CentOS docs - Jan 2015 - Pull Request wiki.c.o/AdditionalResources/Repositories

On Tue, Jan 13, 2015 at 10:04 AM, Karanbir Singh <mail-lists at karan.org>
wrote:
> On 01/09/2015 11:49 PM, Tom Sorensen wrote:
>> KB -- I made those changes several months ago (Sep/Oct I believe), with
>> discussion in IRC. This was after a spate of people in the main channel
>> having issues with Atomic (there's a name that's going to end
up causing
>> problems...) and the continued use of RPMForge/RepoForge, with no
>> indication that they're really really bad. As well as the
recognition of
>> the reality that there are a very few repos that are frequently
>> recommended (and, in the case of EPEL, now easily enabled in CentOS).
>
> I think we should do a bit of work and find a tangiable set of standards
> that a repo needs to meet in order to be 'endorsed' or rated at a
> certain level. Because at the moment it does seem to add value to a repo
> or two over others, based on personal opinion.
>
> I am willing to write code to do this validation, but were going to need
> a set of good rules to implement.
>
> regards and thanks
>
> - KB
>
Maybe it isn't code that needs to be generated.

I obviously wasn't around for any IRC discussion of this page and its
implications.
I don't imagine there is any sort of log of that discussion I could review.

Rhetorical questions and comments:
Is it true some of these repos exist because CentOS wasn't adequate for some
particular purpose, but someone thought they could provide a parallel resource
to easily install additional software?

I imagine another motivation would be a quasi-fork. That is, someone
was offended
that a particular decision was not to their satisfaction, but thought a repo
could be used to eliminate the bad decision and implement their proposal.
>From Tom's comment I infer there was a need to warn people away from
using some
repos that were consuming significant resources to help folks who had
trusted them.

It also appears that somehow Fedora's efforts have been mutually beneficial.

With these thoughts as background for my thinking, my brainstorm is that the
Special Interest Groups and spins may be the path to solving the need to have
additional non-CentOS resources and know they are sufficiently cooperative.

Proposal:
The Third Party Repositories section should not list any other repositories,
but should only note there are difficulties in making several independent
repositories safely usable and give a thorough explaination of what has happened
in the past without naming names.

The offer to help through Special Interest Groups and spins can then be noted.
This would be used to enter into a cooperative arrangement with each group that
wishes to start with the CentOS repositories and "improve" them.

If one implements a set of software tests for the quality of other folks work,
one is effectively committing to maintain those tests, which have no purpose
within the CentOS project. That means a CentOS project member is diverting time
from the CentOS project. Joint ventures, with agreements about how
they will work,
leverage the expertise of all participants for the benefit of all the projects.
Unilateral Quality Assurance tests consume time just generating
debates about the
quality of the tests, even before the maintenance issue comes up.

Atomic has a SIG. Maybe, EPEL needs one.

On Wed, Jan 14, 2015 at 03:09:01PM -0800, PatrickD Garvey
wrote:
> 
> Proposal:
> The Third Party Repositories section should not list any other
repositories,
> but should only note there are difficulties in making several independent
> repositories safely usable and give a thorough explaination of what has
happened
> in the past without naming names.
You are looking for problems to fix where there are none.  The overall state of
that page is and has been fine for many years.  EL requires external third-
party repos.  It has always been this way and it will always continue to
be the case.  Your proposal to remove the listings that are there now
serves no one and will only create more of a support burden on the
people that are volunteering their time.





							John
-- 
Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are, by
definition, not smart enough to debug it.

-- Brian W. Kernighan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL:
<http://lists.centos.org/pipermail/centos-docs/attachments/20150114/ac01acb4/attachment-0002.sig>

On Wed, Jan 14, 2015 at 3:26 PM, John R. Dennison <jrd at gerdesas.com>
wrote:
> On Wed, Jan 14, 2015 at 03:09:01PM -0800, PatrickD Garvey wrote:
>>
>> Proposal:
>> The Third Party Repositories section should not list any other
repositories,
>> but should only note there are difficulties in making several
independent
>> repositories safely usable and give a thorough explaination of what has
happened
>> in the past without naming names.
>
> You are looking for problems to fix where there are none.  The overall
state of
> that page is and has been fine for many years.  EL requires external third-
> party repos.  It has always been this way and it will always continue to
> be the case.  Your proposal to remove the listings that are there now
> serves no one and will only create more of a support burden on the
> people that are volunteering their time.
>
>                                                         John
I view your comments as an opportunity to understand an experience I
have yet to have. Please share which repository you use and how it
depends upon CentOS and how the CentOS community depends upon it.

