FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Ubuntu > Gobuntu Developer

 
 
LinkBack Thread Tools
 
Old 04-18-2008, 11:18 PM
Tim Dobson
 
Default Rethinking Gobuntu

Zeth wrote:
> On 18/04/2008, Tim Dobson <personalwebsite@army.com> wrote:
>> Do you think people would be more willing to work on a freedom
>> emphasising Ubuntu derivative, if things like launchpad[1]
>> [1] I hate to bring it up like a big stick in every debate, but for some
>> it is understandably a big issue.
>
> As far as I know there is nothing in Launchpad that requires a Gobuntu
> desktop to have any type of proprietary software on it. So Launchpad
> is not really relevant to this discussion; as I wrote before on this
> topic, we have to think about what we mean when we say "Launchpad is
> not free" or "Launchpad is not open source":
> https://lists.ubuntu.com/archives/gobuntu-devel/2007-July/000054.html
>
> According to my "smoke and mirrors" conspiracy theory, Launchpad is
> no less free than the Debian homepage, Gentoo homepage or FSF
> homepage.

the fsf's CMS is plone. debian seems to be html but it's not like you
interact with debian's homepage. I don't know much about gentoo so they
are out of my radar.

launchpad is an amazing piece of software. the idea that the whole
community revolves around it while it is kept secret undermines the
goals of gobuntu and free software. If the system to report bugs doesn't
take freedom seriously, then why should anyone else?

>> (Aside) I have been using Gobuntu Hardy now for a fairly long time and I
>> can confirm that annoyingly controversial pieces of software like
>> Firefox, Thunderbird & Sunbird are all available easily.
>
> But not as easily as Epiphany, a more free and generally more superior
> browser, and Evolution a superior mail reader. If one is not
> influenced by branding and spin, why would you want to use slow
> ugly-looking XUL apps over fast beautiful GTK native ones?

because this is a highly opinionated point. it really depends what
people want in a webbrowser.
In general what i want in a webbrowser is adblock plus, tabs and the
knowledge that if i was forced to use a different OS or WIMP that it
would have minimum impact on what software i used.

in reference to thunderbird this is also highly opinionated point.

whether people should or should not want to use firefox or thunderbird,
they will. and whether these are good or bad products, the code to both
of these projects is GPL licenced. however both of them contain blobs
and trademark issues.

> (Sunbird I will concede as it is one of the only free software desktop
> applications that has working CalDaV support).

I know. it's nice (actually i use lighting but it's the same codebase)

>> (Still Aside) Unlike some free software developers and activists, I *do*
>> suffer from latest and greatest syndrome, which is one of the reasons I
>> like Gobuntu.
>
> Yup agreed. Do we actually know that significantly more people are
> using gnewsense than Gobuntu? I doubt it. Show me the numbers. Show me
> the numbers

I actually don't think it really matters which has more users. Gnewsense
currently has people who hold freedom over everything, and gobuntu users
in general would like to be where the gnewsense users are, but would
like to cutting edge free software.

>> Unfortunately, I don't think I would recommend it particularly either,
>
> But is that more a hardware thing? That currently Gobuntu and
> gnewsense do not currently works in many computers or real life
> situations? Rather than Gobuntu is untrustworthy or not free enough?

Not at all. In fact the only hardware issues *I* have had relate to wifi
drivers and since for me, one doesn't exist, free *or* non free, I'm not
devastated. Graphics drivers re: 3d acceleration are also problematic,
but don't reduce productivity.
I wouldn't recommend gnewsense because I think it is vastly outdated - i
couldn't install stuff extensions in burning dog, because it was at 1.5
- ok i should have used epiphany!
I wouldn't recommend gobuntu, not because it doesn't work, but because
of basically all the things we have been talking about.

I love the idea of Free Software Ubuntu. If it *has* to be a derivative
of the main branch, so be it, but I believe, we need to take a close
look at Ubuntu and take freedom seriously.



