Re: AD Time

[Padma Pillay-Esnault <padma.ietf@xxxxxxxxx>]

Ted

I did not mean those options as cure all but as steps to help. Fully agree that this will reduce the breadth of view just as areas did.

To make any kind of progress something will need to give though...

Many suggestions has already floated on the list...

Is it better to keep the breadth or not? 

How can we keep breadth and lighten load? 

Is delegation an option?

We can poll ADs (current and ex) to give 1-3 suggestions that would be helpful for them? 

We can then look at how the process can be enhanced or not for a trial period and see if it actually helps at all. 

My 2 cts 

Padma 

Sent from my iPhone

On Jul 30, 2018, at 13:02, Ted Lemon <mellon@xxxxxxxxx> wrote:

Padma, all of these solutions reduce the breadth of view of each AD, which in turn limits the opportunity for there to be someone in whose head a lot of siloed threads of work are known.   This is why I am arguing that this isn't the right approach to the problem.   When I was AD (and I realize that not everybody considered this a feature) I really felt like if I just relied on directorate reviews without reading the documents myself, I wasn't going to have that overview, and that cross-pollination of ideas would suffer as a result.

I would not say that this is an impossible problem to solve, but what you've described will not solve it.   In order to solve it, we'd need for the directorate reviews to concisely summarize what each document does, with enough detail that the AD who's the recipient of the review has roughly the same mental picture they would have had if they'd read the document in detail; just with less detail.   I think this would be great, but unfortunately this relies on the directorate reviews being of generally _much_ higher quality than they are now.   And I say this having been the recipient of some very good directorate reviews recently: these reviews would not have served the purpose I'm describing.

So if the right answer to this is to make AD a 10% position, we have to figure out how to significantly increase the quality of even already-good directorate reviews.   I'm not saying that's impossible, but we'd have to figure out how it can be possible.

On Mon, Jul 30, 2018 at 3:50 PM, Padma Pillay-Esnault <padma.ietf@xxxxxxxxxx> wrote:

On Mon, Jul 30, 2018 at 11:48 AM, Ted Lemon <mellon@xxxxxxxxx> wrote:

To be clear, I'm suggesting that the IETF pay a stipend so that the AD can pay some or all of their living expenses.   I propose that this would be set at some reasonable but not silicon-valley level, so if you work in Silicon Valley you probably want to get your employer to pay you to do it.   I would think that the right number would be something 33% or 50% of a reasonable silicon valley salary—around USD $100k.   This would be too low for someone living in Silicon Valley to do as a full-time job, unless they live with roommates or take on additional consulting work, for which clear disclosure guidelines would have to exist.   The IETF would also pay travel expenses to meetings and retreats.   And by "the IETF" I am hand-waving a bit, since obviously the IETF doesn't have the budget to do this.

I don't see any point in the IETF paying just travel expenses: either you have a sponsor or you don't.   I think it's really quite unreasonable for the IETF to expect someone to be an AD for free—it's more work than makes sense for anybody who isn't independently wealthy to do as a volunteer.

<Padma2>

To Warren's point, once we open this pandora box, a lot of companies may reduce or even eliminate their sponsorship.

Also, we need to be careful that a new set of biases are not introduced.

The problem seems to be the amount of time ADs spend, there are a number of ways to solve this

1. Reduce number of docs

2. Increase number of ADs

3. Create a supporting structure to reduce workload by delegating

 <Padma2> 

On Mon, Jul 30, 2018 at 2:36 PM, Padma Pillay-Esnault <padma.ietf@xxxxxxxxx> wrote:

From the numerous discussions here,  "paid by the IETF" seems to be interpreted differently.

1. If "paid by the IETF" means compensation as in salary then I foresee a number of difficulties in implementing it

The salary disparity according to regions/countries for one is tricky. Who will decide who much depending on location?

Along with the salary comes the issues of benefits ( in some countries there are health coverage to be integrated for example and others not ...)

There is also the different labor laws for each different country: employer paid benefits, retirement, taxes....

Will IETF hold this payroll structure for potentially 1 employee per region/country for only 2 years? 

2. If "paid by the IETF" means reimbursement of travel expenses then this may be more doable IMHO.

Circling back to the issue of diversity of candidates, then I do not see how 2 will increase diversity. It may still favor for example those who just do not want to incur costs but have other sources of income.

"Diversity" is also open for interpretation ... Gender/age/??

In this context, IMHO, it should include having a balance between "types" of companies/org (operators, edge, network vendors, academia, consulting ... ) not just only "names" of companies. 

my 2 cts

Padma

On Mon, Jul 30, 2018 at 10:34 AM, Ted Lemon <mellon@xxxxxxxxx> wrote:

I guess it sounds like you're proposing a much larger change than I'm proposing.   Right now, you serve the community and Google pays you to serve the community.  I'm suggesting that an AD could be paid by the IETF to serve the community.   I'm not suggesting that an AD would be paid by the IETF to serve the IETF.   Actually, I don't even know what it means to serve the IETF as a separate thing from the community.   The IETF is a weird organization.    But your job description would be to serve the community, not to serve the organization.   We already pay people to serve the organization.   Who do you think would be doing all of those metrics you speak of?   You?   Why would you be doing that?   That wouldn't be serving the community.   Would the IETF chair be tracking those stats?

Organizational leaders are not tracked on the basis of how many widgets they produce.   They are judged on the basis of whether they help the organization to succeed or fail.   This can be a problem when "succeed" and "fail" are defined badly, as is the case with companies that think their job is to serve the immediate needs of the stockholders, but we aren't talking about that here, and I think we all understand these problems pretty well.   So it surprises me how many different ways various people who have joined in in this conversation believe it would be done wrong, when if it is done at all, it would be we who are having this conversation who decide what "done right" and "done wrong" look like.

On Mon, Jul 30, 2018 at 1:15 PM, Warren Kumari <warren@xxxxxxxxxx> wrote:

On Mon, Jul 30, 2018 at 10:35 AM Ted Lemon <mellon@xxxxxxxxx> wrote:
>
> Warren, why would you feel an implicit obligation to get documents out fast rather than good?   That doesn't make sense to me.   The whole point of having ADs is that they get documents out good, not that they get documents out fast.   If we valued getting documents out fast over getting them out good, we would have no ADs.   The recent comments on the session signaling doc are a great example of this: several very smart people took a significant chunk of time to read these documents carefully, and had lots of thoughtful comments.   The document had already been through several layers of review.   What we got from the IESG on this document is exactly what I want out of the IESG: a view from someone smart and careful who otherwise wouldn't have read the document, and who has some ability to make my life less smooth if I try to take the easy way out and not make the requested changes.   That's not the only aspect of the AD job, but in my mind it's one of the most important aspects of it.

Thank you, we try. It's always nice when people appreciate our feedback..

> So the idea that a paid AD would suddenly abdicate that responsibility is the exact opposite of what I'd personally want, and I don't understand why that seems like a natural conclusion to you.   Not saying you're wrong—just not following your logic.

And it's the opposite of what I would want as well -- but I suspect
that I would feel a need to do what is best for the "organization",
which is not the same thing as the "community" - I could see the
organization wanting metrics to evaluate employees, which I suspect
will lead to measuring number of documents progressed (good), amount
of time you held up documents with discusses (bad), number of nits
pointed out (good), WGs chartered (good), etc. Also, I currently feel
that I have the freedom to speak out when I think that the
organization is doing something dumb / not in the interest of the
community - if I'm paid to serve, I'd feel a sense of loyalty to the
IETF, not the community.
I suspect I'm not really articulating this very well...

>
> As for "Google wouldn't let me do the job," do you think that's really true?

I'm not sure, but I suspect it is possible.
If I went to Google and told them that I'd like to spend a significant
amount of my time working for Verisign (NASDAQ: VRSN) helping run
a..root-servers.net they would presumably wonder why Verisign doesn't
just hire someone. Google kindly lets me spend some of my time helping
USC ISI run b.root-servers.net (and before that, helping ISC run f);
some of this is because of the type of organizations these are. I also
use some of my personal time to help run various community (and
similar) networks specifically because they are community / volunteer
type roles - if they hired people to do the same work I'd be much less
inclined to spend my time doing this.
I'll happily spend hours helping sort wood at my local makerspace, but
wouldn't do the same thing for Home Depot.

Somewhat related to this is a *fascinating* NPR podcast -
https://www.npr.org/sections/money/2013/08/02/187373801/episode-386-the-cost-of-free-doughnuts
- well worth a listen.

> Is Google really that dumb?

..... and now you are trolling / that is a loaded question (see, if I

worked for the IETF I would implicitly be working for you, and
wouldn't feel comfortable saying that :-) )

Google isn't dumb, but they *do* pay me well[0] - if the IETF was able
and willing to hire for the AD role, why wouldn't Google prefer that
they do that, and instead put me to work doing something more directly
related? Having organizations each contribute their employees' time as
volunteers creates a feeling that the organizations are all
contributing to the good of the Internet. If the IETF were hiring
people for this role, I think that feeling would change and it seems
likely that employers would rather spend their employees' time
elsewhere, either internally or for organizations who cannot hire
their own.

These are my views, they may be completely wrong...

W
[0]: Hmmm... not sure who's point I'm making here :-)

W

>
> On Mon, Jul 30, 2018 at 10:20 AM, Warren Kumari <warren@xxxxxxxxxx> wrote:
>>
>> On Sun, Jul 29, 2018 at 7:35 PM Ted Lemon <mellon@xxxxxxxxx> wrote:
>> >
>> > A little less catastrophization might make this conversation more fun!  :)
>> >
>> > Seriously, each of the questions you're asking implies a fairly obvious answer.   For example, fundraising: this is straightforward: always have enough money to pay all the ADs for the next year or two.   Figure out how to raise that money.   If it's not available, then this option isn't open to us: end of story.  Once that endowment exists, keep funding it.   If the funding dries up, oh well, we tried.   The only way to find out if this is possible is to try it; the only reason to try it is that we think it's worth trying.   I think this conversation is about whether we think it's worth trying (the running consensus appears to be "no," but we haven't heard much from people who would have tried for IESG if this option were available).
>>
>> Something else to keep in mind is that having the *option* for paid
>> ADs changes the tone of the role, and may make some people unable (or
>> unwilling) to serve.
>>
>> My management is willing to let me serve as an AD because is is a
>> volunteer position (and because I made it clear that I really wanted
>> to serve) - if there was the option for the IETF to "hire" people for
>> the role, it is entirely possible that they would not have let me do
>> so ("Eh, we pay Warren lots of money - if the IETF can hire their own
>> people let them do that, and we'll put Warren to work on "real work"
>> instead"). Also, if I were being paid by the IETF / ISOC / Endowment /
>> Cake Bake Fund I would (personally) feel different about the role --
>> currently I serve because I really like the IETF and want to feel like
>> I'm giving back. If I were being paid (or if others were being paid) I
>> would feel very differently about the organization and it would go
>> from a labor of love to a job. In addition, instead of balloting what
>> I believe, I would feel an implicit obligation to get documents out
>> the door fast (measurable) versus as good as they can be
>> (unmeasurable).
>>
>> Just some thoughts,
>> W
>>
>> >
>> > On Sun, Jul 29, 2018 at 3:29 PM, Theodore Y. Ts'o <tytso@xxxxxxx> wrote:
>> >>
>> >> On Sat, Jul 28, 2018 at 08:59:32PM -0400, Ted Lemon wrote:
>> >> > Ted, it sounds like you're suggesting that right now there's no bias, and
>> >> > if this change were made, it would create bias.   The reality is that if we
>> >> > did exactly the change you suggest, it would indeed shift the bias away
>> >> > from people who can get corporate sponsorship to those who can afford to
>> >> > take bigger risks/work for less money.   Of course, that's not the only way
>> >> > to do it—we could also make it available as an option, while allowing the
>> >> > old form of sponsorship as well.   What's the old quote, "the law, in its
>> >> > infinite grandeur, forbids the rich and poor alike from sleeping under
>> >> > bridges..."
>> >>
>> >> I wasn't referring to the bias that the people might hold, but the
>> >> bias of the sort of people that would stand for selection by Nomcom if
>> >> it required them to resign from their present job and be paid
>> >> non-profit wages by a SDO.
>> >>
>> >> If you are saying that it would be an option (so either their current
>> >> employer could choose to keep them on their payroll, and allow them to
>> >> continue to accrue equity compesantion), *OR* the IETF would somehow
>> >> find the salary for the AD, somehow, then that would avoid decreasin
>> >> the slate of people willing to stand for selection by Nomcom --- but
>> >> that transfers the burden to the organization that needs to be able to
>> >> find the salary for the AD if it turns out to be necessary..  It's hard
>> >> to raise money when it's not clear whether or not it's needed.
>> >> Especially if it turns out if the answer is trying to hold out a tin
>> >> cup and beg for donations (sorry, sponsorships).
>> >>
>> >> Or what other alternative did you have in mind for finding the $$$ to
>> >> pay for a full-time AD's salary?  I hope you're not proposing that the
>> >> IETF start charging hundreds or thousands of dollars for
>> >> fourth-generation xerox copies, ala what was needed to get a hold of a
>> >> (legal) copy of the ASN.1 spec from ANSI....
>> >>
>> >>                                         - Ted
>> >>
>> >>
>> >> >
>> >> > On Sat, Jul 28, 2018 at 8:12 PM, Theodore Y. Ts'o <tytso@xxxxxxx> wrote:
>> >> >
>> >> > > On Sat, Jul 28, 2018 at 06:23:40PM -0400, Ted Lemon wrote:
>> >> > > > ADs don’t choose their terms: nomcom does.
>> >> > >
>> >> > > So this biases the people available to nomcom to those people who are
>> >> > > either (a) consultants, or (b) willing to resign from their well-paid
>> >> > > corporate job to take a job with a non-profit SDO.
>> >> > >
>> >> > > I don't believe this will result increasing the quality of the slate
>> >> > > of candidates available to Nomcom compared to what we have now.  Which
>> >> > > was the whole point of this proposal, was it not?
>> >> > >
>> >> > >                                           - Ted
>> >> > >
>> >
>> >
>>
>>
>> --
>> I don't think the execution is relevant when it was obviously a bad
>> idea in the first place.
>> This is like putting rabid weasels in your pants, and later expressing
>> regret at having chosen those particular rabid weasels and that pair
>> of pants.
>>    ---maf
>
>


--
I don't think the execution is relevant when it was obviously a bad
idea in the first place.
This is like putting rabid weasels in your pants, and later expressing
regret at having chosen those particular rabid weasels and that pair
of pants.
   ---maf