[gentoo-project] Re: Arch teams Resources (round 2) - General issues

[gentoo-project] Re: Arch teams Resources (round 2) - General issues

[Jorge Manuel B. S. Vicetto]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi.

Sometime ago after the issue of "slacking arches" was raised again at a
council meeting, I've started an internal discussion with Gentoo's arch
teams, to determine the current status of the arch teams and to check
how we could help improve it.

I'm now quoting one email I've sent in the last round, hoping to start a
public discussion about two issues raised internally.
I'm bcc'ing the arch teams so they're aware of this new thread and to
give them a chance to chime in.


- -------------

On 06-03-2011 14:40, Jorge Manuel B. S. Vicetto wrote:
...
> Please add a reply under each of the points you're interested. If you
> feel there's anything else I'm missing please add a reply about
> additional points.

7. General issues

Seems I forgot a few general issues that were raised in the previous thread.

 * lack of man power and burn-out

As some arch team members have mentioned, one of the most important
issues affecting arch teams is a general lack of man-power which
frequently leads to developer burn-out.
Anyone has any thoughts to share how to address this? Also, is there
anything we can do to improve the coordination between arch teams and is
there any way we can share some load between teams?

 * lack of statistics

Another issue raised by some developers is the lack of statistics so we
can know how many people use arches and so arch team members can have
any feedback about their work. I agree that numbers are important, so
I'd like to have a way to collect statistics about Gentoo's usage.
I've contacted Sebastian Pipping (sping) about this because of his smolt
project, but haven't heard back from him yet.
Does anyone have any suggestions about this? Does anyone want to work on
this?

- -------------

In the meantime Sebastian has got back to me about smolt.

Dear member of the Gentoo community, if you have any insight about these
issues or if you have anything you'd like to propose, we'd love to hear
from you.


- -- 
Regards,

Jorge Vicetto (jmbsvicetto) - jmbsvicetto at gentoo dot org
Gentoo- forums / Userrel / Devrel / KDE / Elections / RelEng
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=arXm
-----END PGP SIGNATURE-----

Re: [gentoo-project] Re: Arch teams Resources (round 2) - General issues

[David Abbott]

On Tue, May 10, 2011 at 3:02 PM, Jorge Manuel B. S. Vicetto
<jmbsvicetto@gentoo.org> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi.
>
> Sometime ago after the issue of "slacking arches" was raised again at a
> council meeting, I've started an internal discussion with Gentoo's arch
> teams, to determine the current status of the arch teams and to check
> how we could help improve it.
>
> I'm now quoting one email I've sent in the last round, hoping to start a
> public discussion about two issues raised internally.
> I'm bcc'ing the arch teams so they're aware of this new thread and to
> give them a chance to chime in.
>
>
> - -------------
>
> On 06-03-2011 14:40, Jorge Manuel B. S. Vicetto wrote:
> ...
>> Please add a reply under each of the points you're interested. If you
>> feel there's anything else I'm missing please add a reply about
>> additional points.
>
> 7. General issues
>
> Seems I forgot a few general issues that were raised in the previous thread.
>
>  * lack of man power and burn-out
Use Bugzilla to track which stable requests have been tested OK by
arch testers and send a weekly email report similar to the one Robin
sends out for additions and removals. Maybe we can use Target
Milestone: and add "Arch Tested X86" and so on. Most of my time is
spent on finding which packages need testing and if I knew some have
been tested good I could just move to the ones that need work.
[snip]
-- 
David Abbott (dabbott)
Gentoo
http://dev.gentoo.org/~dabbott/

[gentoo-project] Re: Arch teams Resources (round 2) - General issues

[Donnie Berkholz]

[-- Attachment #1: Type: text/plain, Size: 1673 bytes --]

On 19:02 Tue 10 May     , Jorge Manuel B. S. Vicetto wrote:
>  * lack of statistics
> 
> Another issue raised by some developers is the lack of statistics so 
> we can know how many people use arches and so arch team members can 
> have any feedback about their work. I agree that numbers are 
> important, so I'd like to have a way to collect statistics about 
> Gentoo's usage. I've contacted Sebastian Pipping (sping) about this 
> because of his smolt project, but haven't heard back from him yet. 
> Does anyone have any suggestions about this? Does anyone want to work 
> on this?