--
www.tdobson.net
----
If each of us have one object, and we exchange them, then each of us
still has one object.
If each of us have one idea, and we exchange them, then each of us now
has two ideas. - George Bernard Shaw

--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 02:25 AM
Matthew Flaschen
 
Default Rethinking Gobuntu

Zeth wrote:
> On 18/04/2008, Tim Dobson <personalwebsite@army.com> wrote:
>> Do you think people would be more willing to work on a freedom
>> emphasising Ubuntu derivative, if things like launchpad[1]
>> [1] I hate to bring it up like a big stick in every debate, but for some
>> it is understandably a big issue.
>
> As far as I know there is nothing in Launchpad that requires a Gobuntu
> desktop to have any type of proprietary software on it. So Launchpad
> is not really relevant to this discussion; as I wrote before on this
> topic, we have to think about what we mean when we say "Launchpad is
> not free" or "Launchpad is not open source":
> https://lists.ubuntu.com/archives/gobuntu-devel/2007-July/000054.html
>
> According to my "smoke and mirrors" conspiracy theory, Launchpad is
> no less free than the Debian homepage, Gentoo homepage or FSF
> homepage.

Actually, the key Debian server software (such as packages.debian.org)
is free, and /being used/ by Ubuntu.

Matt Flaschen

--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 02:33 AM
"Benj. Mako Hill"
 
Default Rethinking Gobuntu

<quote who="Toni Ruottu" date="Fri, Apr 18, 2008 at 03:19:48PM +0300">
>
> A Debian source code package consists of multiple files:
>
> - dsc file that contains meta-data
> - tar.gz file published by the upstream
> - diff file that does distribution specific changes
>
> See below url for example
> http://archive.ubuntu.com/ubuntu/pool/main/h/hello/hello_2.2-2.dsc
>
> Maybe it would be possible to add a fourth file (say hello_2.2.sanitize)
> defining rules for performing sanitation to make the source fit what
> ever policy. The sanitation file could be optional.

(a) Changes to the source package format are not be taken lightly and
will break compatibility with Debian.

(b) I don't think this would actually solve any real problems.

We can already produce multiple binary packages. We could build, from
the Firefox source package, both FF and un- or rebranded packages. That
would work fine and produce binary packages that Make People Happy but I
think it would still be unacceptable to many because the source packages
would contain unacceptable content.

Regards,
Mako



--
Benjamin Mako Hill
mako@atdot.cc
http://mako.cc/

Creativity can be a social contribution, but only in so far
as society is free to use the results. --GNU Manifesto
--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 02:41 AM
"Benj. Mako Hill"
 
Default Rethinking Gobuntu

<quote who="Dave Crossland" date="Thu, Apr 17, 2008 at 04:36:32PM +0100">
> So the FSF agrees that Mozilla doesn't restrict software freedom with
> its copyright policy, because the icons in question are not functional
> parts of the software.
>
> But none of this is directly relevant to Gobuntu, because Gobuntu's
> goal has always been to exclude bitstreams without freedom-to-modify.
> That includes Firefox artwork and FSF manuals.

That's right, this is a problem withUbuntu and Gobuntu living up its
own stated goals and standards, not a problem with the FSF.

> > > > We don't have the resources in Canonical to maintain two
> > > > different packages of the Firefox code. Nobody else has stepped
> > > > up to do the work.
> >
> > > I take the view that Ubuntu should have a single "meta source"
> > > package that generates Ubuntu and Gobuntu source and binary
> > > packages in the small number of packages in main/universe that
> > > require this, and that security patches therefore happen once.
> >
> > If you could demonstrate that in action for Firefox, the kernel,
> > openoffice and some of the others, then we could seriously consider
> > adopting it.
>
> I guess this means you don't have the resources in Canonical to
> engineer "meta source" packages, and are asking me to step up.

It also sounds like this would require being knee-deep in Launchpad (and
any associated NDAs and such) as well which, I can only imagine, most
Gobuntu activists developers would hesitate to do.

