Re^2: Code tags warning

[Tanktalus]

That's a good point. However, if we¹ can put in heuristics to the PM code to decide when code tags are required (and are not actually present), and we decide these heuristics are accurate (i.e., zero false-positives, few false-negatives), and we decide that this would be beneficial to the community at large, then we could make it mandatory: no "submit" button until all warnings are dealt with.

I would propose that there also be a way around these heuristics. That is, I like the computer being smart. But not too smart. For example, perhaps if you submit a preview without having changed anything, then you get the submit button on the next preview. Or, perhaps, there is a checkbox saying, "ignore warnings" which, if unset (default) would make the submit button act as the preview button.

And then, any time warnings are bypassed, perhaps we could automatically consider the node for those warnings? :-)

¹ I'm using a very, very loose definition of "we" here. As in, "we" == "gods".

[ambrus]

I'm very much against making it mandatory. Both with code and readmore tags.

I can accept a warning, as long as I can turn it off (such as in User Settings), but I don't think there's much point in it.

[ww]

... as with turning off the "preview" option -- which is restricted to those with a certain experience level? I think that's Friar, but too lazy too look it up, since the point is that one who has some significant experience may well have a vested interest in keeping PM useful... and the Monastery neat and tidy.
So, I like Ambrus modification of the OP and first reply... as I like Joost's suggestion that we use css to make the warnings more prominent.

However, I think Ambrus' concern that some folks (/me included) ignore stock warnings is best dealt with by (1) making the warnings specific and content-driven (the generic warning appears in EVERY preview, regardless of whether a warning is relevant) and (2) the Roy Johnson suggestion (at Re^2, well down in the thread) as augmented by Whitehawke's (tongue in cheek?) suggestions for adaptations of 'fortune' (AKA 'cookie').

[Anonymous Monk]

If you can get zero false-positives by using automation, then instead of not having no submit button, add the code tags automatically.

I would propose that there also be a way around these heuristics. That is, I like the computer being smart. But not too smart.

Ah, so that's in the case that the "zero false positives" is actually for values greater than 0? ;-)

[Tanktalus]

Figuring out that there are missing code tags is much easier than figuring out where they should go. For example, if I have: $blah[0] = $blah[1], is that: $blah[0] = $blah[1], or is it $blah[0] = $blah[1], or is it $blah[0] = $blah[1], or is it $blah[0] = $blah[1], or is it ... you get the idea. Figuring out that the ['s and ]'s are supposed to be in code tags is solvable. Figuring out where those codes should have been ... probably is not solvable.

Given the ++/-- score of my post and ambrus' post, it seems that there is no clear majority on this, which, IMO, means the status quo wins. (Without a clear majority, the gods would be ill-advised to spend time on a feature that would likely take much effort for little return.) That's fine by me - I was just expanding on the OP, in the way I understood it anyway.