[seqfan] Re: A139414

[David Wilson]

Yes, but if rankings are implemented in the wrong way, they will be 
skewed. I could click on a sequence I like 1000 times to push it up the 
ranks. I assume Google is not that naive.

Dmitry Kamenetsky wrote:
>>> PS: another idea: implement "votes" for sequences and/or authors - but
>>> that's a little less immediate to implement.
>>>       
>
> I actually really like this idea and I think it can be implemented in an
> automatic fashion. Think of something like Google Rank. Sequence that have
> many links and cross references get a higher ranking than those that have
> few links. Naturally, the less interesting sequences will have fewer links
> and thus will be ranked low, so they will be further down the search and
> rarely seen. Similarly, every time someone clicks on a sequence it should
> reinforce its rank. Notice that the rankings are self-reinforcing, which is
> nice. I think this is a safe way of keeping all the sequences, while not
> throwing too much useless information at the user.
>
> Dmitry
>