Regards,
Mako


--
Benjamin Mako Hill
mako@atdot.cc
http://mako.cc/

Creativity can be a social contribution, but only in so far
as society is free to use the results. --GNU Manifesto
--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 09:38 AM
Toni Ruottu
 
Default Rethinking Gobuntu

On Fri, 2008-04-18 at 22:33 -0400, Benj. Mako Hill wrote:
> <quote who="Toni Ruottu" date="Fri, Apr 18, 2008 at 03:19:48PM +0300">
> >
> > A Debian source code package consists of multiple files:
> >
> > - dsc file that contains meta-data
> > - tar.gz file published by the upstream
> > - diff file that does distribution specific changes
> >
> > See below url for example
> > http://archive.ubuntu.com/ubuntu/pool/main/h/hello/hello_2.2-2.dsc
> >
> > Maybe it would be possible to add a fourth file (say hello_2.2.sanitize)
> > defining rules for performing sanitation to make the source fit what
> > ever policy. The sanitation file could be optional.
>
> (a) Changes to the source package format are not be taken lightly and
> will break compatibility with Debian.
>
> (b) I don't think this would actually solve any real problems.
>
> We can already produce multiple binary packages. We could build, from
> the Firefox source package, both FF and un- or rebranded packages. That
> would work fine and produce binary packages that Make People Happy but I
> think it would still be unacceptable to many because the source packages
> would contain unacceptable content.

Actually it would work just to put all the sanitation files into one web
directory on the repository server. Then it would be possible to check
against the corresponding package name, if such file exists for a
certain package. It requires polling, but I assume it would not be done
that often.

This would still require changes into apt-get. I'm not sure, if that is
equally bad. Changes needed would be:

- checking for sanitation files and applying their rules

- command line option for ignoring them
(apt-get --original source hello)

- command line option for forcing them
(apt-get --sanitize source hello)

- a way of specifying the default behavior
(allowing Gobuntu to sanitize by default and Ubuntu to not)


--Toni



--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 10:24 AM
"Dave Crossland"
 
Default Rethinking Gobuntu

On 16/04/2008, Stanislas Breton <stanislas_breton@yahoo.co.uk> wrote:
> >> This seems to have no sense to me.
> >
> > Please consider the difference between preventing a user from doing
> > something, inviting them to do something, acting as if it is
> > acceptable to do it, and doing it for them.
> > ...
> > Gobuntu contains proprietary software in its source code.
> > ...
> > If a distribution of free software invites users to install
> > proprietary software (such as Ubuntu's "Commercial Software" option in
> > the "Add/Remove Programs" tool, or Firefox's default "proprietary
> > plugin installer" feature) that indirectly recommends proprietary
> > software which is also not an okay thing to do.
> >
> > If a distribution of free software includes references to proprietary
> > software, such as in its manual (as many "community edition" GPL
> > programs do) or such as OpenBSD's ports recipies for downloading say
> > the Opera browser, that also assumes proprietary software is
> > acceptable to use.
>
> Does the same logic apply to the official builds of gcc and GNU Emacs
> for Windows?

Let's repeat the logic and find out:

Please consider the difference between preventing a user from doing
something, inviting them to do something, acting as if it is
acceptable to do it, and doing it for them.

Win32 GCC/Emacs contains no proprietary software in its source code.

If a distribution of free software invites users to install
proprietary software, that indirectly recommends proprietary
software which is also not an okay thing to do. Emacs and GCC do
not invite users to install proprietary software.

If a distribution of free software includes references to proprietary
software, such as in its manual (as many "community edition" GPL
programs do) or such as OpenBSD's ports recipies for downloading say
the Opera browser, that also assumes proprietary software is
acceptable to use. GCC and Emacs do not have crippled "community
editions" under the GPL with proprietary upgrades; there is no
package system for downloading proprietary modes or extensions.

Emacs and GCC manuals do mention Windows though, and are built
to run on Windows. Does this invite people to use proprietary software?

Including references to Windows in manuals is not recommending it,
because it is well known by anyone who uses computers,
and typically they are already using it. Opera is different to this; it
is not well known and mentioning it advertises it.

Making free software available for Windows enables people to try
free software applications, understand directly what it means to
have software freedom, and thus motivate them to switch to a
fully free OS. It is true that some may stick with Windows,
thinking "Well, I have these great programs now, and I can keep
my existing Windows programs without too much bother." and so
invites them to _continue_ using the proprietary software they
already use. But over the long term, if they are using free
software and their data is buliding up in free software formats,
they will find it easier to switch. And they are not being invited
to install any _new_ proprietary software than they used yesterday.