I view the entire FLOSS community as interdependent. I hope to make
this page an asset for that interdependence. That's why I worked on
the link rot.

Karanbir seems to feel that certain phrases in the page unduly favor
some of the repositories and that requires an objective evaluation.

Please help us (me, especially) understand what we may be doing to the
detriment of your use of CentOS and thereby avoid that negative
result.

On 01/15/2015 01:09 AM, PatrickD Garvey wrote:
> On Tue, Jan 13, 2015 at 10:04 AM, Karanbir Singh <mail-lists at
karan.org> wrote:
>> On 01/09/2015 11:49 PM, Tom Sorensen wrote:
>>> KB -- I made those changes several months ago (Sep/Oct I believe),
with
>>> discussion in IRC. This was after a spate of people in the main
channel
>>> having issues with Atomic (there's a name that's going to
end up causing
>>> problems...) and the continued use of RPMForge/RepoForge, with no
>>> indication that they're really really bad. As well as the
recognition of
>>> the reality that there are a very few repos that are frequently
>>> recommended (and, in the case of EPEL, now easily enabled in
CentOS).
>> I think we should do a bit of work and find a tangiable set of
standards
>> that a repo needs to meet in order to be 'endorsed' or rated at
a
>> certain level. Because at the moment it does seem to add value to a
repo
>> or two over others, based on personal opinion.
>>
>> I am willing to write code to do this validation, but were going to
need
>> a set of good rules to implement.
>>
>> regards and thanks
>>
>> - KB
>>
> Maybe it isn't code that needs to be generated.
>
> I obviously wasn't around for any IRC discussion of this page and its
> implications.
> I don't imagine there is any sort of log of that discussion I could
review.
>
> Rhetorical questions and comments:
> Is it true some of these repos exist because CentOS wasn't adequate for
some
> particular purpose, but someone thought they could provide a parallel
resource
> to easily install additional software?
Yes, it is. Some people need newer versions for applications  than what 
the distro can provide. Other need complementary stuff which the distro 
or better said RHEL is not willing or not able to provide.


[...]
>
>  From Tom's comment I infer there was a need to warn people away from
using some
> repos that were consuming significant resources to help folks who had
> trusted them.
Right. Some repos lead to problems similar to:
/
Aug 28 14:27:15 <TheAlien>      hey there, im having trouble updating 
packages on my server. yum update gives a long list of needed updates 
and sumarises 'Install       3 Package(s)//
// / Upgrade     132 Package(s) / Remove        1 Package(s)' but then 
says 'package mysql-5.5.25-1.el5.remi.x86_64 (which is newer than 
mysql-5.0.95-5.el5_9.i386) is already installe//d' --//
//Aug 28 14:27:24 <TheAlien>      plus several messages like 'file 
/usr/bin/mysqlaccess from install of mysql-5.0.95-5.el5_9.i386 conflicts 
with file from package mysql-5.5.25-1.el5.r//
/
Others simply are no longer properly maintained, for instance still 
shipping older versions of applications which have known vulnerabilities.
//

[...]
>
> It also appears that somehow Fedora's efforts have been mutually
beneficial.
Yes, they have but this has nothing to do with the wiki page we speak 
about. Incidentally EPEL aims to a high standard so as to make it 
suitable for an enterprise-grade distro which CentOS aims to be and that 
is why we recommend to all others packagers to follow similar rules ( 
when possible  and if they make sense, of course ). However this is by 
no means a requirement that must be met in order to have a 3rd party 
repository listed in the CentOS wiki.

> With these thoughts as background for my thinking, my brainstorm is that
the
> Special Interest Groups and spins may be the path to solving the need to
have
> additional non-CentOS resources and know they are sufficiently cooperative.
>
> Proposal:
> The Third Party Repositories section should not list any other
repositories,
> but should only note there are difficulties in making several independent
> repositories safely usable and give a thorough explaination of what has
happened
> in the past without naming names.
I can only concur with what John has said. You are looking for problems 
to fix where there are none. 3 months ago Tom has made a great job in 
cleaning the page up ( and several of us assisted him when he asked for 
opinions ) and - AFAIK -  its current shape expresses the views of most 
of the regulars who provide help via the various CentOS support 
channels. If and when needed we modify the page but as it is now it is 
completely satisfactory for its purpose. Even if the feelings of some 
people get hurt by our opinions about the quality of their work ( i.e. 
of the packages they provide ).

