[latexrefman] Looking for feedback on URL scheme for split pages

[Karl Berry]

    less easy to guess, I'd guess.

I was thinking that people basically would never "guess" such a url,
meaning type it in from scratch; they are too long and complicated no
matter what the rules are. They would follow links. Shorter links are
easier to remember or (your word) recognize, seems to me. I surmise the
only url people are at all likely to type in is "latexref.xyz".

However, if you prefer to keep the "and"s and such, I don't mind.
You're the one actually using these urls.

    What I meant is: the script converts "Starting-_0026-ending.html" to

I understand. And what I meant was, right, it's a clash because of
*your script* :). There is no clash with Texinfo itself. And it's not good
to kill urls in general, as I'm sure you agree, but in this case I feel
sure that the number of users of the "Starting-_0026-ending.html" url is
zero :).

    (As an aside grumble: that I could see @anchor(..) always makes a
    file-moved .html file. It'd be nice if there was an @ command that made

No, that would be undesirable at the Texinfo level, because then users
could create an anchor could not be visited as a url in the split-node
output. We wouldn't want Texinfo users to be able to make that mistake.

Since it's your script that is creating the conflicts with anchors, it's
your script's responsibility to resolve them :).

    if the uel changes are not worth it, I could abandon it.

I think nicer urls from Texinfo are a good thing in general, since we
are still supporting the canonical urls. I could even imagine
publicizing this to some small extent.

But anyway, whether we deploy it ... I throw the decision back to you,
since you are the one using them. I am mildly in favor myself. --thanks, karl.