Package "luwak"

Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

We just added a package with the name "luwak".  It browses using lynx.

The name "lynx" is somewhat helpful because it refers in an unmissable
way to the links that make the WWW better than ftp.

By contrast, the name "luwak" is unhelpful.  Wikipedia says "luwak" is
a species of civet; civets are not cats, but fairly close to cats.  If
you know this, it might help you remember "luwak" when you think of
"lynx".  But that relationship is remote and strained.  And it won't
help people who might find luwak useful find out about it.

Let's change the package name now to something that will help users
find out about it and use it.

My suggestion for the name is `lynx'.  That name fits it, since it is
an interface to lynx.  And anyone who knows about `lynx' will guess
what it does.

Does anyone have a better name to suggest?

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Stefan Monnier]

> We just added a package with the name "luwak".  It browses using lynx.
[...]
> Let's change the package name now to something that will help users
> find out about it and use it.

My hope is that this package will disappear (merged with eww.el), so
I think it's OK it stays somewhat obscure until then.


        Stefan

Re: Package "luwak"

[North Year]

[-- Attachment #1: Type: text/plain, Size: 549 bytes --]

Richard Stallman <rms@gnu.org> writes:

> We just added a package with the name “luwak”.  It browses using lynx.
>
> The name “lynx” is somewhat helpful because it refers in an unmissable
> way to the links that make the WWW better than ftp.
>
> My suggestion for the name is `lynx’.  That name fits it, since it is
> an interface to lynx.  And anyone who knows about `lynx’ will guess
> what it does.
>
> Does anyone have a better name to suggest?

Questions: why do we need an interface to lynx when we already have eww?

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > My hope is that this package will disappear (merged with eww.el), so
  > I think it's OK it stays somewhat obscure until then.

Ok, if it does get mered and disappear.
Could we make a decision to revisit the name in April
if it does NOT disappear by then?

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[T.V Raman]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=gb18030, Size: 1028 bytes --]

Richard Stallman <rms@gnu.org> writes:


1+.

As a likely potential user, I looked at that package and didn't see
anything there that would improve EWW after a quick review; 

More generally, could we have some kind of code review and code quality
check by a few folks in general  before adding a package --- taking a
package out tends to have more friction than putting something in over
the long term.


 > [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > My hope is that this package will disappear (merged with eww.el), so
>   > I think it's OK it stays somewhat obscure until then.
>
> Ok, if it does get mered and disappear.
> Could we make a decision to revisit the name in April
> if it does NOT disappear by then?

-- 

Thanks,

--Raman(I Search, I Find, I Misplace, I Research)
7©4 Id: kg:/m/0285kf1  •0Ü8

Re: Package "luwak"

[Will Mengarini]

* Richard Stallman <rms@gnu.org> [22-11/30=We 18:51 -0500]:
>> We just added a package with the name "luwak".  It browses using lynx.
>> [...]
>> Let's change the package name now to something
>> that will help users find out about it and use it.

* Stefan Monnier <monnier@iro.umontreal.ca> [22-11/30=We 19:52 -0500]:
> My hope is that this package will disappear (merged with eww.el),
> so I think it's OK it stays somewhat obscure until then.

The name "eww" is also unhelpful.  Perhaps
the merger should be under the name "lynx".

Re: Package "luwak"

[Jean Louis]

* Will Mengarini <seldon@eskimo.com> [2022-12-05 02:19]:
> * Richard Stallman <rms@gnu.org> [22-11/30=We 18:51 -0500]:
> >> We just added a package with the name "luwak".  It browses using lynx.
> >> [...]
> >> Let's change the package name now to something
> >> that will help users find out about it and use it.
> 
> * Stefan Monnier <monnier@iro.umontreal.ca> [22-11/30=We 19:52 -0500]:
> > My hope is that this package will disappear (merged with eww.el),
> > so I think it's OK it stays somewhat obscure until then.
> 
> The name "eww" is also unhelpful.  Perhaps
> the merger should be under the name "lynx".

But hey, it is abbreviation for GNU Emacs Web Wowser (EWW), quite
clear to me 😇

EWW is not related to Lynx.

I can't see why would luwak be merged with eww, when those packages
use totally different methods to obtain Internet information.

