[DISCUSS] Absent mentors

[Julian Hyde]

The incubator has an ongoing problem with lack of mentor engagement. Mentors are a crucial component of the incubation process. Incubation is the time when projects learn the Apache Way, and they cannot learn in a vacuum.

I’d like to discuss possible solutions to this problem. I’d like to hear from both podlings (PPMC members) and from IPMC members.

(By the way, it’s not just a problem for podlings. As a mentor, I am demoralized when I feel my co-mentors are not pulling their weight, and I get a little closer to burn-out.)

How to detect deadbeat mentors? One solution that has been discussed before is counting mentor sign-offs on podlings’ quarterly reports. Any project that received one or two sign-offs was deemed to be doing just fine. This is an imperfect metric.

Another remedy is to require podlings to be proactive: if they are not receiving adequate supervision, they should reach out to the IPMC and demand a change in mentors. The problem is, podlings have by definition not been through incubation before, so do not know what to expect. They don’t want to rock the boat.

Have your mentors been helpful and responsive? If not, describe what advice or help

It isn't too onerous for the podling, and only embarrasses mentors who deserve to be embarrassed.

What to do about deadbeat mentors? The current thinking is that every project should have three mentors, and if at least one of them is active, that’s OK. I think that the “rule of 3” actually makes the problem worse. It’s difficult to find three motivated individuals (or find enough work for them to do), so a podling will inevitably have one or two inactive mentors. It has become the norm that most mentors are inactive.

I propose that we get rid of the rule of 3. If mentors are not active, they should be encouraged to step down, and if they don’t, the IPMC should remove them. If this leaves the podling with zero or one mentors, then IPMC can step in and appoint new mentors. A podling with two active mentors is probably doing just fine.

Is this problem as serious as I think it is? Would my proposed solutions help?

Julian


---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Ted Dunning]

I think the problem is serious. I also think that signoff rate is a better
metric in practice than it seems it would be.

Adding the additional metric seems like a small step that could help.

Being aggressive about removing non-mentors is a very good idea. It is best
if mentors remove themselves, but it is imperative that the incubator has a
realistic idea about how many mentors there really are.

Post by Julian Hyde
The incubator has an ongoing problem with lack of mentor engagement.
Mentors are a crucial component of the incubation process. Incubation is
the time when projects learn the Apache Way, and they cannot learn in a
vacuum.
I’d like to discuss possible solutions to this problem. I’d like to hear
from both podlings (PPMC members) and from IPMC members.
(By the way, it’s not just a problem for podlings. As a mentor, I am
demoralized when I feel my co-mentors are not pulling their weight, and I
get a little closer to burn-out.)
How to detect deadbeat mentors? One solution that has been discussed
before is counting mentor sign-offs on podlings’ quarterly reports. Any
project that received one or two sign-offs was deemed to be doing just
fine. This is an imperfect metric.
Another remedy is to require podlings to be proactive: if they are not
receiving adequate supervision, they should reach out to the IPMC and
demand a change in mentors. The problem is, podlings have by definition not
been through incubation before, so do not know what to expect. They don’t
want to rock the boat.
I propose another solution. Let’s add a question to the podling report

Have your mentors been helpful and responsive? If not, describe what

advice or help
It isn't too onerous for the podling, and only embarrasses mentors who
deserve to be embarrassed.
What to do about deadbeat mentors? The current thinking is that every
project should have three mentors, and if at least one of them is active,
that’s OK. I think that the “rule of 3” actually makes the problem worse.
It’s difficult to find three motivated individuals (or find enough work for
them to do), so a podling will inevitably have one or two inactive mentors.
It has become the norm that most mentors are inactive.
I propose that we get rid of the rule of 3. If mentors are not active,
they should be encouraged to step down, and if they don’t, the IPMC should
remove them. If this leaves the podling with zero or one mentors, then IPMC
can step in and appoint new mentors. A podling with two active mentors is
probably doing just fine.
Is this problem as serious as I think it is? Would my proposed solutions help?
Julian
---------------------------------------------------------------------

[Dave Fisher]

Hi -

Inline - responses to both.

Post by Ted Dunning
I think the problem is serious. I also think that signoff rate is a better
metric in practice than it seems it would be.

I agree that it is quite serious.

Post by Ted Dunning
Adding the additional metric seems like a small step that could help.
Being aggressive about removing non-mentors is a very good idea. It is best
if mentors remove themselves, but it is imperative that the incubator has a
realistic idea about how many mentors there really are.

Post by Julian Hyde
The incubator has an ongoing problem with lack of mentor engagement.
Mentors are a crucial component of the incubation process. Incubation is
the time when projects learn the Apache Way, and they cannot learn in a
vacuum.
I’d like to discuss possible solutions to this problem. I’d like to hear
from both podlings (PPMC members) and from IPMC members.
(By the way, it’s not just a problem for podlings. As a mentor, I am
demoralized when I feel my co-mentors are not pulling their weight, and I
get a little closer to burn-out.)
How to detect deadbeat mentors? One solution that has been discussed
before is counting mentor sign-offs on podlings’ quarterly reports. Any
project that received one or two sign-offs was deemed to be doing just
fine. This is an imperfect metric.

Sign-off means that Mentors are doing the least they can do which is better than nothing.

Post by Ted Dunning

Post by Julian Hyde
Another remedy is to require podlings to be proactive: if they are not
receiving adequate supervision, they should reach out to the IPMC and
demand a change in mentors. The problem is, podlings have by definition not
been through incubation before, so do not know what to expect. They don’t
want to rock the boat.
I propose another solution. Let’s add a question to the podling report

Have your mentors been helpful and responsive? If not, describe what

advice or help
It isn't too onerous for the podling, and only embarrasses mentors who
deserve to be embarrassed.

