Re: Node.js and it's future in debian

Re: Node.js and it's future in debian

[Patrick Ouellette]

On Wed, May 02, 2012 at 12:13:49PM -0500, Jonathan Nieder wrote:
> Patrick Ouellette wrote:
> 
> > Likewise I can argue the number of people with installed ham radio systems
> > is a good reason NOT to change the current situation.
> 
> You can, yes.  But how does that move things forward at all?

I never said it did.  Clearly both sides have valid reasons to
not change.  Equally clear to me is one side ignored policy
and created an issue to attempt to force a resolution they hope
will be in their favor rather than solve the issue first.

> 
> This is not supposed to be a popularity contest.  I mentioned the
> large pile of scripts because _every one of them would have to be
> changed_ to have a working system.  By contrast, there are two
> configuration files mentioned so far that refer to /usr/sbin/node.

The scripts (on either side) could be changed with a scripted change. 

If it is so simple to change the configuration files for the ham radio
users, why has not a Node.js person put forth code to do this and advocated
it on debian-hams and linux-hams? (The "patch" sent does not address
automatically updating anything)

I've discussed it with other ham radio operators.  They shudder at the
thought of changing the name because of the possible issues that will
come up.

> 
> [...]
> > If it were "easy" to get an exception, why has this not already happened?
> 
> Because you did not ask for one.  Instead you have been wasting time
> arguing and defending against an opponent you seem to assume is not
> going to care or listen to you.

The Node.js people apparently didn't ask for one either
pot - kettle - black

As for the last line, if I thought the opponent did not care or was
not going to listen I would not waste the time putting forth my
position.  Since it would seem that is where we are, I won't
continue to waste my time.

Here is my proposal:
Node.js people, put forth a reasonable and workable plan to allow
hundreds or thousands of ham radio users to transition from 
/usr/sbin/node to /usr/sbin/axnode, including reliable shell scripts
to verify all the files on the system are identified and allowed to
be patched or manually modified.  You created the situation, you 
provide the manpower to resolve it in the way you prefer.


Pat - NE4PO

Re: Node.js and it's future in debian

[Jonathan Nieder]

Patrick Ouellette wrote:

>                                   (The "patch" sent does not address
> automatically updating anything)

This is very funny.  You are putting patch in quotes, but it[1] was a
real patch.  It did not automatically update anything because it was
meant to be a simple patch to get work started.  I volunteered to
write further patches once I got feedback on that one, and then I got
no direct feedback on it, just occasional passive-aggressive comments
like the above.

[...]
>> Because you did not ask for one.  Instead you have been wasting time
>> arguing and defending against an opponent you seem to assume is not
>> going to care or listen to you.
>
> The Node.js people apparently didn't ask for one either
> pot - kettle - black

The pot is presumably me.  But I am not a Node.js person.  The Debian
Node.js package maintainers have been friendly and helpful when I
contacted them, and they seem to be willing to help work on including
a nodejs command upstream and modifying Debian packages to use it.
They also seemed willing to remove the node command from nodejs if a
consensus in the project were to develop that that was needed.

Just for completeness, I should mention that Jaime Robles on the
ham radio package maintenance side has been friendly, too.

You have made it clear that you are more interested in punishing
people than in making wheezy better, so I don't think we have anything
left to talk about.  I'll contact the technical committee and leave
this in their capable hands.

Thanks for some useful clarifications,
Jonathan

[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=20;bug=614907

Re: Node.js and it's future in debian

[David Ranch]

[-- Attachment #1: Type: text/plain, Size: 3696 bytes --]


Hello Gentlemen,

I thought I'd chime in since the linux-hams@vger list was added to the 
thread and give one Packet HAM's perspective.  Specifically, if one 
proposal is to rename the long existing /usr/sbin/node binary to 
/usr/sbin/axnode, why couldn't the "new guy" node.js binary be renamed 
to something like /usr/sbin/nodejs?  The later seems more of a 
reasonable proposal.

 From my experience, many MANY Linux hams have customized scripts that 
startup some very elaborate HAM systems.  For many, these scripts 
weren't written by them and the changing of the node command could be 
very difficult for some.  The other aspect is if this change came into a 
package update that could impact production systems in VERY remote 
sites.  This could cause all kinds ugliness that can be easily avoided.