Thus, making free software available for Windows, or accomodating
Windows, is helping us to use less proprietary software each day, and
eventually use none. I still use proprietary software (mobile phone,
a laptop with Fedora has firmware, my desktop motherboard can run
LinuxBIOS but I didn't get to soldering on a new PRAM chip for it...),
and I'm writing this from GMail. But I am taking steps each month
to reduce the amount of proprietary software I use.

Is it hypocritical for me to strongly advocate software freedom? Yes,
a little. I'm okay with that, because taking small steps _continually_
is what counts. The only person on the planet I know who doesn't use
_any_ and is thus not a hypocrit is RMS.

The first small step for a Windows user is to install Firefox,
OpenOffice, Inkscape etc. Then try WUBI install of Ubuntu. Then
dualboot Ubuntu. Then remove Windows. Then upgrade Ubuntu to Gobuntu
and then gNewSense. Then buy a motherboard that support a free BIOS.
Then a free phone and portable music player and camera and set top
box.

--
Regards,
Dave

--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 10:43 AM
"Dave Crossland"
 
Default Rethinking Gobuntu

On 18/04/2008, Zeth <theology@gmail.com> wrote:
> On 18/04/2008, Tim Dobson <personalwebsite@army.com> wrote:
> > Do you think people would be more willing to work on a freedom
> > emphasising Ubuntu derivative, if things like launchpad[1]
> > [1] I hate to bring it up like a big stick in every debate, but for some
> > it is understandably a big issue.
>
> As far as I know there is nothing in Launchpad that requires a Gobuntu
> desktop to have any type of proprietary software on it. So Launchpad
> is not really relevant to this discussion;

Launchpad is relevant to this discussion because it is part of the
reason people contribute to gNewSense over Gobuntu, and thus why Mark
is "Rethinking Gobuntu."

> as I wrote before on this
> topic, we have to think about what we mean when we say "Launchpad is
> not free" or "Launchpad is not open source":
> https://lists.ubuntu.com/archives/gobuntu-devel/2007-July/000054.html
>
> According to my "smoke and mirrors" conspiracy theory, Launchpad is
> no less free than the Debian homepage, Gentoo homepage or FSF
> homepage.

IMO that theory confuses web-based applications with web-based
publication of documents; a subtle distinction.

Software freedom is about the freedoms of the users of software. To
know if a program is free or not, one must examine the users of the
software.

For example: A user of a GPLv2 program generally has software freedom,
but if that software came preinstalled on a Set Top Box, the hardware
may trample the user's freedom to modify the software and run it on
that hardware. This gives the user software freedom in theory, but not
in practice; it makes a mockery of the GPL, and is one of the major
prompts for the GPLv3 which prevents users' freedom being trampled
like this. However, prominent members of the Open Source community
feel that this is acceptable and see the protection against it offered
by the GPLv3 as of no value.

When examining the question of software freedom for web apps, we first
must distinguish between software that runs on our own computers which
we control, and that which runs on servers not under our control (eg
the Javascript programs and the PHP programs).

For the clientside (Javascript) programs, it is straightforward that
this part runs on our computers, so a webapp with Javascript that is
not free software is not a free web app. (One webapp I know of,
TiddlyWiki, runs purely clientside and is free software.)

For the serverside programs, it is straightforward to answer "Who are
the users?" with "Anyone who browses the site." But I feel this is an
oversimplification; the software does not run on their computer, so it
is not under their direct control, so the social question of "Who
controls this program?" can never be directly them. It is thus
illuminating to ask, "Who controls the server?" and "Who owns the
data?"

When I don't control the server and I don't own the data, there is no
software freedom issue for me. If I control the server there is a
problem for me indentical to that for the software running on my PC.
If I solely own the data there is a problem if I don't control the
server. If I co-own the data, well, this is an unresolved problem :-)

Okay, let's break that down:

Google's public web search engine makes fair-use copies of public web
pages and so they own their mirror of the web. Google does indeed
control their servers, and Google does have software freedom for the
software that indexes and searches it. When I use that search engine
web app to filter the data owned by them, that does not effect my
software freedom.

Google's private web search engine - which they sell as a "Google
Search Appliance" 1U server - is very different to this; you own your
private Intranet web data, and you control your webservers, so you
ought to control the software that indexes and searches your Intranet.
When I use that search engine web app to filter the data owned by you,
that does not effect my software freedom.

