On 28/11/2014 5:49 PM, Scott Gray wrote:
Ron, we get that you really like the idea of sub-projects. The only
positive I can see about a sub-project is that the component gets to
stick with the Apache brand/trademark, I'm pretty sure that's the only
positive I've seen you mention that doesn't also apply to an external
project. With the correct marketing and promotion of an external
project I'm sure we could virtually close that gap.
Cons on the other-hand:
1. Still ultimately the responsibility of the TLP PMC
True but not on a day to day basis
2. Burden of setting up all infrastructure required by the sub-project
This does need to be looked at. Clearly other projects feel that the
benefits outweigh the costs.
3. Sub-projects are still bound by the constraints of the ASF
community. If Adrian or whoever wants to do something in an external
project, he can just do it, no votes, no repetitive discussions
seeking consensus, none of the frustration that comes from a community
driven project. From your point of view that may seem like a negative
but trust me, for a piece of work you plan to take ownership of and
push forward with, it's hugely liberating and allows progress at a
much faster pace. Just ask David Jones, I'm sure he'd agree.
This project is not short of forks.
Anyone can do this but you lose the advantage of building a team
(unless you pay them).
It also leads to duplication of effort and divergent products that
dilute the results and confuses the market.
If Adrian does a private fork of Asset Management, what is the effect on
the users of the OFBiz version.
Will people want to use a "better" private fork from an unknown source
or stick with the Apache version even if it is less functional.
Can the Apache project abandon a component if someone says that they
will make a private fork?
5. Sub-projects would generally (I guess) still be bound to follow
OFBiz "best practices". An external project would have no such
constraint and could look for new ways (for example) to render the UI
through something like
AngularJs/Ember/Backbone/insert-flavor-of-the-month-here or take a
completely different approach in other areas without dirtying the
waters of the core OFBiz project. This could well lead to innovations
making their way back into the OFBiz project.
If that is someone's intention, then setting up a fork would be the best
way to achieve this.
I was thinking of modules that would integrate with the core and
framework in a seamless fashion.
This depends on a certain amount of technological clarity and consistency.
I get that it's unlikely anyone will sway your opinion at this point
but still, it's important to consider the negatives of your approach
as well.
I still think on balance, that it is pretty clear that sub-projects is
the right way to go but it is not a magic solution to the problems in
the OFBiz project and will take some effort to achieve a good result.
There is nothing to stop anyone from doing a private fork but I don't
think that this is a good long-term plan and does nothing to build the
OFBiz community or reduce the workload on the current team.
Regards
Scott
On Sat, Nov 29, 2014 at 10:56 AM, Ron Wheeler
<rwhee...@artifact-software.com> wrote:
!) Sub-projects allow worthwhile projects to find new supporters through
better transparency, focused mission and clearer borders around the
knowledge needed to contribute.
2) I thought that Adrian was suggesting that Asset Management was an effort
that he would support
3) Just when people say that they don't want sub-projects, they turn around
and suggest activities and plans that look like sub-projects.
They just want to invent a new structure outside Apache to do this.
Ron
On 28/11/2014 4:29 PM, Taher Alkhateeb wrote:
Hi Adrian and everyone,
I think this issue was discussed in multiple threads before. There seems
to
be a general agreement that resources are low. The question is then why
sub-projects or forks or spinoffs? Why not just keep specialpurpose in the
project? It's live functioning code even if not updated and it is after
all
secondary to the core applications. If anyone then wants to contribute
they
would be supervised by experts.
IMHO whatever you choose whether sub-projects or forks would probably
just
kill those components.
My 2 cents
Taher Alkhateeb
On Nov 29, 2014 12:15 AM, "Adrian Crum"
<adrian.c...@sandglass-software.com>
wrote:
This conversation has stopped making any sense.
The special purpose components are removed from releases because we don't
have enough resources to maintain them. Now there is interest in putting
them back, but we STILL don't have the resources to maintain them. A
suggestion was made to make them sub-projects, but that requires MORE
resources. So the suggestion was made to spin them off to separate
projects
where they can stand or fall on their own. The sub-project idea (as far
as
I can tell) is dead.
What part of that aren't you understanding?
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 11/28/2014 7:37 PM, Ron Wheeler wrote:
On 28/11/2014 11:23 AM, Adrian Crum wrote:
I agree with Jacopo that OFBiz sub-projects will be nearly impossible
to maintain. That is why I suggested moving special purpose components
to separate projects.
I am willing to move one component to a separate project as a trial
run. I have no interest in being a "chair of a sub-PMC."