Mentors need to be careful of their workload. Having the correct mentors for the community is also important. What I mean about correct will be below.

Post by Ted Dunning

Post by Julian Hyde
What to do about deadbeat mentors? The current thinking is that every
project should have three mentors, and if at least one of them is active,
that’s OK. I think that the “rule of 3” actually makes the problem worse.
It’s difficult to find three motivated individuals (or find enough work for
them to do), so a podling will inevitably have one or two inactive mentors.
It has become the norm that most mentors are inactive.

The rule of 3 was so that there were enough Mentors to provide the 3 +1 (Binding Votes) before we get to the IPMC Vote. Thankfully we have a few experts on the IPMC who are doing the required Voting and releases aren’t getting held up.

Post by Ted Dunning

Post by Julian Hyde
I propose that we get rid of the rule of 3. If mentors are not active,
they should be encouraged to step down, and if they don’t, the IPMC should
remove them. If this leaves the podling with zero or one mentors, then IPMC
can step in and appoint new mentors. A podling with two active mentors is
probably doing just fine.

We really need to have the correct Mentors. I feel uncomfortable as one of only two mentors on Daffodil. It is an experiment of having Two.

Post by Ted Dunning

Post by Julian Hyde
Is this problem as serious as I think it is? Would my proposed solutions help?

I think that we need to also discuss what voting +1 to accept a podling should mean. The value currently is that everyone just +1s because the podling is “cool”.

I think we should discuss these ideas:

(1) Adding more questions to the podling:
- about the number of dependencies. If a lot then we want Mentors who like that part of the process.
- about any registered trademarks. If so then a Mentor with trademark experience is needed.

(2) Think about whether a +1 (binding) VOTE means the IPMC member is willing to Mentor. If we can’t get enough Mentors then we can’t accept a podling.

Regards,
Dave

Post by Ted Dunning

Post by Julian Hyde
Julian
---------------------------------------------------------------------

[Roman Shaposhnik]

Post by Ted Dunning
I think the problem is serious. I also think that signoff rate is a better
metric in practice than it seems it would be.

That problem is indeed pretty serious and also pretty chronic.

As for the metric -- I really think that using mentor turnout on release
voting threads will serve us well.

Post by Ted Dunning
Adding the additional metric seems like a small step that could help.
Being aggressive about removing non-mentors is a very good idea. It is best
if mentors remove themselves, but it is imperative that the incubator has a
realistic idea about how many mentors there really are.

Big +1 on the above. Perhaps if we:
1. get a clear indication on release vote turnout (as part of
Incubator report)
2. add to it the sign-off turnout

We can start at least nagging unresponsive mentors to begin
with and if behaviour doesn't improve -- suggest that podlings
start looking for a replacement.

Thanks,
Roman.

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Willem Jiang]

+1 for taking consideration of release vote.
It's quite important process of the podling.


Willem Jiang

Blog: http://willemjiang.blogspot.com (English)
http://jnn.iteye.com (Chinese)
Twitter: willemjiang
Weibo: 姜宁willem

Post by Ted Dunning

Post by Ted Dunning
I think the problem is serious. I also think that signoff rate is a

better

Post by Ted Dunning
metric in practice than it seems it would be.

That problem is indeed pretty serious and also pretty chronic.
As for the metric -- I really think that using mentor turnout on release
voting threads will serve us well.

Post by Ted Dunning
Adding the additional metric seems like a small step that could help.
Being aggressive about removing non-mentors is a very good idea. It is

best

Post by Ted Dunning
if mentors remove themselves, but it is imperative that the incubator

has a

Post by Ted Dunning
realistic idea about how many mentors there really are.

1. get a clear indication on release vote turnout (as part of
Incubator report)
2. add to it the sign-off turnout
We can start at least nagging unresponsive mentors to begin
with and if behaviour doesn't improve -- suggest that podlings
start looking for a replacement.
Thanks,
Roman.
---------------------------------------------------------------------

[Justin Mclean]

Hi,

Post by Roman Shaposhnik
As for the metric -- I really think that using mentor turnout on release
voting threads will serve us well.

My concern with using that as a metric is people will just vote +1 without doing a thorough check and we may end up with more releases with issues.

Possibly a better metric is how many mentors voted something other than +1 on a RC, most releases (other than very simple ones) go through a couple of RCs before coming to the IPMC.

Another metric is which project releases get a -1 in the IPMC as those issue should of been caught by the projects mentors.

Thanks,
Justin


---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Willem Jiang]

Yeah, it's easy to vote +1, but vote -1 take lots of time to go through the
kits.

To be honestly, I learned a lot of License and Notice stuff thing from
Justin's -1 vote, I really appreciate that.


Willem Jiang

Blog: http://willemjiang.blogspot.com (English)
http://jnn.iteye.com (Chinese)
Twitter: willemjiang
Weibo: 姜宁willem

Post by Justin Mclean
Hi,

Post by Roman Shaposhnik
As for the metric -- I really think that using mentor turnout on release
voting threads will serve us well.

My concern with using that as a metric is people will just vote +1 without
doing a thorough check and we may end up with more releases with issues.
Possibly a better metric is how many mentors voted something other than +1
on a RC, most releases (other than very simple ones) go through a couple of
RCs before coming to the IPMC.
Another metric is which project releases get a -1 in the IPMC as those
issue should of been caught by the projects mentors.
Thanks,
Justin
---------------------------------------------------------------------

[吴晟 Sheng Wu]

Agree, Justin's -1 really helps and pushes SkyWalking to do right things in releasing.
Thanks a lot, even we delayed our release two times (2 weeks)


:) Thanks Justin.


