|Perl: the Markov chain saw|
Re^2: Consider this: What makes a good node title?by sauoq (Abbot)
|on Nov 04, 2005 at 16:34 UTC||Need Help??|
You're arguing to do a disservice to both the original poster (by not gently correcting him/her), and the newbie searcher (by reinforcing bad searches), all in the name of allowing a particular node to be found quickly.
And I don't buy this argument. As far as the original poster goes, well, correction isn't always a service whether it is provided gently or not. On the whole, people aren't coming here looking for correction on their spelling or "correction" on the way they choose to title their online queries and musings. Gratuitous corrections are unnecessary and sometimes viewed as nothing short of impolite. And, as for the newbie searcher, I rather think that requiring him to polish his searching skills before providing him with useful results is doing him more of a disservice.
Meanwhile, we're getting in the way of more advanced searchers
No. Advanced searchers would be using Super Search effectively anyway. How is it that you think retitling nodes is helping them?
You're arguing to cater to the lowest common denominator.
I don't actually agree with that characterization but, even if I did, I would deny that that's a bad thing. That approach is often just the pragmatic approach and I do take it when I feel it's appropriate. For instance, I readily admit that I restrict my lines of code to less than 80 characters to cater to the lowest common denominator...
The reason I don't agree with that characterization in this case, however, is that the word "lowest" doesn't apply and the assumption that it does apply is the root of the problem. The fact is that the best titles are not the ones you think are best. In fact, the whole concept of best in this instance is a local optimization and a function of the searcher. Retitling nodes in an effort to bring them closer to our preconceived notions of what makes an ideal title is necessarily introducing inefficiency. Good search coverage is, essentially, an emergent phenomenon. The assumption that we can help it along by retitling nodes is fallacious.
I would suggest instead that when someone fails to search properly, we, instead, reply and gently Help Super Search Newbies instead.
You are assuming that we will know that they searched first. How many people just look for nodes that answer their question and move on if they can't find one because they are too rushed or maybe just too shy to post? Besides, plenty of people post questions that have been asked and answered before. I'll argue that telling them all how to use Super Search is counterproductive because, after enough of that, all that's going to be found when someon searches is a bunch of nodes telling them how to use Super Search. In other words, if they ask, answer and occassionally mention Super Search. That's the best way to do it and that's the way it tends to be done.
-sauoq "My two cents aren't worth a dime.";