Quick-find shows multiple of the same species.

[MatsP]

When a species has synonyms that only change in genus (e.g. Tatia/Centromochlus romani), the species is shown twice. I think it should only show UNIQUE species.

For example, entering "Syndontis eupt" shows two items: S. euptera and S. eupterus.

To me, this could be quite confusing to users.

--
Mats

[Suckermouth]

Alternatively, synonym results could be somehow marked to say what they are a synonym of.

[MatsP]

Alternatively, synonym results could be somehow marked to say what they are a synonym of.

Indeed, synonyms listed under a single picture along with the correct species name would work too - but it may be difficult to indicate clearly that they are synonyms.

--
Mats

[Jools]

Have a think about this guys, are you sure? I mean, put yourself in the shoes (or mind) of someone who's only beginning to use scientific names. If I have a go at spelling it, and I get a result (or two) that gets me to the fish I know I want, then what's the issue?

The suggestion that we have search name and THE name shown is good in theory, but in practice it looks ugly and I would suggest is confusing.

It is a search, not an index. Its purpose is to quickly find a species. Happy to discuss, and very open to suggestions, but right now I think it's good.

Jools

[Jools]

A really good example is Mats' one but search for "roman" and miss out the final letter i. What I could do is put (synonym) or (common name) on the line below? I would be keen to avoid using the full current name on each entry.

What do you think?

Jools

[MatsP]

So, what you are saying is that someone may know that they have "Tatia romani", and they type in "roman" in the box, and up pops "Centromochlus romani", but not "Tatia romani"? Perhaps we can have soemthing like:
<picture>
Centromochlus romani
(Tatia romani)

--
Mats

[Jools]

Sorry, no, my advice was wrong. Type in romani WITH the i and you get the three results and the set of data I was suggesting looking at.

Jools

[MatsP]

Now, I'm lost. Romani gives two fish, the centromochlus and tatia synonyms. roman (with no i) lists the same two scientific names and "Roman's woodcat", also referring to the same actual species.

It is more annoying if there are several fish that has changed name several times: typing in "pardalis" gives three different pictures of Pterygoplichthys pardalis, along with a few different other species. "seemanni" lists the same species 5 times under different names - it's obvious that it's the same fish, but I'm not sure that is ALWAYS the case.

--
Mats

[Jools]

Now, I'm lost. Romani gives two fish, the centromochlus and tatia synonyms. roman (with no i) lists the same two scientific names and "Roman's woodcat", also referring to the same actual species.

Ignore me, I am being stupid. Roman gives three results (current, synonym & common name). Romani gives the first two results.

It is more annoying if there are several fish that has changed name several times: typing in "pardalis" gives three different pictures of Pterygoplichthys pardalis, along with a few different other species. "seemanni" lists the same species 5 times under different names - it's obvious that it's the same fish, but I'm not sure that is ALWAYS the case.

OK, pardalis is a good example. What's wrong with it? I think there's a conceptual thing here rather than anything functional. Imagine you thought the current species WAS Liposarcus pardalis. So, then you click on that and get your fish. Maybe the point I am making is that, for someone who doesn't know what you do, it works.

I am (I think) willing to put something like...

<picture>
Centromochlus romani

<picture>
Tatia romani
(Synonym)

<picture>
Roman's Woodcat
(Common Name) but I feel that greater use of latin names will make this less clear for the average user.

Jools

[MatsP]

Now, I'm lost. Romani gives two fish, the centromochlus and tatia synonyms. roman (with no i) lists the same two scientific names and "Roman's woodcat", also referring to the same actual species.

Ignore me, I am being stupid. Roman gives three results (current, synonym & common name). Romani gives the first two results.

I thought so, but I was just not sure enough to no question that I hadn't misse something myself.

I would, personally, prefer:
<picture>
Centromochlus romani
Tatia romani (Synonym)
Roman's Woodcat (Common Name)

After all, it is not three different fishes called three different things, so why do we need three pictures.

Some of the possible lists are quite long as it is, and having duplicate pictures is kind of meaningless.

As always, it's your site, your decision - and I appreciate the discussion, even if I don't "win"...

--
Mats

[Jools]

OK, I've had a look at it - what do you think? I think it's clumsy - but open to discussion.

Jools

[MatsP]

I like the idea - there are some details to sort out, I think. The primary one would be that the names probably should line up with the bottom of the picture, rather than the first name lining up and the others going below that name.

Also, for some reason, when I type in "pard", it shows L. pardalis and H. pardalis as synonyms, then further down the P. pardalis. They should all three come together, no matter how many hits there is.