Sheng Wu



---Original---
From: "Willem Jiang"<***@gmail.com>
Date: Sun, Apr 1, 2018 16:17 PM
To: "general"<***@incubator.apache.org>;
Subject: Re: [DISCUSS] Absent mentors


Yeah, it's easy to vote +1, but vote -1 take lots of time to go through the
kits.

To be honestly, I learned a lot of License and Notice stuff thing from
Justin's -1 vote, I really appreciate that.


Willem Jiang

Blog: http://willemjiang.blogspot.com (English)
http://jnn.iteye.com (Chinese)
Twitter: willemjiang
Weibo: œªÄþwillem

Post by Justin Mclean
Hi,

Post by Roman Shaposhnik
As for the metric -- I really think that using mentor turnout on release
voting threads will serve us well.

My concern with using that as a metric is people will just vote +1 without
doing a thorough check and we may end up with more releases with issues.
Possibly a better metric is how many mentors voted something other than +1
on a RC, most releases (other than very simple ones) go through a couple of
RCs before coming to the IPMC.
Another metric is which project releases get a -1 in the IPMC as those
issue should of been caught by the projects mentors.
Thanks,
Justin
---------------------------------------------------------------------

[Hen]

+1 to flagging mentor absence; and I like making it automated otherwise
it's not going to happen (or rather, it'll be up to a podling to flag it
and they're unlikely to feel comfortable doing so).

Justin's two metrics are interesting to me as I (kinda) don't view either
of those as mentor responsibilities.

For me a mentor is:

1) Someone who prods the podling to move along to the next step in its path
in the incubator.
2) Someone who monitors the list for general 'flow'. Is dev happening, does
it seem to be moving along nicely etc.
3) Someone who joins in on exceptional/abnormal threads (ie: when #2 hits
bumps).
4) Someone who deep dives early on in the podlings life to get things
moving.
5) Someone who is reviewing the podling's board report.

For me a mentor is not required to be:

1) Someone who is a coder on the project, or deep in the technology in
question.
2) Someone who votes on how the project chooses to develop, or
3) Someone who votes on the technical choices in the project,
4) Or, someone who is deep diving into release votes once a general cadence
has been set (beyond that need for IPMC +1s).

When a project is ready to graduate, the mentors of the project should be
doing practically nothing with their mentor hat on.

--

Given all that, I would definitely lean towards automated flagging for
mentors when not reviewing the podling's board report.

I'd also have an urge for us to define more specific milestones within
incubation, with more expectation on mentor activity for podlings at
earlier milestones.

Hen

Post by Justin Mclean
Hi,

Post by Roman Shaposhnik
As for the metric -- I really think that using mentor turnout on release
voting threads will serve us well.

My concern with using that as a metric is people will just vote +1 without
doing a thorough check and we may end up with more releases with issues.
Possibly a better metric is how many mentors voted something other than +1
on a RC, most releases (other than very simple ones) go through a couple of
RCs before coming to the IPMC.
Another metric is which project releases get a -1 in the IPMC as those
issue should of been caught by the projects mentors.
Thanks,
Justin
---------------------------------------------------------------------

[Justin Mclean]

Hi,

Post by Hen
Justin's two metrics are interesting to me as I (kinda) don't view either
of those as mentor responsibilities.

Interesting how do you projects you mentor get to work out release profile and what need to done re license and notice? Or are they simple boiler plate license and notice releases and/or already involve ASF people who have those skills and know what to look for?

Post by Hen
For me a mentor is: <snip>

For sure it not just helping out with releases, but some of those things you mentioned are hard to measure in any real way.

Post by Hen
I'd also have an urge for us to define more specific milestones within
incubation, with more expectation on mentor activity for podlings at
earlier milestones.

What would those milestones be? One of them I assume would be a podling first release and vote on the IPMC?

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Hen]

Post by Justin Mclean
Hi,

Post by Hen
Justin's two metrics are interesting to me as I (kinda) don't view either
of those as mentor responsibilities.

Interesting how do you projects you mentor get to work out release profile
and what need to done re license and notice? Or are they simple boiler
plate license and notice releases and/or already involve ASF people who
have those skills and know what to look for?

I subsequently proviso'd it with: "once a general cadence has been set".

Post by Justin Mclean

Post by Hen
For me a mentor is: <snip>

For sure it not just helping out with releases, but some of those things
you mentioned are hard to measure in any real way.

Post by Hen
I'd also have an urge for us to define more specific milestones within
incubation, with more expectation on mentor activity for podlings at
earlier milestones.

What would those milestones be? One of them I assume would be a podling
first release and vote on the IPMC?

Yeah. Half-arsed guess (and these start to feel more like badges in that
they can be in parallel, but I think there is a typical/ideal order):

* Minimum Infrastructure Setup complete
* All ICLAs signed/IPMC setup
* Licensing Sorted
* First (no major issue) Release Complete
* X% new contributors converted to IPMC
* Graduation delta identified (a thread in which general@ agrees on the
remaining items before graduation)

Tempted to throw in these, though feeling more badge-like:

* Website complete (brand, security note, link to Foundation etc)
* Apache Blogged
* Conference presentation
* Board report complete, no comments/concerns

Hen

[Jean-Baptiste Onofré]

That's a fair concern.

I think that the podling should mention in podling report if they need help from
mentors and they don't have it (because the mentor is not active).

