david said: > Compare the manually specified > https://github.com/plaid/sanctuary#either > to the auto-generated > https://github.com/plaid/sanctuary#either--a---c---b---c---either-a-b---c
ok, first of all, if you manually specified this anchor: > https://github.com/plaid/sanctuary#either no way ever would i expect it to go to that location. especially since this link: > https://github.com/plaid/sanctuary#Either goes to another location. (told you about case glitches.) nope, i would expect your manually-specified link to go to the location where this link actually goes: > https://github.com/plaid/sanctuary#either-type (ironically, that header actually has a capital letter, but the anchor is nonetheless in all-lowercase. ha.) no, the thing about "manually specified" anchors is that they will show all kinds of idiosyncrasies. so not only are they work to create, but they are also inconsistent, thus undependable, meaning you have to look at each one to see what it is, so they have almost _no_ redeeming qualities at all. furthermore, the github methodology must be doing some kind of javascript-based search -- which is not a bad idea, as i suggested here: > > https://medium.com/the-bower/lets-please-use-hashtag-terms-to-create-_arbitrary_-deep-links-f9185d5da2f0 (grok that deep-link! hyphens _and_ underbars! and random crap at the end just to confuse you.) except whatever github is doing is implemented very strangely, so is certainly not what i would do. however... *** alan said: > If I were to pick a preferred format for auto-generated IDs, > it would have to be hyphens. Separating words is desirable > as it is more readable. And underscores, though not a terrible option, > can be obscured by underlines when the entire URL is presented to a user. i disagree with several of those points. but, hey, our own personal opinions don't mean jack. i would gladly accept _any_ method, even if distasteful, as long as everyone agreed that that's what we're doing. so what would make a difference is if all the flavors would converge on a mutual-agreed-upon method. but... > Of course, I've been on this list too long to expect > implementations to actually converge on much of anything. right, that's never gonna happen, i think we all know that. so, then, what we need is to poll a million users or so, and see which of the possible solutions they would most prefer, and then let that vote guide us. but that won't happen either. so the guiding principle is this: if it _is_ broke, but you know you can't fix it, don't waste time and energy by even trying. don't even discuss it. pretend it doesn't exist. (that's something all the flavor-makers can do!) i would suggest we have reached this point. thank you for your question, original-poster... -bowerbird _______________________________________________ Markdown-Discuss mailing list Markdown-Discuss@six.pairlist.net https://pairlist6.pair.net/mailman/listinfo/markdown-discuss