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > The name "eww" is also unhelpful.  Perhaps
  > the merger should be under the name "lynx".

Does EWW work using lynx?  If so, that would be a good idea.  We would
replace two unhelpful package names with one name that would be
helpful and easy to remember.

If EWW does not work using lynx, then renaming it to `lynx' would be
misleading at one level.  We might rather make a different choice.

What other packages do we have for looking at web pages?  I see there
is `browse-url'.  Is that an alternative to EWW, or just a higherlevel
confugurable interface to EWW and other options?

Are there any more web-browsing packages that we should consider in
this planning?

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Philip Kaludercic]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > The name "eww" is also unhelpful.  Perhaps
>   > the merger should be under the name "lynx".
>
> Does EWW work using lynx?  If so, that would be a good idea.  We would
> replace two unhelpful package names with one name that would be
> helpful and easy to remember.

No it doesn't, EWW uses shr (Simple HTML Renderer) to display websites.

> If EWW does not work using lynx, then renaming it to `lynx' would be
> misleading at one level.  We might rather make a different choice.

EWW has been around for a while (I belive 24.4), I don't think it is
viable to rename it now.

> What other packages do we have for looking at web pages?  I see there
> is `browse-url'.  Is that an alternative to EWW, or just a higherlevel
> confugurable interface to EWW and other options?

No, browse-url is a generic system for requesting a URL be opened in
some browser, be it Firefox, Chrome, EWW, or whatever the default
settings are on a system.

> Are there any more web-browsing packages that we should consider in
> this planning?

None that I know of.

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

Thanks for the info.

What are the advantages of using shr to look at web pages?


-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Eli Zaretskii]

> From: Richard Stallman <rms@gnu.org>
> Cc: seldon@eskimo.com, monnier@iro.umontreal.ca, emacs-devel@gnu.org
> Date: Tue, 06 Dec 2022 17:37:09 -0500
> 
> What are the advantages of using shr to look at web pages?

It's a package that was written specifically to render HTML.

Re: Package "luwak"

[Philip Kaludercic]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Richard Stallman <rms@gnu.org>
>> Cc: seldon@eskimo.com, monnier@iro.umontreal.ca, emacs-devel@gnu.org
>> Date: Tue, 06 Dec 2022 17:37:09 -0500
>> 
>> What are the advantages of using shr to look at web pages?
>
> It's a package that was written specifically to render HTML.

And the fact that being written in Emacs Lisp, it doesn't require
external dependencies like lynx or w3m.

Re: Package "luwak"

[Andreas Schwab]

On Dez 07 2022, Philip Kaludercic wrote:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> From: Richard Stallman <rms@gnu.org>
>>> Cc: seldon@eskimo.com, monnier@iro.umontreal.ca, emacs-devel@gnu.org
>>> Date: Tue, 06 Dec 2022 17:37:09 -0500
>>> 
>>> What are the advantages of using shr to look at web pages?
>>
>> It's a package that was written specifically to render HTML.
>
> And the fact that being written in Emacs Lisp, it doesn't require
> external dependencies like lynx or w3m.

It depends on libxml, though.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

Re: Package "luwak"

[Yuchen Pei]

On Wed 2022-11-30 18:51:03 -0500, Richard Stallman wrote:

> By contrast, the name "luwak" is unhelpful.  Wikipedia says "luwak" is
> a species of civet; civets are not cats, but fairly close to cats.  If
> you know this, it might help you remember "luwak" when you think of
> "lynx".  But that relationship is remote and strained.  And it won't
> help people who might find luwak useful find out about it.

The naming is explained in the README / package description:

> lynx dump -> feline excretion -> Kopi Luwak

Judging from your talk at EmacsConf, I guess you would still prefer a
different name because this would still fall under word play.

>
> Let's change the package name now to something that will help users
> find out about it and use it.
>
> My suggestion for the name is `lynx'.  That name fits it, since it is
> an interface to lynx.  And anyone who knows about `lynx' will guess
> what it does.

The development of this package is not affiliated with that of lynx in
any way, though it is a lynx client, so I think calling it lynx can
create confusion.

> Does anyone have a better name to suggest?

Common naming for this sort of clients are `lynx-mode` or `lynx.el`. It
is less interesting though but I can live with that. OTOH people can
search for "web browser" in the `list-packages` interface, and it will
hit descriptions.

Best,
Yuchen

-- 
Please join me to support free software for a free society:
https://appeal.fsf.org

PGP Key: 47F9 D050 1E11 8879 9040  4941 2126 7E93 EF86 DFD0
          <https://ypei.org/assets/ypei-pubkey.txt>

Re: Package "luwak"

[Yuchen Pei]

On Fri 2022-12-02 18:49:39 -0800, T.V Raman wrote:

> Richard Stallman <rms@gnu.org> writes:
>
>
> 1+.
>
> As a likely potential user, I looked at that package and didn't see
> anything there that would improve EWW after a quick review;

This package is not intended as a candidate to merge with EWW. It is a
web browser of its own. Stefan Monnier's rationale for merging it with
EWW is that ideally we should have a backend and a frontend systems for
web browsing, the backend retrieves over the net and parses the
response. The frontend decides how to display the parsed results. It is
conceptually a good idea, but requires a lot of work. For starters, lynx
-dump as a backend provides a plaintext version of the html. It seems to
have a long history and widely used for web scraping. There's no dom or
structure, but only links.

Best,
Yuchen

-- 
Please join me to support free software for a free society:
https://appeal.fsf.org

PGP Key: 47F9 D050 1E11 8879 9040  4941 2126 7E93 EF86 DFD0
          <https://ypei.org/assets/ypei-pubkey.txt>

Re: Package "luwak"

[Yuchen Pei]

On Wed 2022-11-30 18:51:03 -0500, Richard Stallman wrote:

> Does anyone have a better name to suggest?

How about `elynxer'? The prefix e stands for emacs or emacs-lisp, and
the suffix er stands for "web browser".

It also sounds a bit liker elixir which is fun.

Best,
Yuchen

-- 
Please join me to support free software for a free society:
https://appeal.fsf.org

PGP Key: 47F9 D050 1E11 8879 9040  4941 2126 7E93 EF86 DFD0
          <https://ypei.org/assets/ypei-pubkey.txt>

Re: Package "luwak"

[Stefan Monnier]

> Does anyone have a better name to suggest?

Luwak sounds perfect to me.
Can we stop wasting time trying to rename things?


        Stefan

Re: Package "luwak"

[Tim Cross]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> Does anyone have a better name to suggest?
>
> Luwak sounds perfect to me.
> Can we stop wasting time trying to rename things?
>

Inf+1

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > > lynx dump -> feline excretion -> Kopi Luwak

  > Judging from your talk at EmacsConf, I guess you would still prefer a
  > different name because this would still fall under word play.

I like wordplay in general, but this wordplay is so obscure that no one
will get the joke, so it isn't very funny.

For names of packages, the priority is to help users find ahwt is
useful to them.  "Luwak" is no good for that, so we should not use
that name.  If we want this ackage in Emacs, we should rename it to a
name that will help users find it and remember it.

That dosn't mean we can't keep this wordplay at all.
We could say, somewhere in the package description, that

  This package was originally called "luwak", making a juke
  lynx dump -> feline excretion -> Kopi Luwak

That would not interfere with higher priorities such as
discocerability and memorability.

But I think that it won't be sufficient to enable most Emacs users to
get the joke.  That would require some more explanation.  But there's
no harm in adding more explanation.

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > > Does anyone have a better name to suggest?

  > Luwak sounds perfect to me.

I think you are not being serious.  If you are, can you explain why
you consider "luwak" better than "lynx"?

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > How about `elynxer'? The prefix e stands for emacs or emacs-lisp, and
  > the suffix er stands for "web browser".

That is better than Luwak.  However, I think plain 'lynx' would be
clearer to users.

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Philip Kaludercic]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > > Does anyone have a better name to suggest?
>
>   > Luwak sounds perfect to me.
>
> I think you are not being serious.  If you are, can you explain why
> you consider "luwak" better than "lynx"?

Honestly, neither of the two is that useful for a modern user
(ie. someone who hasn't investigated text-mode web browsers).  I still
think of the wild cat...  That being said, the discussion has been had
over and over again, and it always results in more or less the same
people stating more or less the same preferences, which is a waste of
time because it is entirely unproductive.

If this is to be resolved in the long term, I think a naming guideline
or recommendations (for ELPA) should be written down.  What are a few
good examples everyone can agree on, what are a few bad examples, and
what makes them bad or good.  What attributes does a good name have
(e.g. easy to remember, easy to write, unique, leaning on convention and
experience, etc.).

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > > I think you are not being serious.  If you are, can you explain why
  > > you consider "luwak" better than "lynx"?

  > Honestly, neither of the two is that useful for a modern user
  > (ie. someone who hasn't investigated text-mode web browsers).

I think you are bluring the question of a name for the package with
the entirely separate question of whether that package is useful.

Since the package is a text-based browser, you may be right that it
isn't useful for people who are not interested in text-based browsers.
Does it follow that it is of no use, that we should not install it?

That could be so -- I don't have a conclusion -- but name calling
(like "a modern user") doesn't prove it.

  > If this is to be resolved in the long term, I think a naming guideline
  > or recommendations (for ELPA) should be written down.

To establish a guideline is a good idea.  The way to come up with a
good guideline is based on some examples.  Thus, I think it is useful
to discuss this example and handle it in a good way.  That will get us
started on more general guidelines.


-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Philip Kaludercic]

Richard Stallman <rms@gnu.org> writes:

> That could be so -- I don't have a conclusion -- but name calling
> (like "a modern user") doesn't prove it.

I can assure you that "modern user" wasn't meant to be a form of name
calling, just a reference to the fact that most people, even Emacs users
have at best little experience with text-based browsers.

>   > If this is to be resolved in the long term, I think a naming
>   > guideline
>   > or recommendations (for ELPA) should be written down.
>
> To establish a guideline is a good idea.  The way to come up with a
> good guideline is based on some examples.  Thus, I think it is useful
> to discuss this example and handle it in a good way.  That will get us
> started on more general guidelines.

Do you have a few examples?

RE: [External] : Re: Package "luwak"

[Drew Adams]

> > That could be so -- I don't have a conclusion -- but name calling
> > (like "a modern user") doesn't prove it.
> 
> I can assure you that "modern user" wasn't meant to be a form of name
> calling, just a reference to the fact that most people, even Emacs users
> have at best little experience with text-based browsers.

If "most people" is meant then consider saying
that, instead of "modern" people or users.

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > I can assure you that "modern user" wasn't meant to be a form of name
  > calling, just a reference to the fact that most people, even Emacs users
  > have at best little experience with text-based browsers.

Maybe you are right about that.  (I wish you'd said it more
straightforwardly.)  Maybe the package proposed under the name Luwak
is not actually worth adding to Emacs.  I don't know enough about it
to have an opinion about that, but I think we should consider that
question, rather than adding every package anyone suggests.

But that's a different question.  We can learn by looking at whether
Luwak is a good name, on the assumption we would install it in Emacs,
whether or not we actually decide to do that.

  > > To establish a guideline is a good idea.  The way to come up with a
  > > good guideline is based on some examples.  Thus, I think it is useful
  > > to discuss this example and handle it in a good way.  That will get us
  > > started on more general guidelines.

  > Do you have a few examples?

I have seen examples of packages we added in the past, whose names seem to
me chosen without attention to how helpful they are to users.
But I mostly don't remember the specifics without something to remind me.

Someone recently mentioned the file bs.el.  That is a rather unhelpful name.
If you don't know what it does, can you guess?

Anyway, the way to learn to think about this kind of issue is to pick
some existing packages, imagine hypothetically that we were going to
give them new names, and discuss possible suggestions.

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: Package "luwak"

[Jean Louis]

* Philip Kaludercic <philipk@posteo.net> [2022-12-15 00:36]:
> Richard Stallman <rms@gnu.org> writes:
> 
> > That could be so -- I don't have a conclusion -- but name calling
> > (like "a modern user") doesn't prove it.
> 
> I can assure you that "modern user" wasn't meant to be a form of name
> calling, just a reference to the fact that most people, even Emacs users
> have at best little experience with text-based browsers.

Many Emacs packages are peculiar for specific tasks, they are there
because they are useful to some people, not to majority of people
and not because they are popular.

Scientists lead the world and they use their peculiar and very much
not popular software. By using Emacs use is empowered. This is for
the reason that Emacs users are not limited to boundaries of
popularity.

Popularity obviously has constraints. When one wish to accommodate
modern users, that would mean to follow the mainstream and
mainstream direction for modern users is strictly private business
of largest companies such as Google or Apple.

Even the opinion of what modern users may like is not truly an
opinion but it is implanted information by what major companies
dictate onto those users.

Regarding text based browsers, since I found them 1999, I still use
them today in 2022. Of course not always, but definitely sometimes.

Very often I use `lynx´ with the `--head´ switch just to find out if
specific long file exists on the server. Being in a country where
Internet bandwidth is not so cheap spares my money! Same applies to
websites with a lot of pictures and distractive elements where I am
interested in the main context.

Then there is use for `lynx´ and `elinks´ and `luwak´, EWW and Emacs
`w3m´ programs where I wish to read some paywall websites. Many of
those websites rely on the modern web browsers to run Javascript.
And instead of disabling Javascript in a GUI browser I tend to use
Emacs `w3m´ or EWW or luwak now when I know about it.

And how about the speed? Sometimes we need text and information in
speedy manner. 

Then sometimes I like to take text from a website and re-format it
for personal notes. Text browser help in doing so.

Especially within Emacs, any kind of text browser is helpful as it
directly brings the text from Internet into the editor, and that
makes it useful.

-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

In support of Richard M. Stallman
https://stallmansupport.org/

Re: Package "luwak"

[Alfred M. Szmidt]

Isn't this sorta tackling the wrong problem?

The main issue that I think Richard wants to tackle is
discoverability.  So while luwak as a name is is meaningless, doesn't
mean that it in itself is a bad name -- it is unique, and easy to
remeber.

In most Emacs files there is the Keywords: field, which lists some
"things that this does".  This would be a better way to make things
discoverable.  If we take a simple example of what I would consider
bad discoverability, it would be eww and shr.

eww.el has:
  Keywords: html

shr.el has:
  Keywords: html

They do do "html" -- but are totally different in what they _actually_
do.  I was looking if there was a command, say package-list-keywords,
package-apropos or apropos-package, but didn't find anything.


Even for rmail, gnus, and mule which are all mail clients and the
names mostly meaningless we have:

rmail.el:
;; Keywords: mail

gnus.el:
;; Keywords: news, mail

mh-eh.el
;; Keywords: mail

While slightly better, but as a user slightly meaningless (what does
"mail" mean? is it a MTA? MUA? does it parse SMTP? mbox files?)
mail-utils.el also has `mail' as a keyword.  A crude grep for `mail'
as a keyword results in 108 files, and I don't think we have 108 mail
readers. :-)


Instead of trying to find "perfect" names (that both explain what a
program does (by "program" i mean something a user directly interactes
with, e.g., rmail, org-mode, gnus, or eww), but are also easy to
remeber and unique); wouldn't it make more sense to have another
field, or extend the Keywords field?  In the case of eww and luwak, or
even a links or lynx mode, the Keywords: (or whatever) could
explicitly say "web-browser".

A command like apropos-package could then list those, or some other
solution.  We have variables like browse-url-browser-function which
could take use of this as well in some form, where "things" could be
marked as "suitable" for use in this variable.

That gets rid of all discussion of "that isn't a very good name of a
package", and instead makes it easier to discover things that might be
useful.

RE: [External] : Re: Package "luwak"

[Drew Adams]

> The main issue that I think Richard wants to tackle is
> discoverability.  So while luwak as a name is is meaningless, doesn't
> mean that it in itself is a bad name -- it is unique, and easy to
> remeber.
> 
> In most Emacs files there is the Keywords: field, which lists some
> "things that this does".  This would be a better way to make things
> discoverable....
>
> I was looking if there was a command, say package-list-keywords,
> package-apropos or apropos-package, but didn't find anything...
> 
> Instead of trying to find "perfect" names (that both explain what a
> program does (by "program" i mean something a user directly interactes
> with, e.g., rmail, org-mode, gnus, or eww), but are also easy to
> remeber and unique); wouldn't it make more sense to have another
> field, or extend the Keywords field?  In the case of eww and luwak, or
> even a links or lynx mode, the Keywords: (or whatever) could
> explicitly say "web-browser".
> 
> A command like apropos-package could then list those, or some other
> solution.  We have variables like browse-url-browser-function which
> could take use of this as well in some form, where "things" could be
> marked as "suitable" for use in this variable.
> 
> That gets rid of all discussion of "that isn't a very good name of a
> package", and instead makes it easier to discover things that might be
> useful.