Regards
JB

Post by Julian Hyde
The incubator has an ongoing problem with lack of mentor engagement. Mentors are a crucial component of the incubation process. Incubation is the time when projects learn the Apache Way, and they cannot learn in a vacuum.
I’d like to discuss possible solutions to this problem. I’d like to hear from both podlings (PPMC members) and from IPMC members.
(By the way, it’s not just a problem for podlings. As a mentor, I am demoralized when I feel my co-mentors are not pulling their weight, and I get a little closer to burn-out.)
How to detect deadbeat mentors? One solution that has been discussed before is counting mentor sign-offs on podlings’ quarterly reports. Any project that received one or two sign-offs was deemed to be doing just fine. This is an imperfect metric.
Another remedy is to require podlings to be proactive: if they are not receiving adequate supervision, they should reach out to the IPMC and demand a change in mentors. The problem is, podlings have by definition not been through incubation before, so do not know what to expect. They don’t want to rock the boat.

Have your mentors been helpful and responsive? If not, describe what advice or help

It isn't too onerous for the podling, and only embarrasses mentors who deserve to be embarrassed.
What to do about deadbeat mentors? The current thinking is that every project should have three mentors, and if at least one of them is active, that’s OK. I think that the “rule of 3” actually makes the problem worse. It’s difficult to find three motivated individuals (or find enough work for them to do), so a podling will inevitably have one or two inactive mentors. It has become the norm that most mentors are inactive.
I propose that we get rid of the rule of 3. If mentors are not active, they should be encouraged to step down, and if they don’t, the IPMC should remove them. If this leaves the podling with zero or one mentors, then IPMC can step in and appoint new mentors. A podling with two active mentors is probably doing just fine.
Is this problem as serious as I think it is? Would my proposed solutions help?
Julian
---------------------------------------------------------------------

[Sid Anand]

Thanks for raising this Julian.

I agree with your take on this situation and with your proposals.
-s (Sid)

Post by Jean-Baptiste Onofré
That's a fair concern.
I think that the podling should mention in podling report if they need help from
mentors and they don't have it (because the mentor is not active).
Regards
JB

Post by Julian Hyde
The incubator has an ongoing problem with lack of mentor engagement.

Mentors are a crucial component of the incubation process. Incubation is
the time when projects learn the Apache Way, and they cannot learn in a
vacuum.

Post by Julian Hyde
I’d like to discuss possible solutions to this problem. I’d like to hear

from both podlings (PPMC members) and from IPMC members.

Post by Julian Hyde
(By the way, it’s not just a problem for podlings. As a mentor, I am

demoralized when I feel my co-mentors are not pulling their weight, and I
get a little closer to burn-out.)

Post by Julian Hyde
How to detect deadbeat mentors? One solution that has been discussed

before is counting mentor sign-offs on podlings’ quarterly reports. Any
project that received one or two sign-offs was deemed to be doing just
fine. This is an imperfect metric.

Post by Julian Hyde
Another remedy is to require podlings to be proactive: if they are not

receiving adequate supervision, they should reach out to the IPMC and
demand a change in mentors. The problem is, podlings have by definition not
been through incubation before, so do not know what to expect. They don’t
want to rock the boat.

Post by Julian Hyde
I propose another solution. Let’s add a question to the podling report

Have your mentors been helpful and responsive? If not, describe what

advice or help

Post by Julian Hyde
It isn't too onerous for the podling, and only embarrasses mentors who

deserve to be embarrassed.

Post by Julian Hyde
What to do about deadbeat mentors? The current thinking is that every

project should have three mentors, and if at least one of them is active,
that’s OK. I think that the “rule of 3” actually makes the problem worse.
It’s difficult to find three motivated individuals (or find enough work for
them to do), so a podling will inevitably have one or two inactive mentors.
It has become the norm that most mentors are inactive.

Post by Julian Hyde
I propose that we get rid of the rule of 3. If mentors are not active,

they should be encouraged to step down, and if they don’t, the IPMC should
remove them. If this leaves the podling with zero or one mentors, then IPMC
can step in and appoint new mentors. A podling with two active mentors is
probably doing just fine.

Post by Julian Hyde
Is this problem as serious as I think it is? Would my proposed solutions

help?

Post by Julian Hyde
Julian
---------------------------------------------------------------------

--
Jean-Baptiste Onofré
http://blog.nanthrax.net
Talend - http://www.talend.com
---------------------------------------------------------------------

[Bertrand Delacretaz]

Hi,

Have your mentors been helpful and responsive? If not, describe what advice or help
you needed, or need:..

I like it but would phrase it slightly differently:

"Have your mentors been helpful and responsive or are things falling
through the cracks? In the latter case, please list open issues that
need to be addressed".

...I propose that we get rid of the rule of 3. If mentors are not active, they should
be encouraged to step down, and if they don’t, the IPMC should remove them...

+1 but for starting a podling I still like having at least two mentors.

-Bertrand

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[吴晟 Sheng Wu]

+1 for the needs of active mentors. As a new podling project PPMC, I and our project committers learn a lot from out mentors about the Apache way. The mentor and communication stuffs need a lot of time, so I totally agree, active mentors are very important for podling project members.


------------------
Sheng Wu
Apache SkyWalking original creator and PPMC member







------------------ Original ------------------
From: "Julian Hyde"<***@apache.org>;
Date: Thu, Mar 29, 2018 06:20 AM
To: "general"<***@incubator.apache.org>;

Subject: [DISCUSS] Absent mentors



The incubator has an ongoing problem with lack of mentor engagement. Mentors are a crucial component of the incubation process. Incubation is the time when projects learn the Apache Way, and they cannot learn in a vacuum.

