It's been less than 3 months since I launched the start of JungleDragon v2, along with its key specie identification feature. I consider it JungleDragon's differentiating feature, a feature to be proud of. I am even more proud however at how the community embraced it. I've developed quite some features that are little used in JungleDragon, but specie identification is not one of them. It's a sport in JungleDragon to identify species on photos or to help others do it.

As a result of this success, we have reached a new milestone: that of 1,000 unique species (1,009 currently to be exact). Combined with a 76% identification rate on photos, that makes for a truly educational link between photos and specie data.

A 1,000 species doesn't sound like that much, but its actually quite a solid base. It's a number that pales against the millions of species known to science, that much is true. Yet, from a practical sense, most species in existence are poorly documented, you unlikely know about them and are unlikely to be looking for them. For example, there are over 30,000 different species of snout beetles. The odds of finding an uncommon one at JungleDragon are poor, yet the most common ones you may be able to find. And that's a reasonable place to be for now. 

Anyway, so celebrate this little milestone, I just made a small donation for the preservation of the Tiger. 

Just a quick little review of the book Adaptive Web Design: Crafting Rich Experiences with Progressive Enhancement...

This is a hugely popular book on modern progressive enhancement strategies for web development that I had in my reading backlog for months. It turns out it's only a short read, yet I consider that a compliment. I'm done with 1000-page tech books.

In this book, the author lays out a very pragmatic strategy for building up a website in layers to maximize reach. For example, one starts with building the content using semantic markup, and that content can be experienced on any device, even without any CSS or Javascript support. A secondary layer adds CSS for a better experience. A third layer adds Javascript for better interactivity, and so on. All of these layers enrich the layers below it yet are setup in such a way that they can fail. The concept falls somewhat inline with recent trends of "content first" and "mobile first", both concept I strongly believe in (yet hard to apply afterwards for existing web projects).

Furthermore, the author makes some really good points on how a similar strategy (graceful degradation) is actually not that similar, and in general a bad strategy to follow. A nice eye opener.

As I have read some related books and follow several web design blogs, few techniques were completely new to me, yet the sum of it all still makes this a worthy read. Recommended reading, unless you have digested these strategies already from similar books.

Update: We had a discussion if embedding photos without explicit approval of photo owners would be a legal risk and concluded that this is the case. At first I removed the embedding feature entirely, but I have now redeployed it. This time, the default setting is "no". This means you need to enable it if you want other users to be able to embed your photos.

The image you see above is a JungleDragon embedded photo, the first one ever published. I've been longing to implement this for a long time and today I finally got to it. Here's how it works...

Photo pages now have a new Embed button, right next to the social sharing widget:

If the owner of the image has photo embedding disabled, this button will not be visible. More on that in a minute. Clicking the embed button takes you to this screen:

From here, you can select from 3 image sizes using the tabs. Just select the size you would like to see embedded. As you can see, I have limited the selection to thumb size formats. The idea is that the embedded photo leads to a click to see it larger within JungleDragon.

The second part of the screen shows the embed code. The first one is the HTML embed code. If you have your own website or blog, you can simply copy and paste that code. The example at the start of this blog entry is embedded this way. 

The second embed code is in the BB format, which is often used at forums on the web. Here's an example of a JungleDragon image embedded in a forum based on phpBB:

Again, it's just a matter of selecting the preferred size and copying and pasting the embed code that fits your needs. 

JungleDragon values the copyrights of photo owners greatly, therefore, should you not wish to see your photos embedded this way, you can disable it using a new setting on your profile:

By default, embedding is disabled.

Spread the love

Do you have a website, blog or participate in a forum? I highly encourage you to embed JungleDragon photos in it, where it makes sense. And of course, there are the existing ways to share JungleDragon content on social networks. Let's build a JungleDragon web!

If you'd ask me what JungleDragon's most important feature is, it would be specie identification (and as a result, specie browsing). It has changed JungleDragon as a whole, lifting it from a niche photo sharing site focusing on wildlife, to a deeper experience where photos have context and users actually learn about wildlife. And sometimes it goes beyond that. Specie identification has not only changed the site, it has also changed people. Some users report that they go about their nature walks and photography different then before. They've developed  knowledge about species and a hunger to know what they are photographing. I consider this behavioral change to be the ultimate compliment to JungleDragon.

So it is evident that species are crucial in JungleDragon, and for this we have the specie engine. It's a complicated piece of software that basically takes unstructured specie information from Wikipedia and attempts to put it intro structured JungleDragon specie entities so that they can be linked to photos. Once that is done, species can be further extended in JungleDragon by adding videos and a geomap.

In practice, the process of specie identification is not perfect, but still pretty good. In about 5-10% of all cases the identification of a new specie results in a blocking error, where the entire identification fails. I continually look for the source of these problems and try to patch the specie engine so that it does not fail for similar cases in the future. 

The other 90% leads to a succesful specie identification. Yet, this category is still not problem free. In about 20 to 30% of these cases, there are (slight) data problems. For example, the taxonomy may not be complete. That means it will not appear correctly in the specie hierarchy (link). Another common problem is malformed data, that contains unexpected characters due to small errors in the parsing of Wiki markup. Those are not really an issue other than looking ugly. And finally, it may be that nothing fails yet Wikipedia simply does not have specific data.

All in all, this means that the specie engine requires manual data corrections. The effort to do so is manageable, but the process to do it has been nasty so far.  I'm using phpMyAdmin to fire manual queries to various tables. Not anymore though. As of now, JungleDragon admins (me) see an edit link below each specie page:

Clicking that link leads to the obvious specie edit form:

From this single screen, I can manage all specie data. Having a friendly UI like this greatly simplifies specie maintenance, which benefits both me and the community.

Just hand over your life and its free!

Continue reading...

	Articles
	Blog entries
	Comments