Announcing ImageWind.Net - Now in Public BetaSubmitted by smartyP on Sun, 10/18/2009 - 23:28
I'm very happy to say that tonight I'm making public my current side project, ImageWind.Net.
ImageWind.Net is written in Silverlight 3, and is a real-time visualizer of the images being posted to Twitter via TwitPic. The stream of images is like a never ending circus of the internet - but watch out, you may not want to see some of this stuff.
The current beta doesn't have much in the way of options yet, but it does have some basic abilities including retweeting an image, viewing the tweet details for an image, and viewing the images in both thumbnail and full views. There is a lot of work left to do, and a good bit of work already done but not yet finished, so please stay tuned for future updates here and in the beta's change log.
Awesome demo :-) And I really
Awesome demo :-)
And I really like it. Its a little bit like watching the whole world at once.
Do you provide source code (is this a codeplex project) or is it "closed source"...
In any case, a small improvement suggestion (perhaps, if feasible): Filter duplicates...
There are a lot of "advertising" images which appear doubled. Its not really bad, but in my opinion, it could be solved easily. Just hash the byte code of the thumbnails, remember it, and discard all thumbnails with the same hash. Not sure about the computing power consumption, but I think it should be ok... Just an idea, really a funny project...
this will not be an open
this will not be an open source project, but if there are things you see in it you would like to better understand i would consider writing up blog posts on them..
as for duplicates, it currently keeps record of all the image url's that are shown so that the same twitpic image will never be shown again that session, however if the same image is posted multiple times with different image urls then there is nothing being done for that right now..
the loading/displaying of the image currently doesnt allow access to the actual bits of the image to do the hash code suggestion you mention.. it could be modified to do that (and was initially), but the memory/cpu overhead of dealing with all the byte arrays for the images directly added significant overhead, enough to where it isn't worth the slowdown.. it may be worth revisiting again though at some point..
thanks for the comment, and stay tuned for further updates!