I¡¯d like to discuss possible solutions to this problem. I¡¯d like to hear from both podlings (PPMC members) and from IPMC members.

(By the way, it¡¯s not just a problem for podlings. As a mentor, I am demoralized when I feel my co-mentors are not pulling their weight, and I get a little closer to burn-out.)

How to detect deadbeat mentors? One solution that has been discussed before is counting mentor sign-offs on podlings¡¯ quarterly reports. Any project that received one or two sign-offs was deemed to be doing just fine. This is an imperfect metric.

Another remedy is to require podlings to be proactive: if they are not receiving adequate supervision, they should reach out to the IPMC and demand a change in mentors. The problem is, podlings have by definition not been through incubation before, so do not know what to expect. They don¡¯t want to rock the boat.

Have your mentors been helpful and responsive? If not, describe what advice or help

It isn't too onerous for the podling, and only embarrasses mentors who deserve to be embarrassed.

What to do about deadbeat mentors? The current thinking is that every project should have three mentors, and if at least one of them is active, that¡¯s OK. I think that the ¡°rule of 3¡± actually makes the problem worse. It¡¯s difficult to find three motivated individuals (or find enough work for them to do), so a podling will inevitably have one or two inactive mentors. It has become the norm that most mentors are inactive.

I propose that we get rid of the rule of 3. If mentors are not active, they should be encouraged to step down, and if they don¡¯t, the IPMC should remove them. If this leaves the podling with zero or one mentors, then IPMC can step in and appoint new mentors. A podling with two active mentors is probably doing just fine.

Is this problem as serious as I think it is? Would my proposed solutions help?

Julian


---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Jim Jagielski]

Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Shane Curcuru]

Post by Jim Jagielski
Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.

Presuming podlings.xml is kept updated:


https://svn.apache.org/repos/asf/incubator/public/trunk/content/podlings.xml

That now sorts by @status first, so current podlings are on top. That's
separate from whimsy cross-checks of what's listed in board reports and
actual signoffs.

[Dave Fisher]

Post by Shane Curcuru

Post by Jim Jagielski
Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.

https://svn.apache.org/repos/asf/incubator/public/trunk/content/podlings.xml
separate from whimsy cross-checks of what's listed in board reports and
actual signoffs.

The clutch runs periodically and generates https://svn.apache.org/repos/asf/incubator/public/trunk/content/clutch/clutchm.ent <https://svn.apache.org/repos/asf/incubator/public/trunk/content/clutch/clutchm.ent>

This file has each mentor listed. There are duplicates lines when people have variations on their names in the date in podlings.xml. Probably there needs to be a fix to use username attribute and not the mentor value in the clutch.py program.

I can see people in the list who probably don’t even think they are a mentor to a podling.

Should we send a note to all mentors recorded asking if they are still engaged?

Regards,
Dave

Post by Shane Curcuru
--
- Shane
Director & Member
The Apache Software Foundation
---------------------------------------------------------------------

[Luciano Resende]

Clutch is available at :
http://incubator.apache.org/clutch/

See Mentors project section for a list of podlings and its mentors.

Post by Jim Jagielski
Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.
https://svn.apache.org/repos/asf/incubator/public/trunk/
content/podlings.xml
separate from whimsy cross-checks of what's listed in board reports and
actual signoffs.
The clutch runs periodically and generates https://svn.apache.
org/repos/asf/incubator/public/trunk/content/clutch/clutchm.ent
This file has each mentor listed. There are duplicates lines when people
have variations on their names in the date in podlings.xml. Probably there
needs to be a fix to use username attribute and not the mentor value in the
clutch.py program.
I can see people in the list who probably don’t even think they are a
mentor to a podling.
Should we send a note to all mentors recorded asking if they are still engaged?
Regards,
Dave
--
- Shane
Director & Member
The Apache Software Foundation
---------------------------------------------------------------------

[Mark Thomas]

Post by Dave Fisher

Post by Shane Curcuru

Post by Jim Jagielski
Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.

https://svn.apache.org/repos/asf/incubator/public/trunk/content/podlings.xml
separate from whimsy cross-checks of what's listed in board reports and
actual signoffs.

The clutch runs periodically and
generates https://svn.apache.org/repos/asf/incubator/public/trunk/content/clutch/clutchm.ent 
This file has each mentor listed. There are duplicates lines when people
have variations on their names in the date in podlings.xml. Probably
there needs to be a fix to use username attribute and not the mentor
value in the clutch.py program.
I can see people in the list who probably don’t even think they are a
mentor to a podling.
Should we send a note to all mentors recorded asking if they are still engaged?

That seems like a sensible first step to me. I suggest asking for
explicit confirmation that they wish to continue as a mentor. That
should hopefully give us a better picture of how things currently stand.

Mark

Post by Dave Fisher
Regards,
Dave

Post by Shane Curcuru
--
- Shane
 Director & Member
 The Apache Software Foundation
---------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Dave Fisher]

Hi -

Post by Mark Thomas

Post by Dave Fisher

Post by Shane Curcuru

Post by Jim Jagielski
Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.

https://svn.apache.org/repos/asf/incubator/public/trunk/content/podlings.xml
separate from whimsy cross-checks of what's listed in board reports and
actual signoffs.

The clutch runs periodically and
generates https://svn.apache.org/repos/asf/incubator/public/trunk/content/clutch/clutchm.ent
This file has each mentor listed. There are duplicates lines when people
have variations on their names in the date in podlings.xml. Probably
there needs to be a fix to use username attribute and not the mentor
value in the clutch.py program.
I can see people in the list who probably don’t even think they are a
mentor to a podling.
Should we send a note to all mentors recorded asking if they are still engaged?

