Oh then maybe I'm misunderstanding "deprecated".

I was thinking of it from the narrow/selfish point of view of a (new
system) package developer. In that system, deprecated can't mean "will
be removed soon", because you can't ever remove things from a package.
(Packages are supposed to be backward compatible, even with a new
major version number. If you want to remove stuff, you're supposed to
make another package. (I'm not trying to re-open that discussion, now,
just follow the status quo.))

So I was thinking of it as purely advice/admonition: "Starting with
version N, you ought to use Shiny instead of Finicky." However,
Finicky will remain forever in that package per se.

I don't know whether that's "deprecated" or not, but it's a thing.

I'm curious to know the answer to Alexis' question: Semantics aside,
how do I simply make something with a standard warning appearance?  In
many cases that's probably sufficient.  I went looking for examples of
this in the docs, so I could go look for its Scribble. But I couldn't
find any. Alexis do you have a link to an example?


On Tue, Oct 13, 2015 at 11:48 AM, Vincent St-Amour
<[email protected]> wrote:
> Someone correct me if I'm wrong, but my understanding is that, in the
> context of Racket at least, "deprecated" implies that we will be
> removing that code in the near-ish future. As we recently did for, e.g.,
> class100.
>
> My understanding is also that we don't want to remove things like
> mzscheme, mzlib, planet, or compatibility. So using the above definition
> of deprecated, we don't want to deprecate those.
>
> With that said, though, I very much agree with your sentiment: we should
> have a clear way of signaling that people shouldn't use those. As we've
> learned from the "unstable" collection, even large warnings are not
> enough for that if there's valuable stuff in there.
>
> With the package system, however, it's possible to remove code from the
> main distribution without breaking code that depends on it (it just
> needs the right dependencies). So it should be possible to essentially
> do the same for the compatibility-lib package (which contains most of
> the things I suspect you'd want to deprecate) as we did for the
> unstable family of packages, and move it out of the main distribution.
> In the process, we can move the last few bits of useful functionality
> that exists in these collections, but not in more modern Racket
> equivalents.
>
> If we do that, then most people won't have that package installed, it
> won't show up when they search the docs, and we win. And we can do that
> without breaking anyone's code (assuming they have the proper
> dependencies already).
>
> (That may be trickier to do with planet, though, as I think it may be
> embedded pretty deeply into our toolchain.)
>
> Doing that for compatibility-lib and other non-recommended libraries
> (like srfis) is on my to-do list, but not currently on the top.
>
> Vincent
>
>
>
> On Mon, 12 Oct 2015 19:01:51 -0500,
> Neil Van Dyke wrote:
>>
>> There's still a problem of people stumbling upon deprecated stuff in the
>> Racket manuals, and going to a lot of trouble to use it without
>> realizing that it's deprecated.
>>
>> For a very quick fix, how about appending "[DEPRECATED]" to certain
>> manual titles, as an easy way of marking each page of the manual?
>>
>> For example, the title: "PLaneT: Automatic Package Distribution"
>> is changed to: "PLaneT: Automatic Package Distribution [DEPRECATED]"
>>
>> This could also be done to individual "compatibility" chapters/section
>> headers of otherwise non-deprecated manuals.
>>
>> Also, move things like the PLaneT manual to the "Legacy Languages and
>> Libraries" category.  Maybe append "[DEPRECATED]" to the titles of the
>> other manuals in that category, as well.
>>
>> Neil V.
>>
>> --
>> You received this message because you are subscribed to the Google Groups 
>> "Racket Developers" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to [email protected].
>> To post to this group, send email to [email protected].
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/racket-dev/561C49EF.3010805%40neilvandyke.org.
>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups 
> "Racket Developers" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
> To post to this group, send email to [email protected].
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/racket-dev/m2twpupy61.wl-stamourv%40eecs.northwestern.edu.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-dev/CAGspUn2FjXnohmT88Tefm5C5QZ54-%2BiVg6nff4-xRkdw7s-TtQ%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to