[teampractices] [Engineering] Feedback requested on proposal for creation of Agile Specialist Group

[Oliver Keyes]

On 13 March 2014 06:56, Dan Andreescu <dandreescu at wikimedia.org> wrote:

>  On Wed, Mar 12, 2014 at 4:42 PM, Oliver Keyes <okeyes at wikimedia.org>wrote:
>>>>
>>>> The objection is pretty self-explanatory; agile is a philosophy that
>>>> dictates putting things we strongly suspect, or even know, to be actively
>>>> buggy, in front of users. When doing so includes replacing or superseding
>>>> core functionality and either forcing or strongly suggesting that users
>>>> should use the buggy replacement, users get, ah, pissed. Users like things
>>>> that work, and when you replace something that works with something that
>>>> doesn't while insisting it'll totally be more usable at some undefined
>>>> point in the future we can't pin down because we don't actually know in
>>>> detail what we'll be doing more than 2 weeks in advance, they start to
>>>> wonder very loudly at our competence.
>>>>
>>>>
>>>
>>> I understand your description of the objection and that some users are
>>> intolerant of the degree to which we break things. However, I think this is
>>> a mischaracterization of agile philosophy, or at least is an interpretation
>>> of the agile manifesto [1]/agile principles [2] that I don't agree with. It
>>> is still possible to create thoroughly tested/QA'd software with minimal
>>> bugs and do it while embracing an agile mindset. I think that we as an
>>> engineering organization place a higher priority on getting experiments and
>>> features in front of our users than we do on polish, but that is not
>>> because to do so is necessarily agile.
>>>
>>
> I think, Arthur, you take too soft of a line here.  When Oliver describes
> agile as  "a philosophy that dictates putting things we strongly suspect,
> or even know, to be actively buggy, in front of users", he's just plain
> wrong.  It's not a potential interpretation of any agile principles that
> I'm aware of.  I mean, sure, people could interpret capitalism as "kill
> everyone until you are the lone survivor - Highlander, there can be only
> one!".  But people ... don't.  And I think it's equally ridiculous to say
> that agile dictates putting bugs in front of users.  I think, Oliver, it's
> likely you had some bad experiences with agile *attempts*, including at
> WMF.  I think this just underscores the importance of training.  So I think
> it would be constructive if, from this point, we left prejudice about agile
> at the door and focused on evaluating the merits of this proposal on the
> values that the agile process tries to foster.  Arthur, perhaps we could
> start a concise list of this - testing, breaking up large projects into
> small chunks, iterative improvement not just of the product but of the
> process and team dynamics, etc.
>
> Evidently I have. Don't get me wrong, I don't think Agile is necessarily
the problem here - I think that the problem is, frankly, overenthusiastic
managers and a weirdly concentrated hierarchical process that leads to a
situation where, combined with human nature, it's really really easy to
ignore the people going 'uh, this is totally an issue, can we please not
deploy this?' But, yes, I have had some bad agile experiences, and more
importantly, the initial question was "Why do the community not like
agile?" This is why the community doesn't like agile - because all of the
failures in our development process are all too often externally justified
with 'it's Agile!' or 'more eyes make shallow bugs!' or similar meaningless
pontificating. And sometimes that is because of agile, and sometimes that's
because we don't take the time to go "okay, is this agile itself, or our
agile-justified process?", but either way it's externally justified the
same way.


> Sure. But if you're an average user, you don't see the development
>> philosophy behind what's changing in the site you rely on 9 times out of
>> 10. When you do, it's because you're on a site that prioritises
>> transparency, like ours. IOW, I wouldn't be shocked to find that most users
>> involved even tangentially in our development processes, as consumers
>> rather than devs, assume that this is just How Agile Works, because most
>> dev teams don't expose their processes.
>>
>
> This I agree with.  But I think if we emphasize agile done properly, we
> can positively impact user perception.  Both by showing progress more
> frequently and by increasing the quality of the progress.  That's the point
> of this proposal I think, and if we execute it properly "How Agile Works"
> will change, because perception follows results not prejudice.
>

Sure; I'm not saying agile can't be done well, or that this wouldn't help
with it. I'm answering the original question, which was about the
*perception* of the agile coaching office. The *perception* of agile in the
community, and their likely reaction to the creation of the agile coaching
office or whatever we're calling it in this particular 30 second chunk of
time, is "badly".


-- 
Oliver Keyes
Product Analyst
Wikimedia Foundation
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wikimedia.org/pipermail/teampractices/attachments/20140313/a5a0a8b6/attachment-0001.html>