Who would you be willing to have as leader and chief architect?
Ron
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 11/28/2014 4:12 PM, Ron Wheeler wrote:
Can someone on the PMC or a current committer find out what has to be
done to set up an Apacahe sub-project in terms of administration
(might
be nothing) and fixing the SCM access so that committers to the
sub-project are not required to be committers to the core and
framework.
This may not be possible from a technical sense but at least it should
be possible to organize the SCM so it is clear what sub-project
committers are supposed to do.
If Adrian is willing to act as Chair of the sub-PMC, that would be a
great start.
I will join on the documentation side to help set up the sub-project
doc
structure.
Ron
On 27/11/2014 10:31 AM, Adrian Crum wrote:
I would be willing to spin off Asset Maintenance to a separate
project. I was thinking it could be a good test-run of the concept.
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 11/27/2014 2:16 PM, Jacques Le Roux wrote:
Hi Jacopo,
I looked a bit back. Even if it's not clearly related I trace this
back
to the slim-down effort. We can forget it since nobody never
complained
(pfew...).
Then you proposed to move some component from specialpurpose to
extras.
As you said, not every people were happy with it (at least Pierre
and in
a less measure I)
I then suggested some components to keep
markmail.org/message/4camcprzximkcftc
<<assetmaint
ecommerce
example*
pos
maybe myportal?
projectmgr
scrum
and maybe webpos?>>
In a very recent thread http://markmail.org/message/ctusiepnuciofc32
I
suggested to associate people with components
<<project manager (Pierre Smits?)
scrum (Hans?)
examples and ext (at least me)
myportal (French people use portals, not sure for myportal?)
When I look now at my 1st list, obviously I can also support the
POS
even if I have less interest in it now.
Pierre at http://markmail.org/message/n23oyye2i24kqzpg suggested
HHFacility, ASSETMAINT, CMSSITE, PROJECTMGR, MYPORTAL, SCRUM, etc.
I don't like the etc. ;) but I can agree to add
HHFacility and CMSSITE to my list
Also in this list birt is missing, clearly at least Chatree has an
interest in it and knows how to maintain it.
I don't know if Anil or/and Adrian have still an interest in
ASSETMAINT
but anyway it seems it's worth to keep it.
HHFacility does not need much work to maintain
For CMSSITE I'm unsure, but it's interesting for the online help
(too
bad BJ is no longer with us)
BTWcmssite/cms/APACHE_OFBIZ_HTML
<https://demo-trunk.ofbiz.apache.org/cmssite/cms/APACHE_OFBIZ_HTML>
is
no longer working (was still in August in trunk demo) I will
investigate
why
At http://markmail.org/message/5dbs3g3vbdfo7dlx I wrote
<<A moment I even thought about Attic for some unmaintained
components
(ebaystore?, googlebase?, googlecheckout?, jetty?, webpos?, ...),
WHO
cares?>>
But this is not a good idea. Obviously we have some responsabilities
with our users.
Now I still wonder about who is really using appserver, ebaystore,
googlebase, googlecheckou, oagis and jetty components...
This is what I can say so far
HTH
Jacques
Le 14/11/2014 14:20, Jacopo Cappellato a écrit :
It was a long discussion that was done in the public lists and I
wouldn't
want to rehash it (you have been part of it for sure): there were
concerns
and discussions about duplicated jars, poor quality code, stale
code,
files
with questionable licenses etc... on the other side there were
people
worried about removing features from the system etc...
I think it would be better to address each component individually
and,
since you would like to "cope with missing specialpurpose
components in
released packages", this is why I am asking you what are the
components
that should be included in the trunk/release branch/releases.
Jacopo
On Fri, Nov 14, 2014 at 1:55 PM, Jacques Le Roux <
jacques.le.r...@les7arts.com> wrote:
I think we need to be sure of what we are doing.
1st question, is why in the 1st place we did that? What pushed us
to
do so?
Jacques
Le 14/11/2014 12:47, Jacopo Cappellato a écrit :
What is your preference? Would you like to see them all in the
release
packages? Some of them only? Which ones?
On Fri, Nov 14, 2014 at 12:38 PM, Jacques Le Roux <
jacques.le.r...@les7arts.com> wrote:
This is the easiest part, I was more thinking about how much
is
downloaded
by users.
Anyway this was just an idea to help user to cope with missing
specialpurpose components in released packages.
Now a question comes to my mind, I don't clearly remember the
reasons we
decided to remove them. Why keeping them in the releases
branches
but not
not in released packages is not clear to me.