--
Mats

[Jools]

Broadly I think you're still looking at this from the point of view of someone familiar with lots of species, common names and synonyms. This feature is for all, not just for such knowledgeable folks.

I like the idea - there are some details to sort out, I think. The primary one would be that the names probably should line up with the bottom of the picture, rather than the first name lining up and the others going below that name.

Not sure you've thought that out. I mean, more than about 4 for one picture and the species name (e.g. that most associated with the picture) will appear above the picture.

Also, for some reason, when I type in "pard", it shows L. pardalis and H. pardalis as synonyms, then further down the P. pardalis. They should all three come together, no matter how many hits there is.

Anyway, it's ordered (logically I thought) on search result (nothing to do with hits). If you search for things with pard in the name, it is in order of result.

Jools

[MatsP]

As to "where does it overflow to when there are too many", you are right, it shouldn't move up above the species picture. I hadn't thought of that, and if going down is the absolutely only way to avoid that, then so be it.

So, what you are saying is that if there is a common name that matches more than one fish, it should clearly be applied to all species that matches.

--
Mats

[MatsP]

Ah, see what you mean by "hits" - not hits on the page, but matches ("hits" matching the word typed in).

So, by what it looks like, you are not iterating through the entire list of matches from the SQL search to see if it's the same species that you've already got. I think you need a double-loop. I'll be back in a bit to show you what I think the code should look like...

--
Mats

[Jools]

OK, well that would allow us to figure it out in a common language. To aide that effort, here are the data items available to you. There is a quick find table (essentially an hourly cache style dump from various source tables) which contains everything we need to work with (I think).

id: numeric ID of the record
type: enum presently species, synonym & common name
search_term: varchar which is searched using the LIKE clause
display_term: what we show on the page of results
image: what image thumbnail is associated with the display term
link: what page is linked from the associated display term

Also, bear in mind this is output as XML results, so I am limited (without an exponential amount of work) in terms of layout (e.g. the align text with the top right of a thumbnail issue but then don't wrap and continue an indent if more results than image height).

Remember to use common names in your thought process (e.g. users are MUCH more likely to search for red tail than Phractocephalus, or L200 than even Phantom).

Jools

[MatsP]

Ok, I will work on something based on those restrictions. I take it that the search results in themselves are available in some sort of array - similar or identical to an SQL search result variable?

It would help, I think, if there was a "species id" along with the link - that would make for a simple term to compare, rather than a long string. [This doesn't matter for the algorithm itself, just for the simplicity of comparing).

Edit: If you sort the search by link, it would automatically put all same species next to each other.

--
Mats

[Jools]

Edit: If you sort the search by link, it would automatically put all same species next to each other.

Yes, but then if your search is on common name, then the results are (usually) not in a logical order - ref my comments above about thinking about other users...

Got the code, thanks. Might be a wee while before I look as it as started new job today and in Lacock. Which is quite a place name and I am sure plays ten tunes of merry hells bells with SPAM filters. Wireless we have, VPN through to my home network struggling.

Jools

[MatsP]

Not sure how the order by link will affect common names - they are pretty much random with regards to what order they are anyways [well, ok, so they come out in alphabetical order, it seems - but if I search for "woodcat", I'm not sure it matters if they are sorted by alphabetical order or not].

--
Mats

[Jools]

Ok, very common search is "red tail", so search for that and you get the results in alphabetic order. If I implemented your suggestion (which would work well for the original issue of multiple hit per species grouping) then we'd get them in an order that wouldn't make sense.

Does that explanation make sense?

Jools

[MatsP]

Sorry to be very pedantic here, but I'm sure you can find a better example - the only two fish that would change order is the Syno and South American variant - the rest (as it happens) are already in alphabetical order per scientific name.

Searching for Bristlenose gives an interesting list - including pictures of regular colour common ones, but the name saying "Albino" or "Piebald"...

--
Mats

[Jools]

Got the code and had a good think about it. It appears to show results in ID order which appears somewhat arbitrary and certainly not very useful. I am not very good with multi dimensional arrays, so I may have it wrong. However, it does not appear to address the issues (and you're right, you were being very pedantic) that I inelegantly described above.

I think it stays as is, but happy to review it again and indeed mock up a page to work on your code with the database. What might be useful (and something that I don't think I can spare the time to do right now) would be to record what people actually search for and indeed how successful that is in terms of getting them to a catelog page.

Jools

[MatsP]

Yes, I agree, it comes out in either link-order or id-order (depending on which one you look at(, which is a drawback. I will have a go at sorting in "search_term" order instead - but a visit to Tesco's first, methinks.

--
Mats

[Jools]

This is resolved (or as good as it gets for now).

Jools