Naming problem for the "Falcon Programming Language" in Ubuntu.
Matt Zimmerman
mdz at ubuntu.com
Thu Feb 7 11:01:05 GMT 2008
I've read over this briefly (have not reviewed the IRC logs in particular)
and would appreciate a response from Dennis with his view.
I've copied in the Ubuntu Technical Board, which is tasked with resolving
disputes like this.
On Mon, Jan 14, 2008 at 01:55:22AM +0100, Giancarlo Niccolai wrote:
> Hello;
>
> I'd like to report a problem that is concerning the naming of two
> packages that are being concurrently racing for inclusion in the very
> next version of Ubuntu.
>
> I would argue that the way MOTU have managed the whole situation is
> questionable under the "Ubuntu Code of Conduct" that we all have signed.
>
> I came to Ubuntu as I wanted to issue my first official release of the
> Falcon Programming Language
>
> http://www.falconpl.org
>
> under the distro which I most respect, for its philosophy of mutual
> help and support which is reflected in the Code of Conduct. The very
> heart of my project is exactly "Respect for Developers", which drove
> my will to write a language written not "to parse logs" or "to keep
> finger warm", but exactly to help people write better programs, and to
> help applications to be better applications.
>
> When I prospected my project in #ubuntu-motu I have been positively
> accepted by the community; so I started writing a module; I searched
> Debian and Ubuntu repositories, and the net in search of debian
> packages named Falcon. Having not found them, I went for "falcon" as
> package name. People in MOTU told me that the naming of the package
> was fine, although some of them knew of a packaging utility written in
> python that was called "falcon.py". I offered to change the name of
> the package to falconpl, that was also the name of the site, but I
> have been told (sorry if I call people by nick), by pochu, persia and
> others that the name was ok, and to proceed. I have talked also with
> Minghua and many others.
>
> So I uploaded the package on December 6 2007 in Revu. I got comments
> through #ubuntu-motu channel, and I updated each time to fix problems
> reported by MOTU reviewers. The records are in
>
> https://bugs.launchpad.net/ubuntu/+bug/174470
> http://revu.tauware.de/details.py?package=falcon
>
> On December the 20th 2007, a Motu (possibly Imbradon, but I am not
> sure), told me that the project called "falcon" had possibly a
> /usr/bin/falcon instance, which the Falcon programming language has
> too (it's the main interpreter, as /usr/bin/python or /usr/bin/ruby).
> As that was a tool used by some Motus to handle packages, although
> there was no package being submitted up to that date, it would have
> been a pity if there was a name clash in future, so I should have
> sorted out the situation.
>
> I immediately mailed the author, sending him this mail:
>
> -------- Original Message --------
> Message-ID: <476AD25B.9060001 at falconpl.org>
> Date: Thu, 20 Dec 2007 21:36:43 +0100
> From: Giancarlo Niccolai <gc at falconpl.org>
> User-Agent: Thunderbird 2.0.0.6 (X11/20071022)
> MIME-Version: 1.0
> To: dennis at kaarsemaker.net
> Subject: "Falcon" name in namespace.
> X-Enigmail-Version: 0.95.0
> OpenPGP: id=12030B86; url=keyserver.ubuntu.com
> Content-Type: text/plain; charset=ISO-8859-1
> Content-Transfer-Encoding: 7bit
>
> hello,
>
> I am the author of an open source software called "The Falcon
> Programming Language". I am submitting a package to Ubuntu, and I
> have been warned about possible namespace collisions.
>
> The main concern is about the "falcon" binary file, which is the
> compiler/interprter of the language (like python), and other binary
> names as libfalcon_engine.so.
>
> Up to date, several MOTUs have checked and reported there is no
> current namespace clash. I am also willing to call my package
> "falconpl", which is also the name of the site:
>
> http://www.falconpl.org
>
> However, it is necessary that we get in contact so that we can see if
> there may be some name clash now or in the future, in order to avoid it.
>
> I am quite open to any proposal.
>
> I usually hang around in #ubuntu-motu or in #falconpl.
>
> Bests,
> Giancarlo Niccolai.