Because Google's public web search engine will index your public site,
you might be tempted to use it to provide search functionality for
your site; http://www.w3.org/ is a good example of this common
practice. This is the same as deploying Google Search Appliance in an
Intranet: There is no difference for me as I browse your site, but
your freedom is being trampled.

Google's word processor web app is different. Google controls their
servers, and has software freedom with the webapp. But if I use it to
write with, my software freedom is being trampled, because my data is
processed by software that I do not control.

(Google Docs isn't antisocial because it doesn't pull people in: It is
easy for Docs users to export their document to ODF, send it to me
where I can work on it with free software running on my computer, and
send it back. But a main purpose of such webapps is collaborative
document authoring, so some other word processor webapps do not offer
such import/export, and require co-authors to use their webapp to edit
the document - Basecamp is a famous one, and I expect the Microsoft
Docs )

Google Blogger is just the same as Docs, and I wouldn't recommend
using it because it tramples your freedom in the same way. However,
while WordPress.com offers a service almost identical to Google
Blogger, it is slightly different: The software is available, from
WordPress.org

If I have a Wordpress.com account, I can download some source code,
hire anyone I like to make changes, and contribute the patch back to
the WordPress developers, who we can suppose will accept it. I am
controlling the server indirectly, so perhaps I have 'indirect'
software freedom..? We might even say that WordPress.com is "open
source."

But to paraphrase Mark Pilgrim, that's like being 'indirectly'
pregnant. "Close doesn't count, and even if people don't notice the
difference today, they'll definitely notice it nine months from now."
The software running on WordPress.com may not be the same as that
available from WordPress.org, and I would still have no idea what
processing is being done with my data.

To have direct control, I will have to install the software on a
server I control, and use it to write with. My software freedom is
thus maintained directly because my data is processed by software
which I control directly. (Popping the stack back to search, that
means running an index on your own server over your own data -
http://www.gnu.org is a good counter example to http://www.w3.org)

What about when I post a comment on your blog, be it at
surname.wordpress.com or surname.org/wordpress? If the input box is
plain and the comment is short, I am passing up a tiny amount of
control over my data, but that is okay because it is so short;
similarly the GPL is not recommended for very short programs, because
the costs of the defence of freedom outweight the benefits. But if the
input box involved a Javascript program, it ought to be free software
Javascript, like TinyMCE. And if the comment is a long one, I ought to
write it on my own blog and link to yours; the 'trackback' mechanism
common in blogging web apps makes this sensible.

If you develop a webapp and make it available under the GPLv3, you
face a copyleft-like problem: Someone else can download your source,
add features, and remove the freedom for their users. If you use a
website for which the clientside software is free, that might not help
all that much, because it might be tied to specific serverside
software.

The Affero GPL seeks to help people in these positions, by requiring
webapps to publish their serverside sourcecode. This hopes to ensure
end-users of a webapp can always install the webapp on their own
server, and that the users of the webapp can re-integrate any changes
made to others' installations that are public-facing. (There is no
requirement for publication, just a view of the user as the web
surfer, not the server administrator.) WordPress and MediaWiki would
do well to upgrade to this license IMO; the excellent webapps from
MySociety already use it, including the one behind
http://petitions.pm.gov.uk/ :-)

OpenClipart.org and Wikipedia are different again, because their data
is not 'owned' by a single (legal) person. The public literally owns
OCAL because its CC-PD, and WP is GFDL so is similar enough. Although
the public do not control the servers as that's not realistic any more
(Once upon a time, there used to be no passwords -
http://www.google.com/search?hl=en&q=stallman+%22there+should+be+no+pass words%22
- but the access control was a social system) the 'indirect freedom'
with WordPress and MediaWiki, please pretty low barriers to
contributing to the work of maintaining the servers, is as good as it
gets.

These centralised repositories have a key benefit to having
archipelagos of clipart archives/encyclopdia wikis: The "network
effect" of aggregated data. Flickr's tag clouds are a great example of
this. Since the data is free culture, its not a problem for different
sites to fork and merge their data (license compatibility permitting).
But that duplication decreases the network effect.

