY comment, D and agreement I have made since 2013 - honest, every one and for every person, many since 2009.
Locate the code that does that and build a query sequence - if that’s what it’s called
I imagine, in my ignorance, that any 2nd year undergraduate Computer Science student could do it.

We actually submit URLs to Google on a regular basis. They are sent the same pages as a user would see.

This is a different thing, this is indeed using the internal database and code and not relying on Google at all which is why it works. It would be possible to produce a search like this for everything but more difficult (and slow) especially if more complex queries are tried which is why we have not built it yet.

Well, yesterday I took some photos of a Pinus cembra, so I guess I should upload them as an ispot post.

Now posted here https://www.ispotnature.org/communities/global/view/observation/855913/pinus-cembra
I wonder when it will show up in the search……

mistakenly typing Puffball into the Sci. ID box gets Lycoperdon - NOT the Puffball one, the Slime mould version

Dictionary anomaly. Echinacea. Plant versus Invertebrate*

This Echinacea plant is in invertebrates albeit the Wikipedia link is for the plant.

The web wasn’t particularly enthusiastic about clarifying this, but it turns out that Echinacea is a superorder of sea urchins (class Echinoidea). The WikiPedia article for this is Echinacea (animal).

Liriope seems to be a plant and an invertebrate. This is the plant but the Likely Banner is for the Invertebrate and needs fixing. I’m not sure what the identification should be?

Sorry, I had not seen this post, and shifted the above to Asparagaceae, as being the only category I could find in the wiki entry, which matched one in the dictionary. I was just so frustrated at looking at every hydroid observation to find a plant. I have been trying to agree or make a sensible comment on Skeleton on a shell | Observation | UK and Ireland | iSpot Nature, but like so many things these have passed me by. Out of my depth on these!

We MUST record dictionary issues when and as we find them, otherwise no-one will be wiser
An update of the dictionary is due in March - who will go through some of these to see if they have been fixed - probably NOT me.
Two occurrences of Pinopsida in the dropdown (and Browser) the second one is invalid
And who needs Conifer seedlings/saplings as a scientific ID?

Why are there no observations for Conifer-seedlings-saplings

2 versions (Spellings); one wrong but has a locked observation
Needs another agreement…

They both appear as separate entries in the NBN atlas:

So its NBN who would have to correct.