> =====================================================================
>
> This my reply on its reply (I think it's useless to repaste it twice)
>
> ------- Original Message --------
> Message-ID: <476BF4B6.8060901 at falconpl.org>
> Date: Fri, 21 Dec 2007 18:15:34 +0100
> From: Giancarlo Niccolai <gc at falconpl.org>
> User-Agent: Thunderbird 2.0.0.9 (Windows/20071031)
> MIME-Version: 1.0
> To: Dennis Kaarsemaker <dennis at kaarsemaker.net>
> Subject: Re: "Falcon" name in namespace.
> References: <476AD25B.9060001 at falconpl.org>
> <1198230256.6930.9.camel at mirage>
> In-Reply-To: <1198230256.6930.9.camel at mirage>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> Content-Transfer-Encoding: 7bit
>
>
> Dennis Kaarsemaker wrote:
> > On do, 2007-12-20 at 21:36 +0100, Giancarlo Niccolai wrote:
> >
> >
> >> I am the author of an open source software called "The Falcon
> >> Programming Language". I am submitting a package to Ubuntu, and I
> >> have been warned about possible namespace collisions.
> >>
> >> The main concern is about the "falcon" binary file, which is the
> >> compiler/interprter of the language (like python), and other binary
> >> names as libfalcon_engine.so.
> >>
> >
> > The places where I see collisions are
> >
> > 1) The package name
> > 2) /usr/{bin,share,share/doc}/falcon
> > 3) Manpage name
> >
> >
> >> Up to date, several MOTUs have checked and reported there is no
> >> current namespace clash. I am also willing to call my package
> >> "falconpl", which is also the name of the site:
> >>
> >> http://www.falconpl.org
> >>
> >
> > Correct, 'my' falcon has not been uploaded to Ubuntu yet for a variety
> > of reasons, so no conflict should exist.
> >
> >
> >> However, it is necessary that we get in contact so that we can see if
> >> there may be some name clash now or in the future, in order to avoid it.
> >>
> >> I am quite open to any proposal.
> >>
> >
> > Renaming the package would unfortunately only remove two naming
> > conflicts (package name and /usr/share/doc). The other conflict can only
> > be resolved by renaming one of the /usr/bin/falcon files.
> >
> > I don't think it's really necessary to do that actually since both
> > falcons serve two distinct niches which will have little overlap, if
> > any. So my suggestion is to only rename the package and not the binary
> > unless you don't mind renaming your /usr/bin/falcon to /usr/bin/falconpl
> >
> > What do you think?
> >
> Renaming the package doesn't cause any trouble. Renaming /usr/bin/falcon
> would cause a bit of harassment as the automated build environment
> creates a lot of dependencies (i.e. windows build system requires
> falcon.exe to be there, double click shortcuts, ini files), not to talk
> about #!/usr/bin/falcon headers on scripts.
>
> About the overlap, if the language gets moment it may be virtually on
> any machine, so it's a thing that we should try to work on.
>
> Bests,
> Giancarlo.
>
> P.s. I should be around somewhere this night as "jonnymind" on IRC.
>
> =====================================================================
>
> I had no other communication. Just, I noticed Imbradon's notice on
> revu's entry
> tonight:
> "
> There is already another project in the archives named "falcon" this
> source and
> binarys will likely have to be renamed ( and make sure both are
> co-installable )
> "
>
> Checking the "falcon" project,
> https://launchpad.net/ubuntu/+source/falcon
>
> I see that the package has been loaded on "2008-01-11"; it has been
> reviewed
> and fixed in the same day, and then immediately published.
>
> I have been in #MOTU about every day, and I offered full and open
> co-operation
> both to the MOTU and to this project author; My project was discussed
> and seen
> by the persons involved in release, yet, they just created a clean
> package and
> immediately published it without even bothering notifying this fact.
>
> The chat session of this night is even more enlightening:
>
> gen 13 20:01:04 jonnymind imbradon:I read your notice about "there
> is already a package named falcon.
> gen 13 20:01:38 jonnymind However, the notice is imprecise.
> gen 13 20:02:24 * amarillion
> (n=martijn at 212-123-137-98.ip.telfort.nl) è entrato in #ubuntu-motu
> gen 13 20:02:30 jonnymind the bug "needs packaging" for my project
> has been opened in 2007, while the other package with the same name
> was opened in 2008.
