incubator-ooo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Terry Ellison <Te...@ellisons.org.uk>
Subject I will try to finish my current tasks before I go
Date Fri, 02 Sep 2011 08:56:12 GMT
This is a response to some content in the replies to
   *  Dennis Hamilton's thread "RE: Administrative controls /  
management of the mailing lists" sent Thu, 1 Sep 2011 12:59:44 -0700
   *  and more specifically to my reply on 01 Sep 2011 23:02:11 +0100.  
However I've broken the references chain, since this isn't the subject.

It was late when I wrote this, so maybe I should have phrased it 
better.  However one of my paragraphs generated a strong reaction and a 
bit of personal criticism, So I just want to explain what my intent was 
below, since the Apache way seems to be to let it all hang out in public.

> ... Quite honestly -- and I can only speak personally -- at the moment 
> I feel that I am caught between a rock and a hard place. 

The context  where I said this was one where Dennis was asked who could 
do a piece of work where:

(i) we have a legacy box running in Oracle infrastructure and its state 
is unknown;
(ii) the maintainers who ran the box have now left Oracle;
(iii) Oracle might be willing to grant temporary root access to this box 
for someone who has the necessary approvals;

I don't have the bandwidth to do this, and we need someone to log onto 
the service; work out what is running and how it is configured; identify 
what applications and content needs to be pulled coordinate with both 
Oracle (to make sure the necessary approvals are obtained) and Apache 
infra (to work out how to map this into a service that fits within the 
Apache support models); to stand up the new service; to pull the content 
and orchestrate the cut-over. So:

> My work is time consuming and the skills are different to the 
> mainstream C++ trained OOo developer, but they are also different to a 
> pure sysAdmin.  In some ways you need to be an expert in *both* these 
> worlds and to be able to integrate this expertise.  I am not talking 
> about enthusiastic newbie volunteers; I am talking about hacks who 
> have done this so many times that it's routine.

What I was trying to do here was define the sort of skills that we need 
to meet the request that Dennis made.  If anyone thinks that this is 
wrong and is willing to volunteer, then please reply to Dennis.

> Again this only my personal experience, but I feel that Apache is 
> unwelcoming to newcomers and this seems to be an endemic culture, 
> albeit strongly advocated by a few individuals.  It is intolerant and 
> often outrightly hostile to domains of expertise outside its comfort 
> area -- even though these may be more relevant to the work and 
> Apache's wider mission.  In short I am being asked to work long hours 
> on technically demanding tasks in a dysfunctional environment.  If I 
> was being paid to do what I am doing now, then I would be seriously 
> thinking about changing jobs -- and this is from a guy who spent 32 
> years working with the same company working to get to its top 
> technical tier -- and also one who is now doing this work pro-bono ...

This was a general comment about my feelings and from a person 
perspective, and how I am struggling to fit it.  I was talking about the 
whole Apache ethos as it appears to me, though the infrastructure 
aspects are one part of this.  My point about "32 years ..." is that I 
am a sticker and I don't give up easily.  Even so, I am also questioning 
how I can contribute effectively here, because I understand that it 
seems to be "the Apache way or no way", and I was trying to make this 
work for me.  For example in the wider (non-infrastructure) discussion 
in this thread we were originally talking about the governance of the 
forums.  My fear here is that if we make some of the suggested 
governance changes to make the forums comply with the Apache open ethos 
then most of the major user contributors will simply leave and the 
forums implode as a service.  I accept that we can't change the Apache 
ethos, but I also find it hard to assist in destroying something that 
I've spent the last four years building.

Tony has talked elsewhere to me about building up Karma and a newbie 
earning way into an organisation.  I understand what he says, and this 
would be a sensible way to go under normal circumstances.  Unfortunately 
here we have Oracle who are keen to decommission the two services: the 
forums and the wiki, and we have the a number of us on the Apache side 
pressing to see this work completed in a few weeks.  This also suits me, 
because I normally spend a month on a Greek island in Sept/Oct and I 
don't really want to put this off.  In the case of the wiki, for 
example, this involves a move of a customised MW 1.15.1 with a lot of 
extensions on Solaris/Coolstack to MW 1.17.0 on a LinuxVM / LAMP.  This 
is a complex piece of work, and one of the challenges is that the Oracle 
guy who worked on the wiki full-time had put off this upgrade for two 
years because of the complexity and work involved competing with his 
other tasks.  We have others, such as the mail service that Dennis asked 
about, that will die sometime because of lack of resources.

Picking up one of Gavin's Qs in the above thread "Please explain what 
you mean by dysfunctional", I assumes he mean "in this context" as the 
word itself is widely understood.  I meant that the system as a whole 
doesn't seem to work properly /as far as my involvement and engagement 
goes/.  I will pick two examples:

    * *Establishing project consensus positions*. Organisations such as
      Wikipedia have a very strong consensus building policy model. 
      IMHO, it works really well.  A shared Wiki page is used for
      proponents to articulate alternative propositions.  Anyone is free
      to add one comment to one of the propositions in support or
      against.  That way anyone can get a feeling of the participant
      community's wishes within a single document.  At this point, a
      small number of moderator will sentence the document and reject
      outliers, then integrate the rest into typically 2 or 3 mutually
      exclusive options into a new iteration and the commenting / voting
      cycle is repeated -- and again if necessary.  At the end of this,
      the majority view is picked and endorsed as consensus which is
      then adopted as policy and the debate ends there.  Policy is
      policy.   Here we seem to have long and rambling nested threads
      which eventually peter out.  OK, if you read the thread then you
      can see some pattern of views, but a few days/weeks later we will
      have another related discussion with exactly the same views being
      articulated.  In the meantime the guy who decides to do the work
      hopefully bares this all in mind and gets on with it (lazy
      consensus) -- until some decides to say stop, often referencing
      the Apache way, and eventually we might get a vote.  I can't see
      any working process for getting issues that should be policy
      explicitly agreed as such -- at least on this project.  That's why
      it doesn't work for me.

    * *Infrastructure processes and practices*.  Watching these at work,
      they clearly work for the core team who do most of the work.  No
      question about this.  But nothing is written down, with everything
      just "commonly understood" by the team.  There are many ways of
      organising and executing this type of infrastructure service, and
      the team has chosen one which can work in small tightly knit
      groups.  However it doesn't scale and it will be very foreign to
      newcomers who are used to different working models.  For example
      anyone working in an F500 company or major government organisation
      is used to working within an ITIL or other QA framework.  So it
      could take newbies months to learn their way into the Apache
      model.  This can cause conflict and tensions if you have project
      goals which involve weeks.  I am a newbie here.  I have weeks.  I
      have a corporate background.  I am used to picking up a rule book,
      reading it, understanding it, then following it.  Here, I seem to
      cross invisible lines and get publicly flamed in response.  That's
      why it doesn't work for me.

This is probably my last public comment on the above.  I don't want to 
start a ping-pong responding to the personal criticisms in the previous 
thread.  They are public and on the record so others can form their own 
opinions.  Whatever happens, I will do my best to migrate the forums and 
wiki, to document these systems and to bring them under proper revision 
control to a standard acceptable to the infrastructure team.  But the 
sensible option for all currently seems to be for me to disengage from 
apache.org at that point.

Regards
Terry Ellison

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message