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.centos.org/pipermail/centos-docs/attachments/20150115/f6dc6366/attachment-0002.html>

On Wed, Jan 14, 2015 at 7:39 PM, Manuel Wolfshant
<wolfy at nobugconsulting.ro> wrote:
> On 01/15/2015 01:09 AM, PatrickD Garvey wrote:
>
> On Tue, Jan 13, 2015 at 10:04 AM, Karanbir Singh <mail-lists at
karan.org>
> wrote:
>
> On 01/09/2015 11:49 PM, Tom Sorensen wrote:
>
> KB -- I made those changes several months ago (Sep/Oct I believe), with
> discussion in IRC. This was after a spate of people in the main channel
> having issues with Atomic (there's a name that's going to end up
causing
> problems...) and the continued use of RPMForge/RepoForge, with no
> indication that they're really really bad. As well as the recognition
of
> the reality that there are a very few repos that are frequently
> recommended (and, in the case of EPEL, now easily enabled in CentOS).
>
> I think we should do a bit of work and find a tangiable set of standards
> that a repo needs to meet in order to be 'endorsed' or rated at a
> certain level. Because at the moment it does seem to add value to a repo
> or two over others, based on personal opinion.
>
> I am willing to write code to do this validation, but were going to need
> a set of good rules to implement.
>
> regards and thanks
>
> - KB
>
> Maybe it isn't code that needs to be generated.
>
> I obviously wasn't around for any IRC discussion of this page and its
> implications.
> I don't imagine there is any sort of log of that discussion I could
review.
>
> Rhetorical questions and comments:
> Is it true some of these repos exist because CentOS wasn't adequate for
some
> particular purpose, but someone thought they could provide a parallel
> resource
> to easily install additional software?
>
> Yes, it is. Some people need newer versions for applications  than what the
> distro can provide. Other need complementary stuff which the distro or
> better said RHEL is not willing or not able to provide.
>
>
> [...]
>
> From Tom's comment I infer there was a need to warn people away from
using
> some
> repos that were consuming significant resources to help folks who had
> trusted them.
>
> Right. Some repos lead to problems similar to:
>
> Aug 28 14:27:15 <TheAlien>      hey there, im having trouble updating
> packages on my server. yum update gives a long list of needed updates and
> sumarises 'Install       3 Package(s)
>  / Upgrade     132 Package(s) / Remove        1 Package(s)' but then
says
> 'package mysql-5.5.25-1.el5.remi.x86_64 (which is newer than
> mysql-5.0.95-5.el5_9.i386) is already installed' --
> Aug 28 14:27:24 <TheAlien>      plus several messages like 'file
> /usr/bin/mysqlaccess from install of mysql-5.0.95-5.el5_9.i386 conflicts
> with file from package mysql-5.5.25-1.el5.r
>
> Others simply are no longer properly maintained, for instance still
shipping
> older versions of applications which have known vulnerabilities.
>
>
> [...]
>
> It also appears that somehow Fedora's efforts have been mutually
beneficial.
>
>
> Yes, they have but this has nothing to do with the wiki page we speak
about.
> Incidentally EPEL aims to a high standard so as to make it suitable for an
> enterprise-grade distro which CentOS aims to be and that is why we
recommend
> to all others packagers to follow similar rules ( when possible  and if
they
> make sense, of course ). However this is by no means a requirement that
must
> be met in order to have a 3rd party repository listed in the CentOS wiki.
>
>
> With these thoughts as background for my thinking, my brainstorm is that
the
> Special Interest Groups and spins may be the path to solving the need to
> have
> additional non-CentOS resources and know they are sufficiently cooperative.
>
> Proposal:
> The Third Party Repositories section should not list any other
repositories,
> but should only note there are difficulties in making several independent
> repositories safely usable and give a thorough explaination of what has
> happened
> in the past without naming names.
>
> I can only concur with what John has said. You are looking for problems to
> fix where there are none. 3 months ago Tom has made a great job in cleaning
> the page up ( and several of us assisted him when he asked for opinions )
> and - AFAIK -  its current shape expresses the views of most of the
regulars
> who provide help via the various CentOS support channels. If and when
needed
> we modify the page but as it is now it is completely satisfactory for its
> purpose. Even if the feelings of some people get hurt by our opinions about
> the quality of their work ( i.e. of the packages they provide ).
>
>
> _______________________________________________
> CentOS-docs mailing list
> CentOS-docs at centos.org
> http://lists.centos.org/mailman/listinfo/centos-docs
>
Thank you for the comments on my conjectures. I will integrate them
with my previous data about how this project works.

I guess I'm learning I should have let Karanbir Singh handle his own
suggestion that if the article seemed to grade the various repos,
someone needed to create an objective yardstick.

I hope you'll remember I offered some up-to-date contents for the link
anchors to those repositories. No one seems to have examined them for
validity. I have never used yum or git. I have no way to evaluate the
utility of the links that are now in the article. Should they be
improved?