That seems like a sensible first step to me. I suggest asking for
explicit confirmation that they wish to continue as a mentor. That
should hopefully give us a better picture of how things currently stand.

I have extracted a list of mentors, emails, and podlings as an ODS file. Attached:

[John D. Ament]

Post by Shane Curcuru
Hi -

Post by Mark Thomas

Post by Dave Fisher

Post by Shane Curcuru

Post by Jim Jagielski
Would it be possible to generate a short list of all current
mentors for all current podlings to see how many podlings
each mentor is signed up for? That would be a good metric
to know.

https://svn.apache.org/repos/asf/incubator/public/trunk/content/podlings.xml
That's

Post by Mark Thomas

Post by Dave Fisher

Post by Shane Curcuru
separate from whimsy cross-checks of what's listed in board reports and
actual signoffs.

The clutch runs periodically and
generates

https://svn.apache.org/repos/asf/incubator/public/trunk/content/clutch/clutchm.ent

Post by Mark Thomas

Post by Dave Fisher
This file has each mentor listed. There are duplicates lines when people
have variations on their names in the date in podlings.xml. Probably
there needs to be a fix to use username attribute and not the mentor
value in the clutch.py program.
I can see people in the list who probably don’t even think they are a
mentor to a podling.
Should we send a note to all mentors recorded asking if they are still engaged?

That seems like a sensible first step to me. I suggest asking for
explicit confirmation that they wish to continue as a mentor. That
should hopefully give us a better picture of how things currently stand.

There are 101 mentors currently on the roster.
52 have 1 podling
24 have 2 podlings
17 have 3 podlings
3 have 4 podlings
2 have 5 podlings
2 have 7 podlings
1 has 8 podlings
The next step would be to discuss the email to send.

Agreed. Anyone want to take a stab at the email text?

Post by Shane Curcuru
Regards,
Dave

[Justin Mclean]

Hi,

Taking a closer look at the data [1] (only accessible by IPMC and ASF members sorry) it may be a good idea to split the mentors/podlings into groups and draft a different email for each? For instance:
- Mentors who are not signing off reports at all or only occasionally
- Mentors who have missed a few as they it looks like they have too many podlings on their plate
- Podlings that have a less that 50% (say) sign off rate

Thanks,
Justin

1. https://whimsy.apache.org/incubator/signoff.cgi#bypodling
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

Coming back to this old thread and coming up with some simple draft emails. Feedback welcome.

Post by Justin Mclean
- Mentors who are not signing off reports at all or only occasionally

(sent to mentor)

I can see you are are the mentor of XXX and haven’t been signing off reports on a frequent basis. Life happens and things change and you may no longer be in the position to be able to put in the same effort that you once did to mentor this podling. If that is the case perhaps you would consider stepping down as mentor and/or help find the podling another mentor.

By my rough count there about 90 mentors :-( who are in this position.

Post by Justin Mclean
- Mentors who have missed a few as they it looks like they have too many podlings on their plate

(sent to mentor)

I can see you are are the mentor of XXX, YYY and ZZZ and haven’t been signing off all reports on a frequent basis. While we appreciate your enthusiasm for helping out at the incubator PMC, perhaps you’ve got busy with other things or you've just taken on too many podlings to mentor. Would it be better for you and your podlings if you cut back a little on the number you mentor and let someone else handle the load?

By my rough count there about 15 mentors who are in this position.

Post by Justin Mclean
- Podlings that have a less that 50% (say) sign off rate

(sent to dev or private list)

It looks like your mentors haven’t been signing off board reports as much as they should. Can you tell us if your mentors are still active or if you need help in any other way?

By my rough count there about 30 podlings who are in this position. That more than half the current podlings.

Those numbers are a lot larger that I expected, but I think we've known that. So they emails might get some mentors re-engaged and they may get some mentors to resign. The question may be where do we find mentors to help podlings out if a large number do decide to step down.

Thanks,
Justin

1. https://whimsy.apache.org/incubator/signoff.cgi#bypodling


---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Julian Hyde]

There is a strong presumption here that a mentor is fulfilling their duty if and only if they sign the reports.

I’m skeptical. I can think of cases where an engaged mentor does not sign reports (because someone else has done it), and vice versa where a disengaged mentor shows up once a quarter to sign the report.

It may be a bit more difficult to measure, but I think a better indicator would be the number of messages sent by the mentor to dev and private lists. It indicates both time spent writing the message and also time spent keeping up with the email traffic to know when the podling needs help.

Julian

Post by Justin Mclean
Hi,
Coming back to this old thread and coming up with some simple draft emails. Feedback welcome.

Post by Justin Mclean
- Mentors who are not signing off reports at all or only occasionally

(sent to mentor)
I can see you are are the mentor of XXX and haven’t been signing off reports on a frequent basis. Life happens and things change and you may no longer be in the position to be able to put in the same effort that you once did to mentor this podling. If that is the case perhaps you would consider stepping down as mentor and/or help find the podling another mentor.
By my rough count there about 90 mentors :-( who are in this position.

Post by Justin Mclean
- Mentors who have missed a few as they it looks like they have too many podlings on their plate

(sent to mentor)
I can see you are are the mentor of XXX, YYY and ZZZ and haven’t been signing off all reports on a frequent basis. While we appreciate your enthusiasm for helping out at the incubator PMC, perhaps you’ve got busy with other things or you've just taken on too many podlings to mentor. Would it be better for you and your podlings if you cut back a little on the number you mentor and let someone else handle the load?
By my rough count there about 15 mentors who are in this position.