I can appreciate Debian's goal to keep things moving forward but I'd 
argue that a binary name of "/usr/sbin/nodejs" would be a lot more 
informative with the two additional characters than just calling it 
"node" (and disrupting a well known binary name for us Linux packet hams).

--David
KI6ZHD


On 05/02/2012 01:04 PM, Patrick Ouellette wrote:
> On Wed, May 02, 2012 at 12:13:49PM -0500, Jonathan Nieder wrote:
>> Patrick Ouellette wrote:
>>
>>> Likewise I can argue the number of people with installed ham radio systems
>>> is a good reason NOT to change the current situation.
>> You can, yes.  But how does that move things forward at all?
> I never said it did.  Clearly both sides have valid reasons to
> not change.  Equally clear to me is one side ignored policy
> and created an issue to attempt to force a resolution they hope
> will be in their favor rather than solve the issue first.
>
>> This is not supposed to be a popularity contest.  I mentioned the
>> large pile of scripts because _every one of them would have to be
>> changed_ to have a working system.  By contrast, there are two
>> configuration files mentioned so far that refer to /usr/sbin/node.
> The scripts (on either side) could be changed with a scripted change.
>
> If it is so simple to change the configuration files for the ham radio
> users, why has not a Node.js person put forth code to do this and advocated
> it on debian-hams and linux-hams? (The "patch" sent does not address
> automatically updating anything)
>
> I've discussed it with other ham radio operators.  They shudder at the
> thought of changing the name because of the possible issues that will
> come up.
>
>> [...]
>>> If it were "easy" to get an exception, why has this not already happened?
>> Because you did not ask for one.  Instead you have been wasting time
>> arguing and defending against an opponent you seem to assume is not
>> going to care or listen to you.
> The Node.js people apparently didn't ask for one either
> pot - kettle - black
>
> As for the last line, if I thought the opponent did not care or was
> not going to listen I would not waste the time putting forth my
> position.  Since it would seem that is where we are, I won't
> continue to waste my time.
>
> Here is my proposal:
> Node.js people, put forth a reasonable and workable plan to allow
> hundreds or thousands of ham radio users to transition from
> /usr/sbin/node to /usr/sbin/axnode, including reliable shell scripts
> to verify all the files on the system are identified and allowed to
> be patched or manually modified.  You created the situation, you
> provide the manpower to resolve it in the way you prefer.
>
>
> Pat - NE4PO
> --
> To unsubscribe from this list: send the line "unsubscribe linux-hams" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[-- Attachment #2: Type: text/html, Size: 4852 bytes --]

Re: Node.js and it's future in debian

[David Ranch]

Hello Gentlemen,

I thought I'd chime in since the linux-hams@vger list was added to the 
thread and give one Packet HAM's perspective.  Specifically, if one 
proposal is to rename the long existing /usr/sbin/node binary to 
/usr/sbin/axnode, why couldn't the "new guy" node.js binary be renamed 
to something like /usr/sbin/nodejs?  The later seems more of a 
reasonable proposal.

 From my experience, many MANY Linux hams have customized scripts that 
startup some very elaborate HAM systems.  For many, these scripts 
weren't written by them and the changing of the node command could be 
very difficult for some.  The other aspect is if this change came into a 
package update that could impact production systems in VERY remote 
sites.  This could cause all kinds ugliness that can be easily avoided.

I can appreciate Debian's goal to keep things moving forward but I'd 
argue that a binary name of "/usr/sbin/nodejs" would be a lot more 
informative with the two additional characters than just calling it 
"node" (and disrupting a well known binary name for us Linux packet hams).

--David
KI6ZHD


On 05/02/2012 01:04 PM, Patrick Ouellette wrote:
> On Wed, May 02, 2012 at 12:13:49PM -0500, Jonathan Nieder wrote:
>> Patrick Ouellette wrote:
>>
>>> Likewise I can argue the number of people with installed ham radio systems
>>> is a good reason NOT to change the current situation.
>> You can, yes.  But how does that move things forward at all?
> I never said it did.  Clearly both sides have valid reasons to
> not change.  Equally clear to me is one side ignored policy
> and created an issue to attempt to force a resolution they hope
> will be in their favor rather than solve the issue first.
>
>> This is not supposed to be a popularity contest.  I mentioned the
>> large pile of scripts because _every one of them would have to be
>> changed_ to have a working system.  By contrast, there are two
>> configuration files mentioned so far that refer to /usr/sbin/node.
> The scripts (on either side) could be changed with a scripted change.
>
> If it is so simple to change the configuration files for the ham radio
> users, why has not a Node.js person put forth code to do this and advocated
> it on debian-hams and linux-hams? (The "patch" sent does not address
> automatically updating anything)
>
> I've discussed it with other ham radio operators.  They shudder at the
> thought of changing the name because of the possible issues that will
> come up.
>
>> [...]
>>> If it were "easy" to get an exception, why has this not already happened?
>> Because you did not ask for one.  Instead you have been wasting time
>> arguing and defending against an opponent you seem to assume is not
>> going to care or listen to you.
> The Node.js people apparently didn't ask for one either
> pot - kettle - black
>
> As for the last line, if I thought the opponent did not care or was
> not going to listen I would not waste the time putting forth my
> position.  Since it would seem that is where we are, I won't
> continue to waste my time.
>
> Here is my proposal:
> Node.js people, put forth a reasonable and workable plan to allow
> hundreds or thousands of ham radio users to transition from
> /usr/sbin/node to /usr/sbin/axnode, including reliable shell scripts
> to verify all the files on the system are identified and allowed to
> be patched or manually modified.  You created the situation, you
> provide the manpower to resolve it in the way you prefer.
>
>
> Pat - NE4PO
> --
> To unsubscribe from this list: send the line "unsubscribe linux-hams" in
> the body of a message tomajordomo@vger.kernel.org
> More majordomo info athttp://vger.kernel.org/majordomo-info.html

RE: Node.js and it's future in debian

[Marius Petrescu]

Now I really subscribe di David's arguments but I would add something else.

Node.js is a server-side JavaScript environment, so the "js" is a
legitimate, even useful addition.
Node on the other hand is exactly what its name states: a communication
node.
axnode is not correct since it also supports pure tcp/ip, netrom and rose
(and virtual any other protocol which allows a character oriented p2p link).

So why get the javascript out of a java script environment and put a
restrictive title to a communication node tool?
And by the way, most know node.js as node.js, so the confusion will be even
greater if you change the name.

Marius, YO2LOJ

-----Original Message-----
From: linux-hams-owner@vger.kernel.org
[mailto:linux-hams-owner@vger.kernel.org] On Behalf Of David Ranch
Sent: Thursday, May 03, 2012 18:34
To: Patrick Ouellette
Cc: Jonathan Nieder; node@packages.debian.org; nodejs@packages.debian.org;
debian-hams@lists.debian.org; linux-hams@vger.kernel.org
Subject: Re: Node.js and it's future in debian


Hello Gentlemen,

I thought I'd chime in since the linux-hams@vger list was added to the
thread and give one Packet HAM's perspective.  Specifically, if one proposal
is to rename the long existing /usr/sbin/node binary to /usr/sbin/axnode,
why couldn't the "new guy" node.js binary be renamed to something like
/usr/sbin/nodejs?  The later seems more of a reasonable proposal.

 From my experience, many MANY Linux hams have customized scripts that
startup some very elaborate HAM systems.  For many, these scripts weren't
written by them and the changing of the node command could be very difficult
for some.  The other aspect is if this change came into a package update
that could impact production systems in VERY remote sites.  This could cause
all kinds ugliness that can be easily avoided.

I can appreciate Debian's goal to keep things moving forward but I'd argue
that a binary name of "/usr/sbin/nodejs" would be a lot more informative
with the two additional characters than just calling it "node" (and
disrupting a well known binary name for us Linux packet hams).

--David
KI6ZHD

Re: Node.js and it's future in debian

[Patrick Ouellette]

On Thu, May 03, 2012 at 07:17:08PM +0300, Marius Petrescu wrote:
> 
> Now I really subscribe di David's arguments but I would add something else.
> 
> Node.js is a server-side JavaScript environment, so the "js" is a
> legitimate, even useful addition.
> Node on the other hand is exactly what its name states: a communication
> node.
> axnode is not correct since it also supports pure tcp/ip, netrom and rose
> (and virtual any other protocol which allows a character oriented p2p link).
> 
> So why get the javascript out of a java script environment and put a
> restrictive title to a communication node tool?
> And by the way, most know node.js as node.js, so the confusion will be even
> greater if you change the name.

I certainly agree with your comments.  The confusion can arise because
Node.js is the project's name, node is one of the names of a binary in the
Node.js package.

Pat

Re: Node.js and it's future in debian

[Patrick Ouellette]

Thanks for the comments!

On Thu, May 03, 2012 at 08:33:43AM -0700, David Ranch wrote:
> 
> From my experience, many MANY Linux hams have customized scripts
> that startup some very elaborate HAM systems.  For many, these
> scripts weren't written by them and the changing of the node command
> could be very difficult for some.  The other aspect is if this
> change came into a package update that could impact production
> systems in VERY remote sites.  This could cause all kinds ugliness
> that can be easily avoided.
> 

One issue here seems to be a perception problem.  The non-ham folks
look at the ax25 how to or similar documentation and see that node is
"usually called from ax25d" or a similar superserver.  This leads them
to believe that the command is only started from one or two "easily 
identified and modified" scripts.

The Node.js side argues that there are many books, tutorials, and scripts
that tell the user to run "node <script>" and so the interaction of the user
and node.js' node command is direct and uncontrollable.

> I can appreciate Debian's goal to keep things moving forward but I'd
> argue that a binary name of "/usr/sbin/nodejs" would be a lot more
> informative with the two additional characters than just calling it
> "node" (and disrupting a well known binary name for us Linux packet
> hams).

I completely agree, but apparently Node.js' upstream has changed the name 
once previously (apparently from a similar problem) and while acknowledging
the name is generic and a poor choice refuses to consider another change.
(According to what I can tell from the Debian discussion.  I have not
talked to Node.js upstream personally.)

FWIW, the commands are NOT in the same path.  The ham radio node command
is /usr/sbin/node and the Node.js command is /usr/bin/node

Debian does NOT allow binaries with the same name to have different functions.
Unfortunately "first come, first served" is not the policy either.


Pat 

Re: Node.js and it's future in debian

[Jonathan Nieder]

Hi again,

Patrick Ouellette wrote:

> I completely agree, but apparently Node.js' upstream has changed the name 
> once previously (apparently from a similar problem) and while acknowledging
> the name is generic and a poor choice refuses to consider another change.
> (According to what I can tell from the Debian discussion.  I have not
> talked to Node.js upstream personally.)

The working title of Node.js was "server" for a few weeks, before
anyone was using it.  When I looked that up in order to understand
what the name "node" was about (in the spirit of [1]) I mentioned this
factoid without making the context sufficiently clear, and I'm sorry
about that[2].

To avoid banging heads against the wall too quickly: I think there are
two aspects that it would be productive to discuss:

 1. Which package should use the name "node" in the long term?  What
    can we do to ensure that happens eventually?

    (My answer is that I hope that neither uses the name "node" in
    the long term.)

 2. What should be the state in Debian's upcoming "wheezy" release to
    provide a smooth upgrade path and not surprise users too much?

    (My answer is that configuration needs to be smoothly migrated:

     - ax25d.conf by the ax25-tools package
     - inetd configuration by the node package
     - other configuration by the sysadmin, after they are notified
       through a note in node's NEWS.Debian file (shown by
       apt-listchanges) and the release notes

    I also would hope that wheezy can include a /usr/sbin/node file
    that prints a message to help people notice they are still using
    it and calls /usr/sbin/axnode, but that is still under discussion.

    Likewise, the Node.js needs some migration to ensure scripts
    installed by Debian packages and from outside use the new name.
    I would hope that wheezy can include a /usr/bin/node synonym for
    compatibility until usage of it fades away, but that is still
    under discussion.)

If you disagree with the long-term goal or have ideas for a smoother
migration, that could be useful.

Hope that helps,
Jonathan

[1] http://wiki.debian.org/WhyTheName
[2] https://lists.debian.org/debian-devel/2011/11/msg00377.html

Re: Node.js and it's future in debian

[Gordon JC Pearc e]

On 03/05/12 19:51, Jonathan Nieder wrote:
. Which package should use the name "node" in the long term?  What
 >      can we do to ensure that happens eventually?
 >
 >      (My answer is that I hope that neither uses the name "node" in
 >      the long term.)

Exactly.  It's a stupidly common term, probably only slightly better 
than calling it "program".

I don't buy into this idea that changing it will break all sorts of 
legacy scripts.  If they are that fragile and undocumented, *get rid of 
them*!  Undocumented fragile code is a liability.  Kill it now, under 
"known" circumstances, and fix it up properly.

-- 
Gordonjcp MM0YEQ

RE: Node.js and it's future in debian

[Hans-Peter F. Oeste]

I thought I'd also chime in with one aspect that perhaps hasn't been
articulated.  This could have a huge impact in 'Health and Public Safety', a
place where many hams use the tools and the debian distribution as an
appliance.  The node and ax25 package is used worldwide providing
communications infrastructure that may people rely on in times of disaster
and emergency, be it a forwarding of medical information to gain insight to
people health problems to local floods or fires to events like Katrina.
Hams often provide the backbone to the communications for these events.

That they treat the Debian distribution as an appliance says that they trust
the distribution to not break down, and this is something to be proud of.
That they use it as an appliance should not be considered a stupidity,.
Consider how many of us nowadays make changes to the engine in their
vehicles themselves.  Not many, it's done by enthusiasts.  The package node
and its attendant  packages is an engine for many ham operators.  So please,
be very careful in your consideration in making changes. The ramifications
could affect you personally inadvertently.

If you are going to muck with it, please consider a phased in approach that
gives warning of potentially at least year, if not longer, so that warning
is give and propagated to the mentors that these ham operators use, so that
customizations can be 'fixed' and any migrations and mitigation is
flawlessly executed.  This may mean that you yourselves may be asked to help
troubleshoot issues/problems that arise.  Please be patient.

While I am a ham, I am also a mediocre programmer, so I can see both sides
of this argument.  In addition, I've used Linux longer than I've been a ham.

Hans Oeste
VE7OES

Re: Node.js and it's future in debian

[David Ranch]

Me again..

AX.25 in Linux has been around for a long time so I can excuse it's 
overly generic "node" name purely based upon it's age but..


> The working title of Node.js was "server" for a few weeks, before
> anyone was using it.

Wow.. that's horrible!  Obviously we don't want stuff like that to 
happen.  Please also consider that all this ISN'T just a *Debian* 
problem.  Its a Linux distro-wide problem.  It's groups like this that 
form and guide aspects of all Linux distributions consistency and 
considering Debian's wide influence, changes here will surely trickle 
into other distributions over time.

It's also worth touching on that I personally appreciate the work that
Patrick Ouellette has done on maintaining the HAM packages for Debian. 
Like always, there are never enough appreciative people in this world 
but once removed / renamed /etc, I'm SURE a lot of people will come out 
of the woodwork to bitch about it.  A *lot* of people use Debian and 
Debian-related distributions with Packet radio.


>   1. Which package should use the name "node" in the long term?  What
>      can we do to ensure that happens eventually?
>
>      (My answer is that I hope that neither uses the name "node" in
>      the long term.)

I personally think that some of it SHOULD be a first come, first served 
thing.  I previously mentioned in the previous email that all of the 
various scripts that people run could/would break.  Probably no big deal 
to many of us on *this* list but trust me, I know a few Linux packet 
people who would be seriously lost because of these changes.

Also consider the tons of documentation, notably the AX.25 HOWTO that 
would be impacted and I highly doubt it would get updated (hasn't been 
since 2001) to reflect these changes.  It's not like things have needed 
to change all that much - http://tldp.org/HOWTO/AX25-HOWTO/


>   2. What should be the state in Debian's upcoming "wheezy" release to
>      provide a smooth upgrade path and not surprise users too much?

Is Node.js a new addition to Debian?  Again, I side with first come 
first served.


>      I also would hope that wheezy can include a /usr/sbin/node file
>      that prints a message to help people notice they are still using
>      it and calls /usr/sbin/axnode, but that is still under discussion.
>
>      Likewise, the Node.js needs some migration to ensure scripts
>      installed by Debian packages and from outside use the new name.
>      I would hope that wheezy can include a /usr/bin/node synonym for
>      compatibility until usage of it fades away, but that is still
>      under discussion.)

If for some reason Debian feels that longstanding packages and their 
well known binary names can be renamed at any given time (I seriously 
disagree with that mentality btw), I'd say then ALSO force the change of 
the "node" in Node.JS name to something more sane.  Don't remove one 
poorly named file for a new poorly named one just because it's new and 
shiny.

--David

Re: Node.js and it's future in debian

[Ray Wells]

All,

Although not currently running a node (although I do run aprs with 
kernel ax25), I started using kernel ax25 back around 1994 (ish) with a 
floppy that booted into ramdisk and provided two ports, one kiss and the 
other baycom. I used node back then but moved on. The disk still works 
but I can't for the life of me remember the logon password :-( Back then 
(kernel 2.0.29) it was ax25-utils-<version> that provided kernel ax25 
apps and tools, in the days prior to ax25-apps and ax25-tools being 
released as separate packages.

When I last ran a node, along with a bbs, I didn't use the node package. 
Instead, I used fpac which provided the facilities I needed - ax25, rose 
and netrom. For many, many years, right up to just last week, I compile 
libax25, ax25-apps & ax25-tools from sources to try to keep as up to 
date as possible with fixes, etc. Although not a user of the packages I 
still appreciate the efforts of the maintainers who produce packages for 
the benefit of users/potential users who don't care to compile from sources.

As for a name change for node, it seems strange that such a long-running 
application should be forced to change simply because of some 
Johnny-come-lately. I think Debian policy in this regard should be 
reviewed to prevent repetition of the current stalemate in the years to 
come. FWIW, I've only ever used Debian or a Debian derivative.

Ray vk2tv

On 04/05/12 14:09, David Ranch wrote:
>
> Me again..
>
> AX.25 in Linux has been around for a long time so I can excuse it's 
> overly generic "node" name purely based upon it's age but..
>
>
>> The working title of Node.js was "server" for a few weeks, before
>> anyone was using it.
>
> Wow.. that's horrible!  Obviously we don't want stuff like that to 
> happen.  Please also consider that all this ISN'T just a *Debian* 
> problem.  Its a Linux distro-wide problem.  It's groups like this that 
> form and guide aspects of all Linux distributions consistency and 
> considering Debian's wide influence, changes here will surely trickle 
> into other distributions over time.
>
> It's also worth touching on that I personally appreciate the work that
> Patrick Ouellette has done on maintaining the HAM packages for Debian. 
> Like always, there are never enough appreciative people in this world 
> but once removed / renamed /etc, I'm SURE a lot of people will come 
> out of the woodwork to bitch about it.  A *lot* of people use Debian 
> and Debian-related distributions with Packet radio.
>
>
>>   1. Which package should use the name "node" in the long term?  What
>>      can we do to ensure that happens eventually?
>>
>>      (My answer is that I hope that neither uses the name "node" in
>>      the long term.)
>
> I personally think that some of it SHOULD be a first come, first 
> served thing.  I previously mentioned in the previous email that all 
> of the various scripts that people run could/would break.  Probably no 
> big deal to many of us on *this* list but trust me, I know a few Linux 
> packet people who would be seriously lost because of these changes.
>
> Also consider the tons of documentation, notably the AX.25 HOWTO that 
> would be impacted and I highly doubt it would get updated (hasn't been 
> since 2001) to reflect these changes.  It's not like things have 
> needed to change all that much - http://tldp.org/HOWTO/AX25-HOWTO/
>
>
>>   2. What should be the state in Debian's upcoming "wheezy" release to
>>      provide a smooth upgrade path and not surprise users too much?
>
> Is Node.js a new addition to Debian?  Again, I side with first come 
> first served.
>
>
>>      I also would hope that wheezy can include a /usr/sbin/node file
>>      that prints a message to help people notice they are still using
>>      it and calls /usr/sbin/axnode, but that is still under discussion.
>>
>>      Likewise, the Node.js needs some migration to ensure scripts
>>      installed by Debian packages and from outside use the new name.
>>      I would hope that wheezy can include a /usr/bin/node synonym for
>>      compatibility until usage of it fades away, but that is still
>>      under discussion.)
>
> If for some reason Debian feels that longstanding packages and their 
> well known binary names can be renamed at any given time (I seriously 
> disagree with that mentality btw), I'd say then ALSO force the change 
> of the "node" in Node.JS name to something more sane.  Don't remove 
> one poorly named file for a new poorly named one just because it's new 
> and shiny.
>
> --David
> -- 
> To unsubscribe from this list: send the line "unsubscribe linux-hams" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: Node.js and it's future in debian