I've got access to some of our web logs, which gave me an idea. What if 
we estimate how many installations were successfully completed on each 
architecture by looking at visits to the last page of the handbook? This 
won't be a perfect measure for various reasons, but it's better than 
nothing.

Here's one day's worth of visits to part 11 of the install guide, which 
is the last page with actual content (adding users, cleaning up):

    190 x86
    150 amd64
     85 sparc
     83 ppc
     82 hppa
     59 ppc64
     54 alpha
     17 mips
     11 arm
      8 ia64

The reason non-x86/amd64 architectures are so close to x86/amd64 is that 
people using "alternative" archs rarely fail to install. The x86/amd64 
numbers start out about 2–3x higher at the beginning of the handbook, 
but most of them don't complete it.

It's tough to know how long an installation sticks around, but that many 
new installs per day is pretty impressive.


-- 
Thanks,
Donnie

Donnie Berkholz
Sr. Developer, Gentoo Linux
Blog: http://dberkholz.com

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-project] Re: Arch teams Resources (round 2) - General issues

["Paweł Hajdan, Jr."]

[-- Attachment #1: Type: text/plain, Size: 948 bytes --]

On 5/10/11 10:56 PM, Donnie Berkholz wrote:
> The reason non-x86/amd64 architectures are so close to x86/amd64 is that 
> people using "alternative" archs rarely fail to install. The x86/amd64 
> numbers start out about 2–3x higher at the beginning of the handbook, 
> but most of them don't complete it.

I just thought about a different possible explanation. When I'm setting
up new chroots or other test installs, I frequently only look up
information from the first pages of the handbook, like what's the exact
command to unpack the stage or what's the right order of steps. Setting
up the logger, cron and so on usually doesn't require referring to the
handbook.

> It's tough to know how long an installation sticks around, but that many 
> new installs per day is pretty impressive.

Yeah, and also it looks like the interest in arches like sparc, ppc,
hppa and so on is much higher than what I expected. That's great!


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 194 bytes --]

Re: [gentoo-project] Re: Arch teams Resources (round 2) - General issues

[Pandu Poluan]

On 2011-05-15, "Paweł Hajdan, Jr." <phajdan.jr@gentoo.org> wrote:
> On 5/10/11 10:56 PM, Donnie Berkholz wrote:
>> The reason non-x86/amd64 architectures are so close to x86/amd64 is that
>> people using "alternative" archs rarely fail to install. The x86/amd64
>> numbers start out about 2–3x higher at the beginning of the handbook,
>> but most of them don't complete it.
>
> I just thought about a different possible explanation. When I'm setting
> up new chroots or other test installs, I frequently only look up
> information from the first pages of the handbook, like what's the exact
> command to unpack the stage or what's the right order of steps. Setting
> up the logger, cron and so on usually doesn't require referring to the
> handbook.
>

... and there are people like me who prints out the handbook once
every handbook change. I'd use the exact same printout for, let's see,
7-8 installations.

So, my guess is that every 'handbook read' for x86/amd64 might
represent more than one installs.

>> It's tough to know how long an installation sticks around, but that many
>> new installs per day is pretty impressive.
>
> Yeah, and also it looks like the interest in arches like sparc, ppc,
> hppa and so on is much higher than what I expected. That's great!
>
>

IMO, the only way to get a precise number is to have a 'call-home'
script every install. But of course there's this question of ethics
and privacy...

--
Pandu E Poluan - IT Optimizer
My website: http://pandu.poluan.info/