Let's use the new datafeeds model and the node groups / query processing
changes as use cases, perhaps?
:-)
On 3/18/16 9:42 AM, Till Westmann wrote:
I’m happy to read and write good documentation. However, I think that
a) it’s quite unclear what we are talking about (is this about javadoc,
design sketches, detailed specifications, …) and
b) I’m not happy with "enforcing" things. I think that to get to good
documentation, we should not ask "other" people to do it, but rather
provide a good example to convince everybody to follow.
So I think that the first step is to provide or point to some
documentation that is good or helpful and to discuss on that example
how we can produce more of it (and where documentation is most urgently
needed).
My 2c,
Till
On 17 Mar 2016, at 13:46, abdullah alamoudi wrote:
Totally agree and this is very much needed and will be worth the
investment.
Regards,
Abdullah.
On Thu, Mar 17, 2016 at 11:39 PM, Chen Li <[email protected]> wrote:
We need to do a good job maintaining documentations of the project. To
achieve the goal, we need to decide a protocol and enforce it.
Chen
On Wed, Mar 16, 2016 at 8:37 AM, Till Westmann <[email protected]>
wrote:
In the spirit of not hijacking the thread I’m trying to create a new
thread here.
On your question:
I think that getting up to speed is not quick and it is difficult.
Even
for those of us with the 10+ years industrial experience the learning
curve
is steep. As you said, the code-base is big and big pieces of it
are not
trivial. There are a number of papers on the high-level concepts,
but the
distance between those and the code is also not small. And the
learning
curve is also steep for people at the universities.
We have been thinking about ways to ease the access to the
code-base e.g.
by adding a few references and design docs to the wiki [1]. And
while I
think that this is a step ahead it is not a solution.
Also, while I’ve found everybody to be friendly and helpful, nobody is
working full-time on easing the access to the code-base for
newcomers -
this is usually a volunteer activity in addition to the day-job.
I think that the way that people have been able to get to the
code-base
is
to
- start at some piece that they think could use some improvement
and then
to
- find our how it works and
- how it can be improved
and then move forward from there.
I’m sorry that I cannot suggest a better way here.
But maybe somebody else on the list has a better idea or can tell
you how
they managed to overcome the initial hurdle?
Cheers,
Till
[1] https://cwiki.apache.org/confluence/display/ASTERIXDB/Home
On 16 Mar 2016, at 7:42, Sandeep Joshi wrote:
Mike,
I asked because I wanted to ascertain how quickly a newcomer
outside the
UCI circle can come up to speed with the vast asterixdb code-base.
I don't want to hijack this thread so maybe we can start another
thread.
I am interested in knowing how one could gain proficiency in
participating
in asterixdb development. What design patterns should one keep in
mind,
what hasn't worked historically, what sequence of tasks can one
take on
to
learn about various parts of the system, etc.
-Sandeep
On Wed, Mar 16, 2016 at 7:36 PM, Mike Carey <[email protected]>
wrote:
And I would guess there are at least 5 individuals with 10+ years of
industrial experience (apiece) on the current team, if that's one of
the
reasons for the inquiry, with a handful more probably in the 2-5-ish
years
range. :-)
On 3/15/16 9:32 PM, Till Westmann wrote:
Hi Sandeep,
On 15 Mar 2016, at 19:13, Sandeep Joshi wrote:
I am curious to know how many committers are from outside the
academic
circle (i.e. those who have not worked in the past on Asterixdb
at UC
Irvine or other academic institutions ) ?
I think that the current number committers without academic UC
heritage
is 3 (it’s an educated guess as I don’t know the CVs of the
committers).
Actually, a nice number for a project that was a university
project 12
months ago.
Cheers,
Till