I think "federation," or bidirectional syndication - as seen in IRC,
Jabber and gnutella networks - could provide a solution here. DVCS
like Bazaar and Git nd web services APIs are paving the way for this
kind of thing. There are lots of photo gallerys run with
http://gallery.sf.net but they is currently no way to link up sets of
them for big tag clouds. But I think it could be done; I'd love to see
the FSF hire programmers to work on software R&D that promotes
software freedom but that free software companies don't think is
profitable to do - like it used to.

So. Launchpad. Who owns the data? Right now, Canonical does, and there
is no way to run your own Launchpad.

The Ubuntu community doesn't feel this is appropriation of their data,
or doesn't mind it, nor the lack of indirect freedom, nor their
inability to run their own launchpads. Mark has said that in the long
term he hopes to do this, and the development of Bazaar and tools for
merging in the patches made by other distributions seems promising.

I think to attract people who won't use Launchpad because they
percieve it as proprietary, Canonical needs to make it possible for
them to set up their own lightweight distro infrastructure to have
direct software freedom with, but which Canonical can pull back
changes from... something Affero licensed and based on Bazaar, I'd
guess.

--
Regards,
Dave

--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 11:31 AM
Zeth
 
Default Rethinking Gobuntu

Fascinating stuff, slightly tangential, but fascinating none the less.

On 19/04/2008, Dave Crossland <dave@lab6.com> wrote:
> For the clientside (Javascript) programs, it is straightforward that
> this part runs on our computers, so a webapp with Javascript that is
> not free software is not a free web app.

Most web sites that do too much in Javascript are annoying anyway. The
web should be usable without Javascript. And in the words of Cory
Doctorow, if your site is only in Flash then you suck.

> illuminating to ask, "Who controls the server?" and "Who owns the
> data?"

Well, with my gmail account, I try to use it for mailing lists, rather
than for personal email which is hosted on my own machine and my own
domain name, etc. Things overlap, but the main reason to use gmail is
that it is easier to have lots of (essentially disposable) mailing
list data on Google's servers rather than on my own machine.

However, the answer to owning data is encryption, if email is
encrypted then it does not matter whose server it is stored on. To
everyone else without the passphrase and private key it is just a load
of random junk.

> Because Google's public web search engine will index your public site,
> you might be tempted to use it to provide search functionality for
> your site;

I do this myself. You can set it up also so that Google will cut you
in on any advertising shown on the results page.

> There is no difference for me as I browse your site, but your freedom is being trampled.

Why? Because Google could in theory fix the results? Censor posts it
didn't like? I'm not sure I understand the problem.

> Google Blogger is just the same as Docs, and I wouldn't recommend
> using it because it tramples your freedom in the same way.

Well I currently use my own blog software based on pyblosxom, but I
have made a Django version also that I am testing.

> The software running on WordPress.com may not be the same as that
> available from WordPress.org, and I would still have no idea what
> processing is being done with my data.

Or if wordpress.com will pull the plug on your blog and censor you
because they don't like your speech.

> If you develop a webapp and make it available under the GPLv3, you
> face a copyleft-like problem:

While a GPL fan, I have found recently that using the GPL for programs
in dynamic interpreted languages is fraught with problems, the GPL is
only general if your programs are in C, but that is another discussion
entirely.

> So. Launchpad. Who owns the data? Right now, Canonical does, and there
> is no way to run your own Launchpad.

But you could approximate it more or less. You could use all the same
dependencies, import the data from it. Write your own glue scripts for
anything that Zope+bzr+loggerhead+storm+etc does not provide. You
would have to make your own templates, but you can still have black
text on white background.

However, the question is why would you want to? What does anyone gain
by having two? At least while Canonical is under your evil threshold,
i.e. Canonical's interests are sufficiently aligned with your own.

Unless "Having your own launchpad" in this context just means having
your own bzr repository on your own server, you can do this now. You
can provide a webfront end (loggerhead), and you can get Launchpad,
and any one else you want, to sync with it.

> inability to run their own launchpads. Mark has said that in the long
> term he hopes to do this,

I think he said he would like to give out all the code so it can be
used in other projects.

The aim would not be for you to run your own launchpads, because that
does not make any sense. The key feature of Launchpad is not the
software, it is that it is a fixed point from which you can
collaborate across projects. It is the hub of the wheel. If you create
your own fixed point then you are replacing not complementing
Launchpad.

If you really don't like the idea that Launchpad is this fixed point,
the answer is not to create another Launchpad, but to create a
protocol that does the same job without the requirement of a fixed
point in the middle.

> The Ubuntu community doesn't feel this is appropriation of their data,

I don't think it is (mis)appropriation. Launchpad are providing a good
service, you can get the data out in usable ways. They are not selling
the data to Microsoft under a patent covenant or anything like that.

Best Wishes,
Zeth

--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 02:17 PM
Mark Shuttleworth
 
Default Rethinking Gobuntu

Dave Crossland wrote:


When I don't control the server and I don't own the data, there is no
software freedom issue for me. If I control the server there is a
problem for me indentical to that for the software running on my PC.
If I solely own the data there is a problem if I don't control the
server. If I co-own the data, well, this is an unresolved problem :-)


In the case of Launchpad, we do view you as a co-owner of the data, so
the resolution of this problem is important to us. As you point out,
there's no really clear best practice that works well and has been
shown to be commercially sustainable. That's different to the GPL (even
v3). I think the Affero GPL is a strong candidate for the front line of
thinking on the subject, and that's what I am inclined to use when we
publish Launchpad's source code.



The first step I hope to address is the sense of real data ownership.
It's weird that you have to screen-scrape ANY site to retrieve data
that you co-own, so Launchpad will shortly offer direct API's where you
can retrieve, as often as you want, all the data you added to the
system, modify it programatically, and even retrieve data entered by
others that you could get via screenscraping. In other words, if you
can read it through the web, or modify it through the web, you should
be able to do so programatically. I think that will be a big step
towards restoring "balance to the force", and I hope others will follow
suit.



Separately, we're working with Trac, Bugzilla, Debbugs and other tools
to get replication in place so that you can in fact run a free software
tool to manage your bugs for your project, and still collaborate with
distributions who use Launchpad (Ubuntu, PLD Linux, Nexenta and others)
as well as upstreams (Zope, AWN and 6,000 others). We're also providing
guidance and support for the folks integrating Bazaar into Savannah, so
that you can use an indisputably free platform for project hosting and
still be part of an efficient code collaboration with other projects
that use Launchpad and Bazaar.




The Ubuntu community doesn't feel this is appropriation of their data,
or doesn't mind it, nor the lack of indirect freedom, nor their
inability to run their own launchpads. Mark has said that in the long
term he hopes to do this, and the development of Bazaar and tools for
merging in the patches made by other distributions seems promising.

I think to attract people who won't use Launchpad because they
percieve it as proprietary, Canonical needs to make it possible for
them to set up their own lightweight distro infrastructure to have
direct software freedom with, but which Canonical can pull back
changes from... something Affero licensed and based on Bazaar, I'd
guess.


I do agree that Affero is a better option than the straightforward GPL.
I am still uncomfortable with the idea of having lots of Launchpad's,
though. Launchpad was designed as a central platform to link different
projects, and having multiple centers would decrease the value of it to
the current users. That's not usually true of software - having more
Gnome users is generally positive for the existing ones. This is a
subtle point and not one that most free software folks will
acknowledge. But then they may not use LP enough to appreciate what it
means. They usually revert to "you got, i want, gimme now" ;-) The
roadmap for Launchpad does get it to the point where it will more
graciously handle that distributed nature, and at that point I will be
more comfortable distributing it.



Mark



--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 
Old 04-19-2008, 02:19 PM
Mark Shuttleworth
 
Default Rethinking Gobuntu

Zeth wrote:


If you really don't like the idea that Launchpad is this fixed point,
the answer is not to create another Launchpad, but to create a
protocol that does the same job without the requirement of a fixed
point in the middle.




Canonical is currently funding just this work in Bugzilla, Trac,
Debbugs, Mantis and RT. I've been asking for it for ages but nobody
stepped up, so we have contracted folks to do the work to make it
possible to talk to all of them using a roughly-standardised interface.



Mark



--
Gobuntu-devel mailing list
Gobuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/gobuntu-devel
 

Thread Tools




All times are GMT. The time now is 08:54 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org