Next Prev



Tags: , , , , , , , , , , , , , , , , ,
Exterior view. Bronze tympanum, by Olin L. Warner, representing Writing above main entrance doors. Library of Congress Thomas Jefferson Building, Washington, D.C. Cropped from the Library of Congress digital version using the GIMP.

Image via Wikipedia

I’m trying out a Zemanta blog post. What it does, apparently, is to suggest ideas for the article you’re currently writing. It’s a semantic blog suggestion feature, and it’s manifested in this instance as a firefox plugin that adds a write widget to my WordPress WYSIWYG editor. IIt updates every 300 characters, and also has ‘semantic features’. There’s an interview over at R/WW, for more information. I’m kind of trying to see what it recommends so need to fill in the 300 characters:

Well it looks like it suggests related articles, and adds a bunch of Zemanta boxes into the blog space. It also finds images from Flickr.

I could see this tool being very handy in future, though I usually blog from a client, and I don’t think this supports ScribeFire or ecto (which is rubbish, by the way.) However, there are a few problems with it:

  1. It generates an unhelpful set of areas in the blog itself. So if you include a Zemanta suggestion, it pastes it where you’re typing, and you end up typing in an alt area in the code… annoying.
  2. It updates every 300 characters. This is annoying because it’s not necessarily that real-time. This is an awkward interface feature. It also places your curser at the top of the post every time it updates, meaning what I just typed appeared above the opening line…

I think this kind of application, however, is prescient of the direction the Read/Write web is heading. It’s active and dynamic, and I’m sure the interface will be ironed out over time. I’m not sure what ‘semantic’ features they’re necessarily incorporating (is this just keyword-searched or is it tyingin with some RDF store somewhere?) but I like the way it’s heading.

I like the fact that it suggests images (all images in this post provided by Zemanta), but I’m not sure about the inclusion of ‘Zemanta’ presence everywhere… I’m also slightly concerned that some of the images it supplies are ‘license unknown’, meaning  you could use one and infringe on copyright. It does, however, have a link saying you can check it yourself, which shows they’re thinking ahead! It’s implementation of images is a bit of a struggle, however, in that you end up typing in the description area without the ability to click out of it. This is balanced by the fact that it automatically adds citations. It only adds a single image, though… so you can’t add a second image to the same blog post.

Now, they just need to make it a bit smoother, and stop jumping to the top of the bloody post 😉

Zemanta Pixie

5 thoughts on “Zemanta”

  1. Jure Cuhalev
     ·  Reply

    Hi, thanks for trying us out. The license is done in a way because Wikipedia/Wikimedia commons don’t have good metadata for all the images, so we can’t really tell you. What we can tell you is that you have to go to that page and read the license for yourself. There are a number of people working on making this better (including ourselves) so in the future you’ll see less and less of this.

    Regarding the jumping of cursor, this is clearly a bug and we’ll fix this as soon as we hunt it down.

    Thanks for you comments, we’re working on addressing all the things you’ve pointed out. This is our second release after the one at the end of March, so hopefully we’ll have something even better in next few weeks for another a bit less major upgrade (of course we’ll fix things like cursor jumping sooner).

    Jure Cuhalev, Zemanta

  2. Zach Beauvais
     ·  Reply

    Hi Jure,

    Thanks for the fast comment, and I look forward to seeing how everything works together in future. Have you thought of using DBPedia instead of Wikipedia for your source? It should have a stronger metadata structure, which might make it easier to use.

  3. Tom Altman
     ·  Reply

    It is a bit “raw” yet, but I really like it. I never posted pictures with my posts BZ (before zemanta) – but I find myself really trying hard to find a good one now. Good mini-review.

  4. Jure Cuhalev
     ·  Reply

    Yes, we looked into dbpedia, but in the end we use something called wikiprep – that we help codevelop that allows us to do similar things to dbpedia or even better. I’m not exactly sure why, but my engineers tell me that there are good technical reasons for doing this our way.

    Btw, we fixed the jumping bug earlier today.

    Jure Cuhalev, Zemanta

Your reply goes here.