[Jonathan Nieder]

David Ranch wrote:

> Is Node.js a new addition to Debian?  Again, I side with first come
> first served.

I think one of Debian's most important roles is to help various
dispersed projects to produce code that works well in combination.
Yes, Node.js is a relatively new addition to the world and Debian
(about 3 years old, not part of squeeze).

Re: Node.js and it's future in debian

[DL1SIG]

Hi,

I think using the alternatives system [1] could be a solution. I guess
that node.js and axnode is not often used on the same system. Therefore
in most of the cases the alternatives auto select mechanism will do the
right thing.

In the case that both programs are installed, the node symlink is auto
selected by the priority so I would suggest to set the priority of
axnode higher because it was first there.

Anyone who want to have node.js as node can simply set the alternatives
to manual selection.

This approach can be combined with the long term goal to have no node
binary. You make node a shell script which prints a warning and then
calls the binary selected by the alternatives system.

73 DL1SIG

1. http://wiki.debian.org/DebianAlternatives

Am Thu, 03 May 2012 13:05:21 -0400
schrieb Patrick Ouellette <pouelle@debian.org>:

> Thanks for the comments!
> 
> On Thu, May 03, 2012 at 08:33:43AM -0700, David Ranch wrote:
> > 
> > From my experience, many MANY Linux hams have customized scripts
> > that startup some very elaborate HAM systems.  For many, these
> > scripts weren't written by them and the changing of the node command
> > could be very difficult for some.  The other aspect is if this
> > change came into a package update that could impact production
> > systems in VERY remote sites.  This could cause all kinds ugliness
> > that can be easily avoided.
> > 
> 
> One issue here seems to be a perception problem.  The non-ham folks
> look at the ax25 how to or similar documentation and see that node is
> "usually called from ax25d" or a similar superserver.  This leads them
> to believe that the command is only started from one or two "easily 
> identified and modified" scripts.
> 
> The Node.js side argues that there are many books, tutorials, and
> scripts that tell the user to run "node <script>" and so the
> interaction of the user and node.js' node command is direct and
> uncontrollable.
> 
> > I can appreciate Debian's goal to keep things moving forward but I'd
> > argue that a binary name of "/usr/sbin/nodejs" would be a lot more
> > informative with the two additional characters than just calling it
> > "node" (and disrupting a well known binary name for us Linux packet
> > hams).
> 
> I completely agree, but apparently Node.js' upstream has changed the
> name once previously (apparently from a similar problem) and while
> acknowledging the name is generic and a poor choice refuses to
> consider another change. (According to what I can tell from the
> Debian discussion.  I have not talked to Node.js upstream personally.)
> 
> FWIW, the commands are NOT in the same path.  The ham radio node
> command is /usr/sbin/node and the Node.js command is /usr/bin/node
> 
> Debian does NOT allow binaries with the same name to have different
> functions. Unfortunately "first come, first served" is not the policy
> either.
> 
> 
> Pat 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-hams"
> in the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [Pkg-javascript-devel] Node.js and it's future in debian

[Emilien Klein]

