argument names for define-obsolete-*-alias

argument names for define-obsolete-*-alias

[Juanma Barranquero]

Is everybody happy with the names of the `make-obsolete*' and
`define-obsolete-*-alias' arguments?

The alias being created is called FUNCTION or VARIABLE, and the symbol
(or definition) being aliased, NEW. Now, I understand the rationale:
NEW because it is the new, shinier name.

However, and even knowing as I know that all defining functions and
macros get the defined symbol as first argument, today I tried to use
`define-obsolete-variable-name' and on first reading of the docstring
I just put the arguments reversed. From the Emacs history POV, NEW is
indeed newer; from the use case of define-obsolete*, however, NEW is
"old" because it already exists.

Wouldn't be better to use OBSOLETE and CURRENT (or MODERN), or ALIAS
and ORIGINAL, or some other pair that best defines the relationship
between the arguments?

-- 
                    /L/e/k/t/u

argument names for define-obsolete-*-alias

[Nick Roberts]

 > Is everybody happy with the names of the `make-obsolete*' and
 > `define-obsolete-*-alias' arguments?
 > 
 > The alias being created is called FUNCTION or VARIABLE, and the symbol
 > (or definition) being aliased, NEW. Now, I understand the rationale:
 > NEW because it is the new, shinier name.
 > 
 > However, and even knowing as I know that all defining functions and
 > macros get the defined symbol as first argument, today I tried to use
 > `define-obsolete-variable-name' and on first reading of the docstring
 > I just put the arguments reversed. From the Emacs history POV, NEW is
 > indeed newer; from the use case of define-obsolete*, however, NEW is
 > "old" because it already exists.

NEW is an argument, not part of the function name, so I wouldn't expect
it to create a new symbol. These functions/macros are also defined in
XEmacs with similar argument names which makes it clearer that they
really do the same thing.

 > Wouldn't be better to use OBSOLETE and CURRENT (or MODERN), or ALIAS
 > and ORIGINAL, or some other pair that best defines the relationship
 > between the arguments?

I think these are confusing. OBSOLETE and CURRENT would only be 
obsolete and current _after_ the function/macro has been evaluated.
Maybe FROM and TO would work. 

On a related note, I find the arguments for defvaralias and defalias more
confusing than those for define-obsolete-*-alias.

Nick

Re: argument names for define-obsolete-*-alias

[Juanma Barranquero]

On 6/13/05, Nick Roberts <nickrob@snap.net.nz> wrote:

> NEW is an argument, not part of the function name, so I wouldn't expect
> it to create a new symbol.

I don't understand this.

> These functions/macros are also defined in
> XEmacs with similar argument names which makes it clearer that they
> really do the same thing.

In which way being the same on XEmacs makes it clearer what they mean? :-)

> I think these are confusing. OBSOLETE and CURRENT would only be
> obsolete and current _after_ the function/macro has been evaluated.
> Maybe FROM and TO would work.

Whatever works.

> On a related note, I find the arguments for defvaralias and defalias more
> confusing than those for define-obsolete-*-alias.

For me is just the opposite.

-- 
                    /L/e/k/t/u

Re: argument names for define-obsolete-*-alias

[Daniel Brockman]

I just thought I'd chip in and say that I fully agree with Juanma; the
current names are really confusing.  I think almost any other pair of
symbols would be an improvement.

-- 
Daniel Brockman <daniel@brockman.se>

   ``Why fix an old bug if you can write three new ones
     in the same time?'' --- David Kastrup

Re: argument names for define-obsolete-*-alias

[Richard Stallman]

    Is everybody happy with the names of the `make-obsolete*' and
    `define-obsolete-*-alias' arguments?

    The alias being created is called FUNCTION or VARIABLE, and the symbol
    (or definition) being aliased, NEW. Now, I understand the rationale:
    NEW because it is the new, shinier name.

I can see how this could be confusing.  I suggest using names
OLD-OBSOLETE-NAME and CURRENT-RECOMMENDED-NAME.

Re: argument names for define-obsolete-*-alias

[Richard Stallman]

    On a related note, I find the arguments for defvaralias and defalias more
    confusing than those for define-obsolete-*-alias.

defalias seems clear to me, but defvaralias is unclear.  For
defvaralias, I suggest the names NEW-ALIAS and BASE-VARIABLE.

If using similar names in defalias would make it clearer to some people,
we could change that too--it costs little to change this.

Re: argument names for define-obsolete-*-alias

[Juanma Barranquero]

> I can see how this could be confusing.  I suggest using names
> OLD-OBSOLETE-NAME and CURRENT-RECOMMENDED-NAME.

I like these, but they are a bit too long for docstrings, etc.
Wouldn't just OBSOLETE-NAME and RECOMMENDED-NAME (or CURRENT-NAME)
suffice?

-- 
                    /L/e/k/t/u

Re: argument names for define-obsolete-*-alias

[David Kastrup]

Juanma Barranquero <lekktu@gmail.com> writes:

>> I can see how this could be confusing.  I suggest using names
>> OLD-OBSOLETE-NAME and CURRENT-RECOMMENDED-NAME.
>
> I like these, but they are a bit too long for docstrings, etc.
> Wouldn't just OBSOLETE-NAME and RECOMMENDED-NAME (or CURRENT-NAME)
> suffice?

How about just OBSOLETE and CURRENT?

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: argument names for define-obsolete-*-alias

[Juanma Barranquero]

> How about just OBSOLETE and CURRENT?

That was my first suggestion...

-- 
                    /L/e/k/t/u

Re: argument names for define-obsolete-*-alias

[David Kastrup]

Juanma Barranquero <lekktu@gmail.com> writes:

>> How about just OBSOLETE and CURRENT?
>
> That was my first suggestion...

Well, IMHO the more complicated proposals don't provide noticeable
improvement.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: argument names for define-obsolete-*-alias

[Richard Stallman]

    > I like these, but they are a bit too long for docstrings, etc.
    > Wouldn't just OBSOLETE-NAME and RECOMMENDED-NAME (or CURRENT-NAME)
    > suffice?

    How about just OBSOLETE and CURRENT?

That is more cryptic--good enough to remind people, but why not
add the extra 5 chars to make it clearer?

Re: argument names for define-obsolete-*-alias

[Juanma Barranquero]

> That is more cryptic--good enough to remind people, but why not
> add the extra 5 chars to make it clearer?

I installed it yesterday with OBSOLETE-NAME and CURRENT-NAME.

-- 
                    /L/e/k/t/u

Re: argument names for define-obsolete-*-alias

[Nick Roberts]

 > > That is more cryptic--good enough to remind people, but why not
 > > add the extra 5 chars to make it clearer?
 > 
 > I installed it yesterday with OBSOLETE-NAME and CURRENT-NAME.

Please could you also change the entries in the lisp reference manual to use
the same argument names.

Nick

Re: argument names for define-obsolete-*-alias

[Juanma Barranquero]

On 6/15/05, Nick Roberts <nickrob@snap.net.nz> wrote:

> Please could you also change the entries in the lisp reference manual to use
> the same argument names.

Done.

-- 
                    /L/e/k/t/u