> gen 13 20:03:18 * cassidy` si è disconnesso (Remote closed the
> connection)
> gen 13 20:03:24 jonnymind Moreover, I was negotiating with those
> person about the package and binary names. We were talking about what
> to do, then the conversation stopped and the other package has been
> started.
> gen 13 20:03:33 jonnymind I think we should discuss a bit the
> namespace question.
> gen 13 20:03:57 * cassidy (n=cassidy at dhansak.collabora.co.uk) è
> entrato in #ubuntu-motu
> gen 13 20:06:02 jonnymind And also, frankly decide what to do
> basing on an objective criterion.
> gen 13 20:08:16 jonnymind imbradon:?
>
> <rip>
>
> en 13 23:18:06 ScottK jonnymind: Conflicts is more usually used for
> packages that provide equivalent functionality.
> gen 13 23:18:37 jonnymind ScottK: that is absolutely true.
> gen 13 23:19:13 ScottK jonnymind: Have you discussed this with
> imbrandon?
> gen 13 23:19:26 * jekil (n=alessand at 151.82.16.79) è entrato in
> #ubuntu-motu
> gen 13 23:19:38 pochu jonnymind: (I'm not blaming you, just in case
> I'm not expressing well)
> gen 13 23:20:46 Nafallo Seveas and imbrandon.
> gen 13 23:20:52 * Knightlust si è disconnesso (Read error: 104
> (Connection reset by peer))
> gen 13 23:20:59 TheMuso s/c
> gen 13 23:21:03 TheMuso ugh damn keyboard
> gen 13 23:21:32 jonnymind ScottK: No, I have discussed with the
> original author of the other falcon.
> gen 13 23:21:45 pochu That's Seveas.
> gen 13 23:21:46 jonnymind Well, I have started the discussion and
> replied to its e-mail.
> gen 13 23:21:55 jonnymind In which he stated he wasn't going to
> make a package.
> gen 13 23:22:07 ScottK jonnymind: Right, and he didn't
> gen 13 23:22:15 pochu (which is true, the package was done by
> imbrandon)
> gen 13 23:22:24 pochu ScottK: you beat me
> gen 13 23:22:33 jonnymind Ah.
> gen 13 23:22:42 * pochu is eating pizza, so he's not that fast ;)
> gen 13 23:22:50 ScottK jonnymind: My concern right now (not having
> a great interest in either package) is that if you upload your package
> as falcon to Debian, then it's going to cause trouble between Ubuntu
> and Debian that it would be better to avoid.
> gen 13 23:22:52 jonnymind So Imbradon knew there was a falcon
> package made in december.
> <rip>
> gen 13 23:23:15 Seveas packages for 'the other falcon' have existed
> since 2006
> <rip>
> gen 13 23:25:52 jonnymind Seveas: again; the package here is named
> falconpl now.
> gen 13 23:25:59 jonnymind That closes the question.
> gen 13 23:26:30 Nafallo both packages use /usb/bin/falcom? :-)
> gen 13 23:26:36 Nafallo s/m/n
> gen 13 23:26:42 jonnymind I do.
>
> <rip>
>
> gen 13 23:59:19 pochu And out of curiosity: a package needs two
> ACKs to be uploaded to Universe... who did ACK falcon other than
> imbrandon?
> gen 13 23:59:22 jonnymind Nafallo: I am not changing 6 system
> installation scripts and 300 doc pages for this.
> gen 13 23:59:30 Nafallo pochu: *sigh* you probably know exactly
> what I mean, so why even argue about it? :-)
> gen 13 23:59:36 jonnymind Someone will find a way to package falcon
> when it is included in the other distros.
> gen 13 23:59:57 Seveas pochu, could be persia, I did the final
> checks with them
> gen 14 00:00:26 jonnymind I started packaging from ubuntu because I
> beleived in ubuntu philosophy of respect.
> gen 14 00:00:28 jonnymind ...
> gen 14 00:00:37 pochu Seveas: I'd be surprised if it was persia
> since persia was reviewing jonnymind's falcon package.
> gen 14 00:01:42 pochu And since there's no REVU upload for it to
> look at, nor ACK in the needs-packaging bug...