2012/5/4 DL1SIG <dl1sig@hw42.dyndns.org>:
> Hi,
>
> I think using the alternatives system [1] could be a solution. I guess
> that node.js and axnode is not often used on the same system. Therefore
> in most of the cases the alternatives auto select mechanism will do the
> right thing.
>
> In the case that both programs are installed, the node symlink is auto
> selected by the priority so I would suggest to set the priority of
> axnode higher because it was first there.
>
> Anyone who want to have node.js as node can simply set the alternatives
> to manual selection.
>
> This approach can be combined with the long term goal to have no node
> binary. You make node a shell script which prints a warning and then
> calls the binary selected by the alternatives system.
>
> 73 DL1SIG
>
> 1. http://wiki.debian.org/DebianAlternatives
>

Seems to me like a good idea.
   +Emilien

Re: Node.js and it's future in debian

[Patrick Ouellette]

On Fri, May 04, 2012 at 02:31:29PM +0200, DL1SIG wrote:
> 
> I think using the alternatives system [1] could be a solution. I guess
> that node.js and axnode is not often used on the same system. Therefore
> in most of the cases the alternatives auto select mechanism will do the
> right thing.
>

While the alternatives system could be used in this manner, we are
back to a policy type issue. From the wiki:

"The Debian alternatives system creates a way for several programs 
that fullfill the same or similar functions to be listed as alternative 
implementations that are installed simultaneously but with one 
particular implementation designated as the default."

The packages in question are worlds apart in what they do.  They are not
alternatives or substitutes.

Pat

Re: Node.js and it's future in debian

[walter harms]

Am 02.05.2012 22:04, schrieb Patrick Ouellette:
> On Wed, May 02, 2012 at 12:13:49PM -0500, Jonathan Nieder wrote:
>> Patrick Ouellette wrote:
>>
>>> Likewise I can argue the number of people with installed ham radio systems
>>> is a good reason NOT to change the current situation.
>>
>> You can, yes.  But how does that move things forward at all?
> 
> I never said it did.  Clearly both sides have valid reasons to
> not change.  Equally clear to me is one side ignored policy
> and created an issue to attempt to force a resolution they hope
> will be in their favor rather than solve the issue first.
> 
>>
>> This is not supposed to be a popularity contest.  I mentioned the
>> large pile of scripts because _every one of them would have to be
>> changed_ to have a working system.  By contrast, there are two
>> configuration files mentioned so far that refer to /usr/sbin/node.
> 
> The scripts (on either side) could be changed with a scripted change. 
> 
> If it is so simple to change the configuration files for the ham radio
> users, why has not a Node.js person put forth code to do this and advocated
> it on debian-hams and linux-hams? (The "patch" sent does not address
> automatically updating anything)
> 
> I've discussed it with other ham radio operators.  They shudder at the
> thought of changing the name because of the possible issues that will
> come up.
> 
>>
>> [...]
>>> If it were "easy" to get an exception, why has this not already happened?
>>
>> Because you did not ask for one.  Instead you have been wasting time
>> arguing and defending against an opponent you seem to assume is not
>> going to care or listen to you.
> 
> The Node.js people apparently didn't ask for one either
> pot - kettle - black
> 
> As for the last line, if I thought the opponent did not care or was
> not going to listen I would not waste the time putting forth my
> position.  Since it would seem that is where we are, I won't
> continue to waste my time.
> 
> Here is my proposal:
> Node.js people, put forth a reasonable and workable plan to allow
> hundreds or thousands of ham radio users to transition from 
> /usr/sbin/node to /usr/sbin/axnode, including reliable shell scripts
> to verify all the files on the system are identified and allowed to
> be patched or manually modified.  You created the situation, you 
> provide the manpower to resolve it in the way you prefer.
> 
>

 hi par,
the problem of scripts would dissolve with time. I am worried about node.js itself.
I do think js is a problem not a solution, my experience with webservices is like
that described with other on the net .. slow, ugly, unmaintainable in the long run.
Unfortunately there is now a tendency in programming that everything should go web
and that understanding of code and algorithms is not longer important since fast
computers will make every rubbish look like usable.

My only hope is that even perl was unable to displace shell programming and this
time it will not work either.

re,
 wh