Post by Justin Mclean
- Podlings that have a less that 50% (say) sign off rate

(sent to dev or private list)
It looks like your mentors haven’t been signing off board reports as much as they should. Can you tell us if your mentors are still active or if you need help in any other way?
By my rough count there about 30 podlings who are in this position. That more than half the current podlings.
Those numbers are a lot larger that I expected, but I think we've known that. So they emails might get some mentors re-engaged and they may get some mentors to resign. The question may be where do we find mentors to help podlings out if a large number do decide to step down.
Thanks,
Justin
1. https://whimsy.apache.org/incubator/signoff.cgi#bypodling
---------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

Post by Julian Hyde
There is a strong presumption here that a mentor is fulfilling their duty if and only if they sign the reports.

It’s an imperfect metric but one we have easy access to. If we send out an email asking for if you’re AWOL and a mentor responds that they are active then all is good. Of course it’s not going to catch mentors who sign reports and do nothing else but I hope there’s only a few incidences of that.

Post by Julian Hyde
I’m skeptical. I can think of cases where an engaged mentor does not sign reports (because someone else has done it), and vice versa where a disengaged mentor shows up once a quarter to sign the report.

IMO The reports should be signed off by all mentors where possible not just one or two, that way the mentor only signing reports (and doing nothing else) is less of an issue.

Post by Julian Hyde
It may be a bit more difficult to measure, but I think a better indicator would be the number of messages sent by the mentor to dev and private lists. It indicates both time spent writing the message and also time spent keeping up with the email traffic to know when the podling needs help.

Given we have 100+ people to check and some of them are involved in multiple projects, that’s a lot of lists to check but it would be possible. I could put it up in a shared doc and if other people help out then it wouldn’t be too much effort.

Alternately I could take 10 mentors at random if any of them are active on their mailing lists, I’m guessing there a strong correlation between not signing off reports and absent mentors.

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

Post by Justin Mclean
Alternately I could take 10 mentors at random if any of them are active on their mailing lists, I’m guessing there a strong correlation between not signing off reports and absent mentors.

A data point. Not mentioning any names but I checked a single mentor who had not recently signed-off reports in 10 podlings that had a low sign off rate, of those mentors 7 had no email to the lists in the last 6 months, 2 had sent a single email and one is active to a limited extent.

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Julian Hyde]

I hear what you’re saying. As a metric it’s better than nothing, but let’s not assume it is telling us everything.

I believe that asking all mentors to sign off is a change in policy or at least in practice. (Neither the “Mentors’ guide”[1] nor the “Guide to being a mentor”[2] talks about signing off.) This could be a useful opportunity to remind mentors of what is expected of them.

One gray area has always bugged me. How hard should mentors push a podling to submit a report? If the podling fails to submit a report, or submits it late, it is impossible for mentors to sign off, and therefore counts as a black mark against the mentors. I believe that mentors should facilitate and guide, but pushing and nagging is counterproductive.

Julian

[1] https://incubator.apache.org/guides/mentor.html <https://incubator.apache.org/guides/mentor.html>

[2] https://community.apache.org/guide-to-being-a-mentor.html <https://community.apache.org/guide-to-being-a-mentor.html>

Hi,

Post by Julian Hyde
There is a strong presumption here that a mentor is fulfilling their duty if and only if they sign the reports.

It’s an imperfect metric but one we have easy access to. If we send out an email asking for if you’re AWOL and a mentor responds that they are active then all is good. Of course it’s not going to catch mentors who sign reports and do nothing else but I hope there’s only a few incidences of that.

Post by Julian Hyde
I’m skeptical. I can think of cases where an engaged mentor does not sign reports (because someone else has done it), and vice versa where a disengaged mentor shows up once a quarter to sign the report.

IMO The reports should be signed off by all mentors where possible not just one or two, that way the mentor only signing reports (and doing nothing else) is less of an issue.

Post by Julian Hyde
It may be a bit more difficult to measure, but I think a better indicator would be the number of messages sent by the mentor to dev and private lists. It indicates both time spent writing the message and also time spent keeping up with the email traffic to know when the podling needs help.

Given we have 100+ people to check and some of them are involved in multiple projects, that’s a lot of lists to check but it would be possible. I could put it up in a shared doc and if other people help out then it wouldn’t be too much effort.
Alternately I could take 10 mentors at random if any of them are active on their mailing lists, I’m guessing there a strong correlation between not signing off reports and absent mentors.
Thanks,
Justin
---------------------------------------------------------------------

[Justin Mclean]

Hi,

I believe that asking all mentors to sign off is a change in policy or at least in practice. (Neither the “Mentors’ guide”[1] nor the “Guide to being a mentor”[2] talks about signing off.)

See [1] I’d say "and report its status” covers signing off reports.

Thanks,
Justin

1. https://incubator.apache.org/policy/roles_and_responsibilities.html#mentor
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

Post by Julian Hyde
[2] https://community.apache.org/guide-to-being-a-mentor.html <https://community.apache.org/guide-to-being-a-mentor.html>

BTW this link refer to another mentorship program, not being a mentor of a podling.

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

I made a list of possible inactive mentors here based on how often they sign off board reports [1]. It may contain mistakes so apologies in advance if your name is listed and you are actually active in the podling listed. There are 78 names listed of those 50 have not checked off any board reports recently, the other 18 may be reasonably or occasionally active but I don’t know. Hopefully this will show how poor a metric of signing off reports is :-)

Note that some mentors may be active in one podling but not in others and where that is the case I’ve just listed the possible inactive podling(s).There's some well know names there and I think some people may just be a little over worked / have taken on a little too much. That may not be healthy long term for them or their podlings. This are 19 people I’ve listed as possibly having too many podlings.

