Re: A Roadmap for TexInfo without Info

[Assaf Gordon]

Hello,

> On Nov 25, 2016, at 04:00, Gavin Smith <address@hidden> wrote:
> 
> On 24 November 2016 at 19:55, Assaf Gordon <address@hidden> wrote:
>> one minor annoyance that stems from the "info" format: the references 
>> (@ref/@xref/@pxref).
> 
> As Eli said, the main problem with the above is the printed output,
[...]

Perhaps I should rephrase: the problem is in the "texinfo/info" eco-system, 
where, if I want to adhere to the best-practices, these kind of references are 
discouraged (by the official manual).

IMHO, the distinction between "print" and "online" should be refined, as both 
HTML and PDF can be used for printing and for online viewing. The distinction 
should be between formats/displays/interfaces that support embedded linking, 
and those that don't.
Additionally, with modern HTML capabilities we can also know if it being viewed 
on screen or being printed. 

I'd like to be able to do the following. Perhaps it's already possible, I'm not 
sure how.

I want to be able to write something like:

    The @option{-E} option turns on @linkref{extended regular expression} mode
    @printref(extended regular expression}.

Where "@linkref" would generate an embedded link, with the text being "extended 
regular expression" and look natural as part of the sentence. It should work in 
HTML, PDF when viewed on screen, and all other interactive formats.
When printed, it should just be the text "extended regular expression" without 
additional text (e.g. no "see" or pages) - effectively looking like normal text 
and losing its cross-ref attributes.

And "@printref" is something that generates a stand-alone expression, with 
parenthesis, suitable for inclusion in printed material (such as PDF, or HTML 
with special media=print tags), and it should be something like "(see Extended 
Regular Expressions, Section X, Page Y)".

This is just a rough idea, I'm sure there are edge-cases I haven't considered.
But my point is - I want to take advantage of convenient cross-referencing with 
media that allows it,
and not be limited by having to consider the lowest-common-denominator of "how 
it will look in info".

> The Info output may be suboptimal, but you are free to ignore it, as I said.

I suspect others (Eli for example) would not agree that "I'm free to ignore 
it"...


regards,
 - assaf