> gen 14 00:01:51 * keffie_jayx si è disconnesso (Connection timed out)
> gen 14 00:04:25 Nafallo pochu: I just check my logs. persia :-)
> gen 14 00:04:41 * Kmos (n=gothicx at unaffiliated/kmos) è entrato in
> #ubuntu-motu
> gen 14 00:04:47 pochu Crap.
> gen 14 00:04:59 * pochu kicks persia :)
>
> ...
>
> =================================================
>
> I have a "falconpl" ubuntu package with the "Conflict: falcon" entry
> in debian/control sitting on my hard disk;
> it was a mediated solution suggested and required by some enlighted
> MOTUs, and I gladly accepted;
> but in the moment I was going to upload it I received this notification:
>
> ------ Original Message --------
> Return-path: <bounces at canonical.com>
> Envelope-to: gc at niccolai.cc
> Delivery-date: Sun, 13 Jan 2008 14:55:44 -0800
> From: Nafallo Bjälevik <nafallo at magicalforest.se>
> <rip headers>
> X-Generated-By: Launchpad (canonical.com)
>
>
> This package source and binary needs to be renamed. We already have
> falcon. Please make sure they are co-installable.
>
> ======================================================
>
> This was right in the middle of our discussion about name clashes and
> Conflict: field.
>
>
>
> The point is: I have opened a falcon package in Debian, and I will do
> it with other distros as well, as Fedora,
> Suse and Mandrake. The falcon programming language is already
> installed in production environments under Windows and
> MacOS; I would have gladly mediated some solution for Ubuntu
> "/usr/bin/falcon" naming problem, but this
> possibility has been willfully snatched away for unknown reasons. As
> the space for a mediation has been
> removed by this behavior which is intentionally breaking the Code of
> Conduct, I am forced to resume the
> technical reasons why I ask to be granted the /usr/bin/falcon program
> name:
>
> 1) Falcon P.L. is a stand-alone binary application; falcon.py is a
> python script proxied by an alias command.
> 2) The "falcon" (langauge) command is common to several architectures,
> including Windows and MacOS, some
> of which have special handling of application registration which is
> part of our project.
> 3) Falcon codebase is several hundred thousands lines and growing.
> Many of this lines refer to the name of
> the main interpreter. falcon.py script is a relatively small
> application.
> 4) Falcon language (to be found as "falcon" command) is currently used
> in production environments, and changing
> its name may affect third party applications.
> 5) Falcon language documentation is several hundreds pages long, and
> reviewing it in search of the main interpreter
> name would be an extensive activity.
> 6) Falcon language copyright goes since 2004, with activity actually
> started in 2003. Falcon.py has been started in
> 2005-2006.
> 7) Falcon is known as script language name in Wikipedia
> (http://en.wikipedia.org/wiki/Scripting_language)
> 8) Falcon language Packages for other distros (i.e.) RPM will be
> organized having /usr/bin/falcon.
> 9) Although user count between the two projects is currently not
> directly comparable, the number of potential
> users of a repository utility for a specific subset of debian
> distributions and a multiplatform scripting
> language is not a match.
> 10) Users already using Falcon would be impacted by the main
> interpreter name change, as it is the base to
> launch scripts and other applications, falcon.py script is only
> conventionally summoned with /usr/bin/falcon
> at the command line by end-users.
>
> It seems there isn't a single technical/objective reason why the
> falcon.py script project should be granted
> /usr/bin/falcon name, except for a direct involvement of some of the
> MOTUs in its release.
>
> I openly admit that I have opened a package request on Debiabn with
> the name "falcon" as I knew about this
> package being just snapped in without an open discussion. I am still
> ready to negotiate the package name,
> as *I* do beleive in the principles written on the Code of Conduct,
> that I have signed, and I don't want
> to cause unnecessary problems in a project that I still regard as a
> bright example of what mankind should
> be one day.
>
> Yet, the "falcon" command will be present in many distributions, and
> it is now available in many environments
> and O/S. I hope the community discuss openly and peacefully this
> argument considering the technical
> aspects of the question.
>
> Best regards,
> Giancarlo Niccolai.
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> --
> ubuntu-devel mailing list
> ubuntu-devel at lists.ubuntu.com
> Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
>
--
- mdz
More information about the ubuntu-devel
mailing list