A single question of "What regex do I use to strip <x> out of $my_html" is going to receive all the usual answers, and spark the usual arguments of HTML::CoolSplitter vs s/<x>//g;...regex when you know exactly what's in the markup, module if not etc. By the time somebody's asked a variant of the same question 3 times though, and code has been posted that obviously comes from the same project, one is able to state with greater certainty "in this case you should be using X".
In this particular case, it was rather obvious (due to the frequency of posts and CB discussion etc.) that there was a link between the questions, allowing alert monks to realise that posts about general strategy were more applicable than specific answers. My suspicion though is that there are often cases where that is missed.
We've all seen SOPW where we suspect that the root of the problem lies elsewhere - not in the question that is asked, but maybe in the understanding of the general principle, or in another unposted piece of code. And indeed, when we can be bothered to do the research (click on user, read their homenode, click on 'Writeups', search for questions, read a few, read the given answers...), we often find that this is the case...the same person asked a similar question 3 months ago and has been 'cargo-culting' the best of the answers for all this time, without understanding *why*.
What would be useful is some way of automating that research. I hesitate to offer specific suggestions for all the usual reasons ("The DB'll cannae take it Cap'n - she's gonna bloowww!.."), but it seems that possibilities range from an (optional?) single 'see user writeups' link next to all "posted by" links, through to some kind of fully-fledged-topic-tracking-thread-linking-project-indication mechanism. But knowing nothing about The Engine, I'm keeping my mouth shut. :)
Just a thought.
Ben.
In reply to Context Indication by benn
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |