You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@roller.apache.org by Ted Husted <te...@gmail.com> on 2006/02/23 15:59:07 UTC
Re: Roller Incubator TODOs
On 12/22/05, Dave Johnson <da...@rollerweblogger.org> wrote:
>
> On Dec 21, 2005, at 11:44 AM, Ted Husted wrote:
> > I was looking over the incubator TODO list for Roller.
> > * http://incubator.apache.org/projects/roller.htm
>
> Thanks Ted! Further comments below...
>
> > If you don't mind my committing some patches, I'm getting up to speed
> > on the Roller architecture, and I could insert the license blurb in to
> > the Java source files, as a part of a general code review.
>
> Please do.
>
> > A related issue is author tags. The "best" practice is to omit them,
> > but some projects have chosen to retain the author tags. For example,
> > Struts removed them, but Velocity kept them. If the Roller committers
> > decided to omit author tags, this would be a good time to remove them.
> > I could do it while inserting the license text.
>
> I could go either way. Anybody have an opinion on this?
Now that I'm close to getting PlanetStruts up, I'd like to get back to
helping with the "All code ASL'ed" item.
To comply with ASF norms and requirements, we should have the "short
form" license at the top of each source code file, which includes the
ASF copyright. We should also remove individual copyright statements
for anyone with a CLA on file (which will hopefully be everyone). If
there is any copyrighted material from somone without a CLA, we will
need to obtain a grant. As a best practice, we should also remove the
@author tags, and be sure to credit all new contributions in the
commit logs. Finally, we should rename README.txt as NOTICE.txt to
comply with the ASF norm and the terms of the Apache Software License.
:) And by "we" I do mean"me" :)
-Ted.
Re: Roller Incubator TODOs
Posted by Ted Husted <te...@gmail.com>.
On 2/23/06, Anil Gangolli <an...@busybuddha.org> wrote:
> Well this is just one of tens of files that have not had the ASF
> copyright applied. It is copyrighted in the historical style of Roller
> contributions; that is, it bears the copyright of the author and grants
> the Roller Weblogger License. Being the current copyright holder of
> record on that particular file, I can say it is not an indication of any
> reluctance to apply the ASF copyright and grant the Apache license; it
> has just not happened.
>
> We all understood this had to happen at some point when joining the
> incubator. I guess there has been some doubt whether we would end up
> having to leave the incubator due to the LGPL issues and no satisfactory
> resolution either on our end (switching off Hibernate) or the Apache end
> (finding an agreement with FSF on interpretation/compatiblity).
>
> If and when we are sure we have committed to resolve issues and push to
> graduate, the copyrights should be applied. Regarding the mechanics of
> this, I have been secretly hoping Dave or one of the other committers
> (possibly me) would go through and do this in scripted fashion in one
> swell foop.
The middle road might be to start removing the individual copyrights
for people who have filed CLAs. It does look like not having a
copyright in every file will be OK with the ASF, so long as there is a
COPYRIGHT.txt file at the root. We could use the list in README.txt
for the COPYRIGHT.txt for now, and add the ASF copyright to the list,
until the other issues are resolved.
If anyone did want to run a script later, it would be easier if the
head of each file were consistent (e.g. no notices at all). I would
like to start a code review, for my own erudition, and could do that
as I go.
We also need to be sure that any notices that are already in files are
from people who have filed CLAs or Grants, or make a list of who still
needs to file.
But, I'll tell you what, if it comes down to replacing with Hibernate
with something, I'll happily do it with Cayenne or iBATIS.
* http://www.objectstyle.org/cayenne/
-Ted.
Re: Roller Incubator TODOs
Posted by Anil Gangolli <an...@busybuddha.org>.
Comments inline (sorry).
Ted Husted wrote:
> On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
>
>> On the one hand it's bad to have non-ASF copyrighted code in the ASF
>> repository; on the other hand it might make the rollerweblogger
>> releases from the committers harder to do.
>>
>
> I guess I don't understand the problem. Do we have ASF committers that
> are adament that individual copyrights must be retained in material
> that is donated to the foundation? (We did have that case with
> another project.)
>
> Even if the board relaxes the policy that every file licensed to the
> ASF include our copyright, we still need to deal with ambigous files
> like
>
> * https://svn.apache.org/repos/asf/incubator/roller/trunk/src/org/roller/business/AutoPingManagerImpl.java
>
> which bear only the copyright of an individual and don't mention the
> ASF at all.
>
Well this is just one of tens of files that have not had the ASF
copyright applied. It is copyrighted in the historical style of Roller
contributions; that is, it bears the copyright of the author and grants
the Roller Weblogger License. Being the current copyright holder of
record on that particular file, I can say it is not an indication of any
reluctance to apply the ASF copyright and grant the Apache license; it
has just not happened.
We all understood this had to happen at some point when joining the
incubator. I guess there has been some doubt whether we would end up
having to leave the incubator due to the LGPL issues and no satisfactory
resolution either on our end (switching off Hibernate) or the Apache end
(finding an agreement with FSF on interpretation/compatiblity).
If and when we are sure we have committed to resolve issues and push to
graduate, the copyrights should be applied. Regarding the mechanics of
this, I have been secretly hoping Dave or one of the other committers
(possibly me) would go through and do this in scripted fashion in one
swell foop.
--a.
> Even then, I believe the policy would be to either place 1 copyright
> notice in licensed files running soley to the ASF, -OR- 0 copyright
> notices in licensed files, and a COPYRIGHT.txt listing all parties.
>
> Right now, we seem to be dong neither :)
>
> What is the Roller PPMC's policy?
>
> * If the Board relaxes the "every file" policy, will Roller choose to
> omit the standard copyright lines in favor of a root COPYRIGHT.txt
> file.
>
> * If a file already contains an individual copyright notice, and the
> holder is a ASF committer, have all the ASF committers agreed that we
> can we remove the individual notice?
>
> And, do we have paperwork on file for "past members" Mindaguas Idzelis
> and Jaap Van Der Molen? (I couldn't find any.)
>
This is a good point. I think we'd need some kind of grant from them.
> -Ted.
>
>
>
Re: Roller Incubator TODOs
Posted by Henri Yandell <fl...@gmail.com>.
On 3/2/06, David M Johnson <Da...@sun.com> wrote:
>
> On Feb 23, 2006, at 9:58 PM, David M Johnson wrote:
> > On Feb 23, 2006, at 9:31 PM, Henri Yandell wrote:
> >> Is any of the Roller code "Copyright Sun" now?
> >
> > I guess the answer is yes, but that is *only* because I add the
> > words "Copyright Sun" in some of the files I committed. Nobody told
> > me to do that, I just figured that it was the right thing to do.
> >
> > I would assume for now that Sun is OK with having all source be
> > copyright Apache, since that's what Sun does on other Apache
> > projects -- but I will verify that.
>
> I spoke with Sun's Craig McCanahan and Craig Russell and both of them
> use the Apache copyright when creating new files for the ASF projects
> that they are involved with. That's what Allen and I should do as well.
>
> So, I think the TODOs here are:
> For every file that has no copyright message, add copyright Apache
+1
> For every file created by Dave with a Sun copyright, change it to
> copyright Apache
+1 - I think this falls under your CCLA, comes down to your
relationship with your employer if anything ever comes of it.
> For files with other copyrights, leave the copyright in place and add
> to list in CREDITS* file
For those files with copyrights owned by other Roller committers; ask
them to replace their copyright with the Apache copyright.
For those files with copyrights owned by non-Roller committers, ask
those people for permission to remove their copyright - anyone
listening know if we have to do a fax-a-form dance for that?
Hen
Re: Roller Incubator TODOs
Posted by Ted Husted <te...@gmail.com>.
On 3/5/06, Allen Gilliland <Al...@sun.com> wrote:
> I would think that the sooner the better, but I doubt there is any real
> pressure to do it immediately.
I need to focus on WebWork (Struts Action2) for the next two weeks,
but after that, I'd like to help square this way.
-Ted.
Re: Roller Incubator TODOs
Posted by Allen Gilliland <Al...@Sun.COM>.
On Sun, 2006-03-05 at 09:35, Anil Gangolli wrote:
> David M Johnson wrote:
> >
> > On Feb 23, 2006, at 9:58 PM, David M Johnson wrote:
> >> On Feb 23, 2006, at 9:31 PM, Henri Yandell wrote:
> >>> Is any of the Roller code "Copyright Sun" now?
> >>
> >> I guess the answer is yes, but that is *only* because I add the words
> >> "Copyright Sun" in some of the files I committed. Nobody told me to
> >> do that, I just figured that it was the right thing to do.
> >>
> >> I would assume for now that Sun is OK with having all source be
> >> copyright Apache, since that's what Sun does on other Apache projects
> >> -- but I will verify that.
> >
> > I spoke with Sun's Craig McCanahan and Craig Russell and both of them
> > use the Apache copyright when creating new files for the ASF projects
> > that they are involved with. That's what Allen and I should do as well.
> >
> > So, I think the TODOs here are:
> > For every file that has no copyright message, add copyright Apache
> > For every file created by Dave with a Sun copyright, change it to
> > copyright Apache
> I think this should also be the case for all current committers that
> have personal copyright notices on the files (like me).
> > For files with other copyrights, leave the copyright in place and add
> > to list in CREDITS* file
> >
> > * or whatever the appropriate file is
> >
> > Comments?
> >
> now or later?
I would think that the sooner the better, but I doubt there is any real
pressure to do it immediately.
-- Allen
> >
> > - Dave
> >
> >
> >
>
Re: Roller Incubator TODOs
Posted by Anil Gangolli <an...@busybuddha.org>.
David M Johnson wrote:
>
> On Feb 23, 2006, at 9:58 PM, David M Johnson wrote:
>> On Feb 23, 2006, at 9:31 PM, Henri Yandell wrote:
>>> Is any of the Roller code "Copyright Sun" now?
>>
>> I guess the answer is yes, but that is *only* because I add the words
>> "Copyright Sun" in some of the files I committed. Nobody told me to
>> do that, I just figured that it was the right thing to do.
>>
>> I would assume for now that Sun is OK with having all source be
>> copyright Apache, since that's what Sun does on other Apache projects
>> -- but I will verify that.
>
> I spoke with Sun's Craig McCanahan and Craig Russell and both of them
> use the Apache copyright when creating new files for the ASF projects
> that they are involved with. That's what Allen and I should do as well.
>
> So, I think the TODOs here are:
> For every file that has no copyright message, add copyright Apache
> For every file created by Dave with a Sun copyright, change it to
> copyright Apache
I think this should also be the case for all current committers that
have personal copyright notices on the files (like me).
> For files with other copyrights, leave the copyright in place and add
> to list in CREDITS* file
>
> * or whatever the appropriate file is
>
> Comments?
>
now or later?
>
> - Dave
>
>
>
Re: Roller Incubator TODOs
Posted by David M Johnson <Da...@Sun.COM>.
On Feb 23, 2006, at 9:58 PM, David M Johnson wrote:
> On Feb 23, 2006, at 9:31 PM, Henri Yandell wrote:
>> Is any of the Roller code "Copyright Sun" now?
>
> I guess the answer is yes, but that is *only* because I add the
> words "Copyright Sun" in some of the files I committed. Nobody told
> me to do that, I just figured that it was the right thing to do.
>
> I would assume for now that Sun is OK with having all source be
> copyright Apache, since that's what Sun does on other Apache
> projects -- but I will verify that.
I spoke with Sun's Craig McCanahan and Craig Russell and both of them
use the Apache copyright when creating new files for the ASF projects
that they are involved with. That's what Allen and I should do as well.
So, I think the TODOs here are:
For every file that has no copyright message, add copyright Apache
For every file created by Dave with a Sun copyright, change it to
copyright Apache
For files with other copyrights, leave the copyright in place and add
to list in CREDITS* file
* or whatever the appropriate file is
Comments?
- Dave
Re: Roller Incubator TODOs
Posted by David M Johnson <Da...@Sun.COM>.
On Feb 23, 2006, at 9:31 PM, Henri Yandell wrote:
> Is any of the Roller code "Copyright Sun" now?
I guess the answer is yes, but that is *only* because I add the words
"Copyright Sun" in some of the files I committed. Nobody told me to
do that, I just figured that it was the right thing to do.
I would assume for now that Sun is OK with having all source be
copyright Apache, since that's what Sun does on other Apache projects
-- but I will verify that.
- Dave
Re: Roller Incubator TODOs
Posted by Allen Gilliland <Al...@Sun.COM>.
On Thu, 2006-02-23 at 18:31, Henri Yandell wrote:
>
> > > And, do we have paperwork on file for "past members" Mindaguas Idzelis
> > > and Jaap Van Der Molen? (I couldn't find any.)
> >
> > No we don't.
>
> Definitely a new todo then. Obtain software grants (I assume, rather
> than ICLAs) from Mindaguas and Jaap.
what code is owned by them? i think it's worth checking how much code we are talking about and if it's really needed anymore. i remember Jaap doing some i18n stuff and i am actually planning to rip out some of that stuff and replace it in the not too distant future.
>
> > When I started Roller, I thought that all contributions would become
> > copyright by me. Later I realized that was not the open source norm and
> > I started leaving copyrights on the files.
>
> Legal norm :)
>
> > Recently, I've be adding "Copyright Sun" on code I write for Sun and
> > "Copyright Dave Johnson" for other code I contribute (e.g. the Atom
> > protocol implementation I wrote in off-hours for my book).
>
> Is any of the Roller code "Copyright Sun" now?
I am the absolute un-authority on copyright/licensing issues, but I find it hard to believe that Sun expects code from contributers like myself and Dave to fall under a Sun copyright.
but that's just my feeling and definitely wouldn't hold up in a court of law ;)
-- Allen
>
> Hen
Re: Roller Incubator TODOs
Posted by Henri Yandell <fl...@gmail.com>.
On 2/23/06, David M Johnson <Da...@sun.com> wrote:
>
> Comments inline...
>
>
> On Feb 23, 2006, at 11:52 AM, Ted Husted wrote:
> > On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
> >> On the one hand it's bad to have non-ASF copyrighted code in the ASF
> >> repository; on the other hand it might make the rollerweblogger
> >> releases from the committers harder to do.
> >
> > I guess I don't understand the problem. Do we have ASF committers that
> > are adament that individual copyrights must be retained in material
> > that is donated to the foundation? (We did have that case with
> > another project.)
>
> I don't believe so. However, when I was researching the move to
> Apache some people told me that copyright DID NOT have to be
> signed over to Apache. Despite those capital letters, I don't really
> have a strong opinion on this topic.
>
> Has Sun insisted on keeping Sun copyright on committer contributions
> on other Apache projects?
Interesting. I've never heard of Apache code not being copyrighted to
the ASF - ignoring the grey period of Incubation.
Personal copyrights seem even more of a pain in the arse to maintain
in the long term than
> > Even if the board relaxes the policy that every file licensed to the
> > ASF include our copyright, we still need to deal with ambigous files
> > like
> >
> > * https://svn.apache.org/repos/asf/incubator/roller/trunk/src/org/
> > roller/business/AutoPingManagerImpl.java
> >
> > which bear only the copyright of an individual and don't mention the
> > ASF at all.
> >
> > Even then, I believe the policy would be to either place 1 copyright
> > notice in licensed files running soley to the ASF, -OR- 0 copyright
> > notices in licensed files, and a COPYRIGHT.txt listing all parties.
> >
> > Right now, we seem to be dong neither :)
> >
> > What is the Roller PPMC's policy?
>
> Good question ;-)
Very interested in hearing if there are any reasons for things not to
eventually be copyright ASF.
> > And, do we have paperwork on file for "past members" Mindaguas Idzelis
> > and Jaap Van Der Molen? (I couldn't find any.)
>
> No we don't.
Definitely a new todo then. Obtain software grants (I assume, rather
than ICLAs) from Mindaguas and Jaap.
> When I started Roller, I thought that all contributions would become
> copyright by me. Later I realized that was not the open source norm and
> I started leaving copyrights on the files.
Legal norm :)
> Recently, I've be adding "Copyright Sun" on code I write for Sun and
> "Copyright Dave Johnson" for other code I contribute (e.g. the Atom
> protocol implementation I wrote in off-hours for my book).
Is any of the Roller code "Copyright Sun" now?
Hen
Re: Roller Incubator TODOs
Posted by David M Johnson <Da...@Sun.COM>.
Comments inline...
On Feb 23, 2006, at 11:52 AM, Ted Husted wrote:
> On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
>> On the one hand it's bad to have non-ASF copyrighted code in the ASF
>> repository; on the other hand it might make the rollerweblogger
>> releases from the committers harder to do.
>
> I guess I don't understand the problem. Do we have ASF committers that
> are adament that individual copyrights must be retained in material
> that is donated to the foundation? (We did have that case with
> another project.)
I don't believe so. However, when I was researching the move to
Apache some people told me that copyright DID NOT have to be
signed over to Apache. Despite those capital letters, I don't really
have a strong opinion on this topic.
Has Sun insisted on keeping Sun copyright on committer contributions
on other Apache projects?
> Even if the board relaxes the policy that every file licensed to the
> ASF include our copyright, we still need to deal with ambigous files
> like
>
> * https://svn.apache.org/repos/asf/incubator/roller/trunk/src/org/
> roller/business/AutoPingManagerImpl.java
>
> which bear only the copyright of an individual and don't mention the
> ASF at all.
>
> Even then, I believe the policy would be to either place 1 copyright
> notice in licensed files running soley to the ASF, -OR- 0 copyright
> notices in licensed files, and a COPYRIGHT.txt listing all parties.
>
> Right now, we seem to be dong neither :)
>
> What is the Roller PPMC's policy?
Good question ;-)
> And, do we have paperwork on file for "past members" Mindaguas Idzelis
> and Jaap Van Der Molen? (I couldn't find any.)
No we don't.
When I started Roller, I thought that all contributions would become
copyright by me. Later I realized that was not the open source norm and
I started leaving copyrights on the files.
Recently, I've be adding "Copyright Sun" on code I write for Sun and
"Copyright Dave Johnson" for other code I contribute (e.g. the Atom
protocol implementation I wrote in off-hours for my book).
I'm comfortable going with Apache best practice on the copyright
and @author tag stuff -- as long as the team agrees to that.
- Dave
Re: Roller Incubator TODOs
Posted by Henri Yandell <fl...@gmail.com>.
On 2/24/06, Ted Husted <te...@gmail.com> wrote:
> On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
> > We shouldn't be leaping in to make all the source ASF-based while
> > there are still questions over the resolution of the LGPL
> > dependencies.
>
> What I would suggest is that we
>
> * place a single COPYRIGHT.txt file in the root of the project with
> whatever copyright assertions we want to make now,
>
> * remove *all* the copyright notices and other cruft from the top of
> the files that are asserted by the ASF Committers, and
>
> * note any other copyright assertions (e.g. Sun) in the COPYRIGHT.txt file.
>
> It hasn't been made altogether official yet, but it looks like leaving
> all copyright notices out of the files and putting them in a single
> COPYRIGHT file is probably to be an acceptable alternative to the ASF.
> Even if it isn't, it would be much easier to have a script update all
> the source files if the heads were already clean.
>
> Even if the committers trudged back to java.net, using a single
> COPYRIGHT.txt instead of individual statements scattered in some of
> the files seems like a better approach for everyone. Given a volunteer
> to do the work (me! me!), it's win-win either way.
Bit wary of diving in and redoing the copyright to how we think it
will be. It may get bogged down in legal realities in the same way the
licensing issue has. I'd definitely like to get our copyright
questions resolved (see below); and then probably go ahead and move to
the standard ASL copyright at the top of each file method unless we've
heard otherwise. I'd like to check with the Incubator PMC that
changing the copyrights to the ASF wouldn't have much effect on the
ability to release at rollerweblogger.
> Of course, even if all the files were marked with the usual ASF
> statement, the team would still be able to trudge back to java.net and
> setup shop. The ASF statement is non-exclusive and doesn't prevent
> anyone from continuing work elsewhere. If Roller fails incubator, no
> one is going to assert that the "Roller" name is still under the ASF
> copyright. We really are all friend here :)
>
> But, I'm not suggesting that we insert the ASF statement at this time,
> just that we remove *all* the statements, and use a single
> COPYRIGHT.txt file instead.
+1 to starting the work to solve the copyright. ie) finding out how
much code is affected by the two committers who are not ASF
committers, getting their permission if need be, being happy about the
Sun copyright statements from David and being able to zap them all.
Then regroup before actually making the changes to see what our situation is.
> > that it makes it harder for the Roller committers to make releases at
> > rollerweblogger.org. Currently it does as much as possible to not look
> > like an Apache release.
>
> Under the Apache License, anyone can roll a release of any of our
> codebases and distribute it wherever they like. The only stipulation
> is that no one else can call it an Apache or ASF release.
Right. So grab the source; remove any reference to Apache Xxxx on the
products published interface; add one that points to Apache as the
owners of the code this is rereleased from; ideally change all package
names from org.apache.
It also changes the community sell of it. We're not forking the Apache
version, we're squatting in the Apache infrastructure and releasing a
version from the Roller community.
> > Hibernate is the easy one to focus on, but we have other licensing
> > issues to deal with too. There are other LGPL components, and we need
> > to do an audit of our Javascript; I want to make sure that we're good
> > there license-wise.
>
> Is that it, then? Hibernate and the Javascript?
>
> If another alternative is not forthcoming, I'll do whatever it takes
> to replace Hibernate with iBATIS or Cayenne.
There are some other LGPL components; I think they mostly come down to
one of the editors that we've talked about dropping (2.1? 2.2? not
sure when it's going).
There are also other licenses that the new policy is going to affect.
And the LGPL issue is going to be lessened. Give me just a little more
time to continue being a pain in the arse to Cliff before using up too
much energy on the license fixing? :)
> Aside from my own commitment, I'm sure some of the Cayenne team would
> be very keen on helping us move Roller to Cayenne.
Craig's doing a JPOX one isn't he?
Sorry if the above is disjoint;
Generally we need to:
* Fix copyright
* Fix distribution for licensing
* Get a zone setup for the Roller website to be migrated over -
involves a db migration from mysql (I think) to Derby.
* Migrate Roller wiki (JSPWiki to moinmoin?)
* Jira is another in a holding pattern - flat out not possible to
migrate projects from one Jira into another Jira.
Hen
Re: Roller Incubator TODOs
Posted by Ted Husted <te...@gmail.com>.
On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
> We shouldn't be leaping in to make all the source ASF-based while
> there are still questions over the resolution of the LGPL
> dependencies.
What I would suggest is that we
* place a single COPYRIGHT.txt file in the root of the project with
whatever copyright assertions we want to make now,
* remove *all* the copyright notices and other cruft from the top of
the files that are asserted by the ASF Committers, and
* note any other copyright assertions (e.g. Sun) in the COPYRIGHT.txt file.
It hasn't been made altogether official yet, but it looks like leaving
all copyright notices out of the files and putting them in a single
COPYRIGHT file is probably to be an acceptable alternative to the ASF.
Even if it isn't, it would be much easier to have a script update all
the source files if the heads were already clean.
Even if the committers trudged back to java.net, using a single
COPYRIGHT.txt instead of individual statements scattered in some of
the files seems like a better approach for everyone. Given a volunteer
to do the work (me! me!), it's win-win either way.
Of course, even if all the files were marked with the usual ASF
statement, the team would still be able to trudge back to java.net and
setup shop. The ASF statement is non-exclusive and doesn't prevent
anyone from continuing work elsewhere. If Roller fails incubator, no
one is going to assert that the "Roller" name is still under the ASF
copyright. We really are all friend here :)
But, I'm not suggesting that we insert the ASF statement at this time,
just that we remove *all* the statements, and use a single
COPYRIGHT.txt file instead.
> that it makes it harder for the Roller committers to make releases at
> rollerweblogger.org. Currently it does as much as possible to not look
> like an Apache release.
Under the Apache License, anyone can roll a release of any of our
codebases and distribute it wherever they like. The only stipulation
is that no one else can call it an Apache or ASF release.
> Hibernate is the easy one to focus on, but we have other licensing
> issues to deal with too. There are other LGPL components, and we need
> to do an audit of our Javascript; I want to make sure that we're good
> there license-wise.
Is that it, then? Hibernate and the Javascript?
If another alternative is not forthcoming, I'll do whatever it takes
to replace Hibernate with iBATIS or Cayenne.
* http://www.objectstyle.org/cayenne/
The Cayenne team is considering an Incubator proposal of their own.
* http://www.mail-archive.com/tapestry-dev@jakarta.apache.org/msg04867.html
Aside from my own commitment, I'm sure some of the Cayenne team would
be very keen on helping us move Roller to Cayenne.
-Ted.
Re: Roller Incubator TODOs
Posted by Henri Yandell <fl...@gmail.com>.
On 2/23/06, Ted Husted <te...@gmail.com> wrote:
> On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
> > On the one hand it's bad to have non-ASF copyrighted code in the ASF
> > repository; on the other hand it might make the rollerweblogger
> > releases from the committers harder to do.
>
> I guess I don't understand the problem. Do we have ASF committers that
> are adament that individual copyrights must be retained in material
> that is donated to the foundation? (We did have that case with
> another project.)
Anil has the right of it.
We shouldn't be leaping in to make all the source ASF-based while
there are still questions over the resolution of the LGPL
dependencies. Roller entered the Incubator amidst a euphoric wave of
joy and belief that there was no legal issue in using LGPL'd works -
long a thorny topic. However that was followed by the realisation that
distributing LGPL'd works really screws up the conditions set on an
ASF product.
I think the rest of the list of TODOs are blocked by our inability to
release an official ASF version of Roller. Anil points out one
negative side of this - that if Roller had to leave the ASF, it would
be a lot harder if lots of these things were done; but another one is
that it makes it harder for the Roller committers to make releases at
rollerweblogger.org. Currently it does as much as possible to not look
like an Apache release.
Thus the no-mans land in which Roller is stuck. The community
maintains momentum and I spend my 'Roller time' keeping up to date
with the licensing issues on the private lists and hassling Cliff
Schmidt mercilessly :) Also offering help etc so that he doesn't hire
hitmen.
Hibernate is the easy one to focus on, but we have other licensing
issues to deal with too. There are other LGPL components, and we need
to do an audit of our Javascript; I want to make sure that we're good
there license-wise.
Hen
Re: Roller Incubator TODOs
Posted by Ted Husted <te...@gmail.com>.
On 2/23/06, Henri Yandell <fl...@gmail.com> wrote:
> On the one hand it's bad to have non-ASF copyrighted code in the ASF
> repository; on the other hand it might make the rollerweblogger
> releases from the committers harder to do.
I guess I don't understand the problem. Do we have ASF committers that
are adament that individual copyrights must be retained in material
that is donated to the foundation? (We did have that case with
another project.)
Even if the board relaxes the policy that every file licensed to the
ASF include our copyright, we still need to deal with ambigous files
like
* https://svn.apache.org/repos/asf/incubator/roller/trunk/src/org/roller/business/AutoPingManagerImpl.java
which bear only the copyright of an individual and don't mention the
ASF at all.
Even then, I believe the policy would be to either place 1 copyright
notice in licensed files running soley to the ASF, -OR- 0 copyright
notices in licensed files, and a COPYRIGHT.txt listing all parties.
Right now, we seem to be dong neither :)
What is the Roller PPMC's policy?
* If the Board relaxes the "every file" policy, will Roller choose to
omit the standard copyright lines in favor of a root COPYRIGHT.txt
file.
* If a file already contains an individual copyright notice, and the
holder is a ASF committer, have all the ASF committers agreed that we
can we remove the individual notice?
And, do we have paperwork on file for "past members" Mindaguas Idzelis
and Jaap Van Der Molen? (I couldn't find any.)
-Ted.
Re: Roller Incubator TODOs
Posted by Ted Husted <te...@gmail.com>.
On 2/23/06, Allen Gilliland <Al...@sun.com> wrote:
> > * http://tinyurl.com/mw7t6
>
> this is a total side note, but i also don't understand the tinyurl thing.
Long URLs, like this:
* http://mail-archives.apache.org/mod_mbox/www-community/200306.mbox/%3C20030609234538.GA22335@lyra.org%3E
tend to break in e-mail threads, so it's polite to use a TinyURL instead.
-T.
Re: Roller Incubator TODOs
Posted by Allen Gilliland <Al...@Sun.COM>.
On Thu, 2006-02-23 at 10:19, Ted Husted wrote:
> The topic of author tags has been discussed in literally hundreds of
> ASF posts on various mailing lists. In the end, the ASF board
> determined that removing the tags is a "best practice", but it is not
> a requirement.
fair enough.
>
> As to why, Greg Stein sums it up well here:
>
> * http://tinyurl.com/mw7t6
this is a total side note, but i also don't understand the tinyurl thing.
>
> As to "who to contact first", any and all development conversations
> should happen on a mailing list. We should not contact each other
> about code, we should go through the list. If the topic is "sensitive"
> for some reason, then we can use the PMC list instead, at least until
> we get past the sensitive bits. But it is a very bad practice for one
> committer to email another with a question about code.
>
> The ASF considers the mailing lists to be our communal memory. We want
> the entire decision making process to happen over the lists so that it
> is made part of the ASF archives. That's one reason why all the
> commits, and issue ticket postings, and wiki changes, are all
> funnelled through the mailing lists. (I think another TODO may be to
> configure JIRA to post changes to dev@.)
>
> As to the idea of a "primary maintainer", while things like that
> happen, it's not an idea that the ASF encourages. Every member of the
> PMC is jointly and severally responsible for all of the code. No PMC
> member needs another member's permission to make a change, and any PMC
> member can veto a change on technical grounds. (Which is why we are
> careful about who we invite to be a committer.) Once the code is
> committed to the repository and donated to the ASF, it belongs to all
> of us, and we all the authors now.
I think all of that sounds ideal, but isn't always reality. I think that with roller we do all of that stuff, but it doesn't change the fact that people tend to deliver focused pieces of functionality to the application and hence they tend to know that piece better than everyone else. I don't like the idea of a "primary maintainer" either, but we are a small project with only a small group of people contributing most of the code so it's inevitable that some parts of the code base are likely only truly maintained by one person.
I don't really care that much about @author tags, so I'm fine with removing them.
-- Allen
>
> -Ted.
>
>
> On 2/23/06, Allen Gilliland <Al...@sun.com> wrote:
> > I am still a bit confused as to why this is a best practice.
> >
> > I actually like having the @author tags so that I can know who "maintains" that piece of code. To me it's not about giving credit it's more about responsibility and it helps me to know who owns what pieces of code so that if I have questions about it I know who to contact first. I am not particularly tied to the @author javdoc markup if we don't want the author to show up in our javadocs, but I do like having that indication somewhere in the file.
> >
> > I feel that svn doesn't do quite as good a job of that because often times commit notes are not very specific and even if the last few commits are from user XXX that doesn't mean that XXX is the primary maintainer of that code. I also think it's more of a pain to go back and lookup svn commit logs versus just seeing the @author tag in the code.
Re: Roller Incubator TODOs
Posted by Ted Husted <te...@gmail.com>.
The topic of author tags has been discussed in literally hundreds of
ASF posts on various mailing lists. In the end, the ASF board
determined that removing the tags is a "best practice", but it is not
a requirement.
As to why, Greg Stein sums it up well here:
* http://tinyurl.com/mw7t6
As to "who to contact first", any and all development conversations
should happen on a mailing list. We should not contact each other
about code, we should go through the list. If the topic is "sensitive"
for some reason, then we can use the PMC list instead, at least until
we get past the sensitive bits. But it is a very bad practice for one
committer to email another with a question about code.
The ASF considers the mailing lists to be our communal memory. We want
the entire decision making process to happen over the lists so that it
is made part of the ASF archives. That's one reason why all the
commits, and issue ticket postings, and wiki changes, are all
funnelled through the mailing lists. (I think another TODO may be to
configure JIRA to post changes to dev@.)
As to the idea of a "primary maintainer", while things like that
happen, it's not an idea that the ASF encourages. Every member of the
PMC is jointly and severally responsible for all of the code. No PMC
member needs another member's permission to make a change, and any PMC
member can veto a change on technical grounds. (Which is why we are
careful about who we invite to be a committer.) Once the code is
committed to the repository and donated to the ASF, it belongs to all
of us, and we all the authors now.
-Ted.
On 2/23/06, Allen Gilliland <Al...@sun.com> wrote:
> I am still a bit confused as to why this is a best practice.
>
> I actually like having the @author tags so that I can know who "maintains" that piece of code. To me it's not about giving credit it's more about responsibility and it helps me to know who owns what pieces of code so that if I have questions about it I know who to contact first. I am not particularly tied to the @author javdoc markup if we don't want the author to show up in our javadocs, but I do like having that indication somewhere in the file.
>
> I feel that svn doesn't do quite as good a job of that because often times commit notes are not very specific and even if the last few commits are from user XXX that doesn't mean that XXX is the primary maintainer of that code. I also think it's more of a pain to go back and lookup svn commit logs versus just seeing the @author tag in the code.
Re: Roller Incubator TODOs
Posted by Allen Gilliland <Al...@Sun.COM>.
> > As a best practice, we should also remove the
> > @author tags, and be sure to credit all new contributions in the
> > commit logs.
>
> To clarify, this is up to the community, but most have gone towards
> removing the @author tags. Personally I just never put my name in one
> anymore and avoid the discussions, but I'm +1 to not bothering with
> @authors.
I am still a bit confused as to why this is a best practice.
I actually like having the @author tags so that I can know who "maintains" that piece of code. To me it's not about giving credit it's more about responsibility and it helps me to know who owns what pieces of code so that if I have questions about it I know who to contact first. I am not particularly tied to the @author javdoc markup if we don't want the author to show up in our javadocs, but I do like having that indication somewhere in the file.
I feel that svn doesn't do quite as good a job of that because often times commit notes are not very specific and even if the last few commits are from user XXX that doesn't mean that XXX is the primary maintainer of that code. I also think it's more of a pain to go back and lookup svn commit logs versus just seeing the @author tag in the code.
-- Allen
Re: Roller Incubator TODOs
Posted by Henri Yandell <fl...@gmail.com>.
On 2/23/06, Ted Husted <te...@gmail.com> wrote:
> On 12/22/05, Dave Johnson <da...@rollerweblogger.org> wrote:
> >
> > On Dec 21, 2005, at 11:44 AM, Ted Husted wrote:
> > > I was looking over the incubator TODO list for Roller.
> > > * http://incubator.apache.org/projects/roller.htm
> >
> > Thanks Ted! Further comments below...
> >
> > > If you don't mind my committing some patches, I'm getting up to speed
> > > on the Roller architecture, and I could insert the license blurb in to
> > > the Java source files, as a part of a general code review.
> >
> > Please do.
> >
> > > A related issue is author tags. The "best" practice is to omit them,
> > > but some projects have chosen to retain the author tags. For example,
> > > Struts removed them, but Velocity kept them. If the Roller committers
> > > decided to omit author tags, this would be a good time to remove them.
> > > I could do it while inserting the license text.
> >
> > I could go either way. Anybody have an opinion on this?
>
> Now that I'm close to getting PlanetStruts up, I'd like to get back to
> helping with the "All code ASL'ed" item.
The issue we face is that we're in a holding pattern (but still
wanting to do releases) as the policy on whether Roller can release is
still in formulation (it's very, very close to being public, if I can
catch Cliff I'll find out if I can post the draft link to here).
On the one hand it's bad to have non-ASF copyrighted code in the ASF
repository; on the other hand it might make the rollerweblogger
releases from the committers harder to do. I'll ask our board
representative (well, the board member who last had to ask me a
question about things).
> To comply with ASF norms and requirements, we should have the "short
> form" license at the top of each source code file, which includes the
> ASF copyright. We should also remove individual copyright statements
> for anyone with a CLA on file (which will hopefully be everyone). If
> there is any copyrighted material from somone without a CLA, we will
> need to obtain a grant.
+1, eventually (see above).
> As a best practice, we should also remove the
> @author tags, and be sure to credit all new contributions in the
> commit logs.
To clarify, this is up to the community, but most have gone towards
removing the @author tags. Personally I just never put my name in one
anymore and avoid the discussions, but I'm +1 to not bothering with
@authors.
> Finally, we should rename README.txt as NOTICE.txt to
> comply with the ASF norm and the terms of the Apache Software License.
+1, that should happen right away.
Hen