Comments on the Technorati redesign
July 26, 2004 at 6:32 PM
So the latest buzz about redesigns is the Technorati redesign. Adaptive Path worked with Tim Gasparak to develop the initial CSS/XHTML standardized look and feel, then Tantek Çelik took templates and code and reworked them to develop what you see on the site today. The sad part is that none of the IA of the site was altered as far as I can tell and the performance continues to be pretty lacking. There is a lot of common IA mistakes going on here in the results page, but as I understand it they are working fixing these issues in the next phase of the redesign. They were a little rushed this time around. Which will hopefully include better performance optimization and some better architecture.
What "common IA mistakes"? Keen to know from the point of view of an academic exercise and not flaming them. Thanks.
Erm, doesn't tantek deserve a mention - it is my understanding theat it was he who pulled the adaptive path design in and dealt with all the validation issues etc to give us the site we see today... http://tantek.com/log/2004/07.html#redesigned-technorati Although i may be mistaken
Yeah it looks nice. But 50% of the time the service doesn't work for me. IA and usability add up to a whole hill of beans when the website doesn't work.
I find that there is insufficient contrast between type colors and background colors in several of the areas of the page. Especially in pull qoute boxes. Light grey on white, or vice-versa, is especially difficult to read using your fonts.
Common IA Mistakes: - Mystery meat navigation (icons that don't really mean much) - chunking of information (highlighting key points instead of all the same size-ed, non-bold, text) - Unintuitive links (multiple links but noting indicating if they link to more details or the actual site) - Content formatting (lack of any formatting to indicate if the link is from just a blogroll or a blog post) - Random photos (why did they do this? It's annoying) - No Filtering (I can't distill the search results at all) - Poor labeling ("freshness" "authority" ..how about using "date" and "rank" instead?) - Not enough depth ("n links from n sources" ...show me the links? show me the sources?) - Poor grouping of information (created and updated date/time stamps should appear side by side... right now it's confusing to have one above the result content and one below) - Full disclosure (show me the URL, the titles are often misleading) - Special issues with pagination links (how about separating "next" from the page number links).