I recognize some of my struggles in that piece. People get tired of me saying it, but I don't care:
"Chose your semantics wisely, for they do have consequences."
Ironic that in a discipline which requires such a high degree of precision in the use of its symbols and tokens consistently prescribes obscure and contradictory rules for their usage.
Although not required, success in this field is enhanced if one possesses the ability to mentally manage symbol sets with a fairly agnostic attitude towards their clarity and consistency.
When I was a newbie, I recall more than one veteran C programmer giving me a quizzical look when I asked who the hell thought it was a good idea to use the same symbol to express assignment and equality.
Yes,the Register is shrill to the point of borderline tabloid status. But they do put out a significant amount of original work.
Just to see for myself, I went to the Wiki page for Gary Weiss' bio and saw numerous attempts had been made to add a link to this Register article and all had been undone immediately.
Hopefully, Citizendium (like Wikipedia, but no anonymous editors and there is an editorial board) will be able to gain some traction:
I am new to Wolfram's work, so I am at a disadvantage, but I find the idea interesting.
Seems unlikely that the "viral potential" for a given random web app could be gleaned by perusing the code or that there will soon be technology available to simulate a user. And if you were to reduce the randomness by somehow requiring the code or the app to conform to some standard in order to make it easier to automate the perusal, well we already have a whole SEO industry dedicated to gaming the current automated search technology, don't we?
At some point don't you think it is likely that the social networking application model will be co-opted by market testing enterprises who will design ways to entice lots of people to test software applications while having their physiological reactions captured online. With a record of human reactions directly correlated to a specific application, predictive results from automated searching would then be feasible.
Overall everything worked well. However, I did find the navigation in the "Results" page a bit counter-intuitive:
I entered "beer" as the seek term and was taken to the "Results" page and I clicked on the first entry in the results--"Beers" which is a list--and got a new page with a view of the items in that list. But when I clicked on the second entry in the results--"beer" which is an item--I didn't get a new page with a view of the item. Instead I got a preview of the item in the frame on the right-side of the Results page.
After further review, I saw the link to the right of the "beer" entry labeled "view" and was able to view the page for that item. I really don't think labeling the links to the right of each results entry differently--"preview" for a list and "view" for an individual item--will prevent the user from being surprised by this behavior.
Seeklr is a very intriguing site. The presentation is very attractive and implemented nicely. Good job and good luck.