I believe Jacopo kind of answered at
http://markmail.org/message/
w3xw6lipifdeks3z
Actually we need to clarify 1st which components to keep active
in
release
branches. For now it seems only ecommerce which is for me too
restrictive.
And then discuss about why not doing the same in released
packages
(sorry
if I missed some arguments here).
For that we need first to exactly know which components affect
which
ones.
I believe at this stage we don't want to send any specialpurpose
component
to Attic, but this might be discussed also.
Jacques
Le 13/11/2014 22:51, Pierre Smits a écrit :
That is not difficult to assess. Do a download from trunk,
and
see how
many Mb's are transferred. Do a ./ant clean-all. Subsequently
remove all
hidden files in .svn folders. Finally do a zip of the cleaned
download
and
compare the original amount of Mb's with the size of the zip
file.
That
difference is what is saved on storage and transfer cost of
trunk
code.
Now multiply that with the number of branches you had in mind.
Verstuurd vanaf mijn iPad
Op 13 nov. 2014 om 22:32 heeft Jacques Le Roux <
jacques.le.r...@les7arts.com> het volgende geschreven:
Le 13/11/2014 21:25, Ron Wheeler a écrit :
Is it Apache's concern that while people may be free to
choose,
ASF
server capacity is not free nor unlimited?
I doubt that OFBiz really puts a big load on the ASF
infrastructure
but
users are not supposed to be downloading from the SVN.
They are supposed to get downloads from local mirrors.
You said it :) At the moment I don't fear any overload on
the svn
server
from users downloading from releases branches instead of
released
packages.
OFBiz is not Tomcat ;)
But I must say I have no measures, so you got a point
until-we/if-we-can
discover that. Because users can already do that, I think it's
fair to
use
this method as long as it's reasonable.
Of course, having that suggested in a TLP project could be
viewed
as an
abuse from the Board, but let's be pragmatic, numbers should
tell us
the
truth (if can get them)
That may be the practical side of Apache's urging to get
the
releases
following their guidelines.
Yes for Tomcat, HTTPD or such that's understandable. For
OFBiz I
"fear"
it's not a problem. Can we discuss with the board in case,
instead of
hiding behind unknown numbers?
Jacques
Ron
On 13/11/2014 3:13 PM, Jacques Le Roux wrote:
Le 13/11/2014 20:03, Ron Wheeler a écrit :
Does this solve ASF's issue about having users access the
main
servers?
I don't try to solve an issue, just to propose an
alternative.
It's a
free user choice, but with more elements
What do you put on the mirrors and how do you stop users
from
accessing the development SVN which is ASF's concern?
Things stay as they are, it's only that we inform our
users
than
another way is possible and we give them enough elements of
comparison to
choice, it's called freedom
Jacques
Ron
On 13/11/2014 1:55 PM, Jacques Le Roux wrote:
For the licence free issues (an other related stuff) we
could
put a
disclaimer in the wiki page where all alternatives would
be
explained
Jacques
Le 13/11/2014 12:38, Jacopo Cappellato a écrit :
In the past the ASF Board asked to the OFBiz PMC to fix
the
release
strategy of the project by providing officially voted
release
files
thru
the ASF mirrors: at that time we were pushing the users
to
get the
trunk.
Officially asking the user to use a release branch would
be
better
than the
trunk but would bring back similar concerns: an official
vote is
required
to publish a product to the outside of the project in
order to
guarantee
License free issues etc...
On Thu, Nov 13, 2014 at 11:38 AM, Jacques Le Roux <
jacques.le.r...@les7arts.com> wrote:
Hi,
In a recent user ML threadhttp://markmail.org/
message/ivjocjr2ull7lwqe I
suggested we could propose our users to use a release
branch
strategy
rather than downloaded packages.
And that we could expose this way of doing in our
download
page,
or maybe
better with a link to an explaining page (in details)
in the
wiki.
I know it's not the recommended way of doing at the ASF.
But we
all know
the OFBiz differences when compared with other TLPs
which
are
mostly libs,
and even mostly jars.
Most of us are actually using this way in their custom
projects
and I have
a feeling it would not only help our users but also us
to
support
them.
What do you think?
Jacques
--
Ron Wheeler
President
Artifact Software Inc
email: rwhee...@artifact-software.com
skype: ronaldmwheeler
phone: 866-970-2435, ext 102
--
Ron Wheeler
President
Artifact Software Inc
email: rwhee...@artifact-software.com
skype: ronaldmwheeler
phone: 866-970-2435, ext 102