Good initiative.  The point truly is _discoverability_.

A file/library name can sometimes help discovery a little bit,
but that effect and possibility is quite limited (doesn't
scale).  Nothing wrong with trying to have better library
names, but that isn't much of a solution for the problem of
finding libraries.

Discoverability can no doubt be improved.  And a file's
`Commentary' field `Keywords' is, yes, one thing that could
be leveraged better.  (That would of course depend on library
maintainers actually adding and maintaining that field.)

There's library `finder.el'.  It's all about leveraging field
`Keywords' in library Commentary.  It has commands
`finder-by-keyword' and `finder-list-keywords'. (For some
reason those seem to be exactly the same, without aliasing -
the first just calls the second.  Why?).

There could/should be a `finder-apropos', which lets you use
multiple keywords.  (You mention the need for a command such
as `apropos-package', which would be essentially the same
thing.)

`finder.el' is quite old.  It could be beefed up and perhaps
extended to do some more (?) for "packages".  IOW, it, - and
field `Keywords' - could use some more love.

We might imagine some semi-automatic way (suggestions for a
maintainer) of coming up with relevant keywords, to help make
the job of adding and maintaining `Keywords' easier and more
fruitful.

___

[ My library `finder+.el' enhances `finder.el' in a few minor
  ways, such as using its own syntax table, adding font-lock,
  and naming the `*Finder*' buffer after the library instead
  (so you can have multiple such buffers) - e.g., for library
  `cl-lib' the buffer is named `*Commentary, cl-lib*'.

  https://www.emacswiki.org/emacs/download/finder%2b.el
]

Re: [External] : Re: Package "luwak"

[Alfred M. Szmidt]

   Good initiative.  The point truly is _discoverability_.

Thanks, but this is only useful someone picks up the taks, and does
something ... ideas are easy.

   We might imagine some semi-automatic way (suggestions for a
   maintainer) of coming up with relevant keywords, to help make
   the job of adding and maintaining `Keywords' easier and more
   fruitful.

Back in the day, we (the GNU nitpicker gang --
https://cvs.savannah.gnu.org/viewvc/womb/gnits/gnits.texi?view=log)
had a glossary of sorts, with tables of commonly used options, and
what nots.  It wasn't semi-automatic, very much manual, and even so
more so that hard to know about it.

Maybe this could be automated somewhat, a M-x identify-package like
command that asks the user stuff, and tries to match things against
some file.

Ofcourse, these are just random ideas, someone would have to implement
something ... alas I'm out of general time to do this.

Re: Package "luwak"

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > The main issue that I think Richard wants to tackle is
  > discoverability.

The two goals I am concerned about are (1) discoverability and (2)
understandability.  "Luwak" is bad on both of those dimensions.  Even
if you read about the package 6 months ago, you are likely not to
recall what it does, because the name does nothing to help you recall.

                      So while luwak as a name is is meaningless, doesn't
  > mean that it in itself is a bad name -- it is unique, and easy to
  > remeber.

In my view, meaningless names are bad names.  Mere uniqueness is not
enough to make up for the lack of discoverability and
understandability.

There can be unusual cases where a unique, meaningless name might
suffice: for a feature that is so important and so often used that
everyone will encounter the feature every week, perhaps we can expect
all users to learn and remember what it does.

In those cases, perhaps it won't matter than the name is meaningless.

But such cases are unusual.

  > Instead of trying to find "perfect" names

By saying "perfect", that sentence criticizes bvy exaggeration.  I am
not calling for "perfect" names.  What I said is nothing like that.

I am calling for an effort to look for a name that is at lesst
somewhat helpful -- rather than adopting a totally unhelpful name
that has no reasons to prefer it.

    wouldn't it make more sense to have another
  > field, or extend the Keywords field?

That might be a good idea; how about posting a proposal
in a different thread with a different subject line.

But that wouldn't eliminate the benefit of a name that is meaningful
and helpful.


-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)