If IPMC members could double check and filled in if they (or others) are actually active or not that would be a great help. Also feel free to add comments or any other information that you think is useful.

Thanks,
Justin

1. https://docs.google.com/spreadsheets/d/1oR2kymzDTLOy-P-QLYM1wlHFES6LG9X-V2SMgwohvR4/edit?usp=sharing
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

For those interested here are the numbers of possible absent mentors by podling, again this may contain mistakes. If you listed here and your mentors are active again you have my apologies in advance, but if your mentors are absent please speak up. That goes for if you are a mentor on a project and could do with a bit more help.

No Podling
5 gearpump
4 nemo
3 weex
3 toree
3 spot
3 slider
3 singa
3 sdap
3 rya
3 omid
3 netbeans
3 mxnet
3 joshua
3 hivemall
3 heron
3 annotator
2 tephra
2 taverna
2 superset
2 skywalking
2 senssoft
2 samoa
2 s2graph
2 openwisk
2 odftoolkit
2 milagro
2 livy
2 htrace
2 griffin
2 echarts
2 datafu
2 batchee
2 ariatosca
1 wave
1 unomi
1 tamaya
1 soma
1 servicecomb
1 ratius
1 ratis
1 quickstep
1 qucikcstep
1 pulsar
1 ponymail
1 plc4x
1 pagespeed
1 myriad
1 livyedgent
1 imparla
1 impala
1 h2trace
1 gossip
1 gobklin
1 freemarker
1 edgent
1 dubbo
1 druid
1 daffodil
1 crail
1 amaterasu
1 airflow

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

So I went through the tedious task of checking activity on the mailing list in the last 6 months for all possible absent mentors (being those who have missed a number of reports). I may of made some mistakes but it look like that missing 3 reports n a row correlates quite well with mentors being absent. (about 80% or so).

While this doesn’t tell who is just signing off report and otherwise being absent but I think that a good start.

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Julian Hyde]

Thanks for doing this, Justin. It must have taken quite a while.
You've convinced me that missing sign-off is a reasonable proxy for
mentor disengagement.

Just curious: As you were reviewing lists looking for mentor activity,
did you learn anything about patterns of mentor behavior? Are there
any lessons we can learn?

Julian

Post by Justin Mclean
Hi,
So I went through the tedious task of checking activity on the mailing list in the last 6 months for all possible absent mentors (being those who have missed a number of reports). I may of made some mistakes but it look like that missing 3 reports n a row correlates quite well with mentors being absent. (about 80% or so).
While this doesn’t tell who is just signing off report and otherwise being absent but I think that a good start.
Thanks,
Justin
---------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

Post by Julian Hyde
Thanks for doing this, Justin. It must have taken quite a while.

Not as long as you may think due to pony mail's ability to search multiple lists at once. However it can be in some cases hard to match names and not every one uses their email address recorded against their name.

Post by Julian Hyde
Just curious: As you were reviewing lists looking for mentor activity,
did you learn anything about patterns of mentor behavior? Are there
any lessons we can learn?

I was only looking at people we thought might have low levels of activity so that may skew things a little.

I noticed a couple of minor things:
- Some mentors are only active in a podlings starts up phase (but that's probably still a useful thing)
- There a number of mentors that look to have taken on too many podling and are spread a bit thinly
- Some mentors only seem to turn up to vote +1 on things with little or no other activity
- A couple of mentors are quite active but don’t sign off reports

I also noticed that recently discussions around this here have seen to woken up a few mentors. :-)

Thanks,
Justin


---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Dave Fisher]

Hi Justin,

Thanks!

That’s a pretty good correlation. If the process is not too tedious then maybe the incubator should trigger a look into the situation and finding replacement mentors.

Maybe a 7 day ACK to the Mentor. “It looks like you are no longer able to mentor <podling>. We understand that you are a volunteer and circumstances may be fluid. Let us know immediately if this is not so and if you would like to continue as a mentor to <podling>. Otherwise we will seek a replacement from the IPMC and Apache Membership”

Regards,
Dave


Regards,
Dave

Sent from my iPhone

Post by Justin Mclean
Hi,
So I went through the tedious task of checking activity on the mailing list in the last 6 months for all possible absent mentors (being those who have missed a number of reports). I may of made some mistakes but it look like that missing 3 reports n a row correlates quite well with mentors being absent. (about 80% or so).
While this doesn’t tell who is just signing off report and otherwise being absent but I think that a good start.
Thanks,
Justin
---------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org

[Justin Mclean]

Hi,

Post by Dave Fisher
That’s a pretty good correlation. If the process is not too tedious then maybe the incubator should trigger a look into the situation and finding replacement mentors.

That’s the idea :-)

Post by Dave Fisher
Maybe a 7 day ACK to the Mentor. “It looks like you are no longer able to mentor <podling>. We understand that you are a volunteer and circumstances may be fluid. Let us know immediately if this is not so and if you would like to continue as a mentor to <podling>. Otherwise we will seek a replacement from the IPMC and Apache Membership”

Thanks for that. I’ve already suggested some text to send mentors earlier in this thread, we can add/adapt the above to add to it, when we get that that point I’ll start a new thread as we’ve probably lost some people in this one due to it’s length.

Thanks,
Justin
---------------------------------------------------------------------
To unsubscribe, e-mail: general-***@incubator.apache.org
For additional commands, e-mail: general-***@incubator.apache.org