flink-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Max Michels <...@data-artisans.com>
Subject Re: Drafting a roadmap for Flink
Date Tue, 20 Jan 2015 09:41:01 GMT
Hi Kostas,

Thanks for putting this into the wiki. I added the JIRA link for the
off-heap memory. Now the wiki displays:

> Error rendering macro 'jira' : com.atlassian.confluence.macro.MacroExecutionException:
java.lang.RuntimeException: Not Found

If persistent, we should report this.

Is this the Flink streaming roadmap you were referring to?
https://cwiki.apache.org/confluence/display/FLINK/Flink+Streaming

I agree, there should only be one roadmap. As far as I see, a lot of
points in the streaming roadmap are also contained in the general
roadmap. We should further join them were possible and leave the
remaining parts of the streaming roadmap for future reference.

Best regards,
Max

On Mon, Jan 19, 2015 at 7:57 PM, Kostas Tzoumas <ktzoumas@apache.org> wrote:
> Hi everyone,
>
> I started a wiki page about this:
> https://cwiki.apache.org/confluence/display/FLINK/Flink+Roadmap
>
> If you are working on one of these features, could you insert the
> corresponding JIRA ticket and expand the description if you think it's not
> informative enough?
>
> I saw that there is a streaming roadmap page as well, I think we should
> have only one. The styles are currently a bit different though. Perhaps we
> could isolate some JIRAs/specific features from the projects mentioned in
> the streaming roadmap and insert them in the general roadmap? What do you
> think?
>
> Best,
> Kostas
>
> On Thu, Jan 8, 2015 at 9:16 AM, Kostas Tzoumas <ktzoumas@apache.org> wrote:
>
>> +1 for indicating the person currently working on the issue, we can just
>> open a JIRA issue for each of these. And we can clearly indicate that other
>> features are not being currently worked on.
>>
>> How about indicating rough time goals (quarters) for issues that are
>> currently being worked on (of course with the concern of the assignee)?
>>
>> I have a problem with priorities: the only priorities I see right now are
>> P1 (someone is working on this) and P2 (noone is working on this), and this
>> information is already conveyed by the JIRAs. We can come up with a more
>> detailed priority scheme, but would this be easier to implement than
>> date-to-complete goals?
>>
>> Kostas
>>
>>
>>
>> On Thu, Jan 8, 2015 at 11:39 AM, Fabian Hueske <fhueske@apache.org> wrote:
>>
>>> Adding a responsible person sounds good. We should make sure that this
>>> role
>>> is clearly communicated though.
>>>
>>> How about adding priorities instead of time estimates?
>>> That would help to see how the priorities are set in Flink and which
>>> features to expect next (without having a date assigned to it though).
>>>
>>> 2015-01-08 11:30 GMT+01:00 Stephan Ewen <sewen@apache.org>:
>>>
>>> > Okay, I see what you are going. Some issues have people working on them,
>>> > some do not.
>>> >
>>> > How about we add a "responsible person" to the items that have someone
>>> in
>>> > charge already, and mark others as open?
>>> >
>>> > Associating a responsible person (that need not be the one that does all
>>> > the work, but the one that supervises the issue) may be a good idea in
>>> > general.
>>> >
>>> > On Thu, Jan 8, 2015 at 11:26 AM, Fabian Hueske <fhueske@apache.org>
>>> wrote:
>>> >
>>> > > Having an estimate assigned to an issue might give the impression
>>> that it
>>> > > is already assigned to somebody.
>>> > > This would not help to find external contributors who are interested
>>> in
>>> > > helping with a certain feature.
>>> > >
>>> > > Issues without estimates are still useful as they show in which
>>> direction
>>> > > the project plans to evolve.
>>> > >
>>> > > 2015-01-08 11:13 GMT+01:00 Stephan Ewen <sewen@apache.org>:
>>> > >
>>> > > > The Pig Journal is nice, we could have this evolve into something
>>> like
>>> > > > that.
>>> > > >
>>> > > > I think that we need to give some time estimate on the features
/
>>> > issues.
>>> > > > Otherwise, it is of rather little value - all it says is that
people
>>> > > > thought about that, no one knows when you can plan with it.
>>> > > >
>>> > > > We can coarsen the time estimates, though...
>>> > > >
>>> > > > Stephan
>>> > > >
>>> > > >
>>> > > > On Thu, Jan 8, 2015 at 10:57 AM, Robert Metzger <
>>> rmetzger@apache.org>
>>> > > > wrote:
>>> > > >
>>> > > > > Hi,
>>> > > > >
>>> > > > > I very much like the "PIG Journal" here:
>>> > > > > https://cwiki.apache.org/confluence/display/PIG/Pig+Journal
>>> > > > > Its basically a nice view (however outdated in that case)
on whats
>>> > > going
>>> > > > on
>>> > > > > in the PIG community. You can see finished features on the
top,
>>> > current
>>> > > > > features being developed in the middle and ideas in the end.
>>> > > > >
>>> > > > > The document posted by Stephan is a good start to create
a "Flink
>>> > > > Journal".
>>> > > > > I agree with Fabian that the estimates are very optimistic.
>>> > > Implementing
>>> > > > > all these features including unit tests, documentation and
>>> testing a
>>> > > > takes
>>> > > > > a lot of time.
>>> > > > > I would suggest to only add estimates (finish dates) to features
>>> > which
>>> > > > are
>>> > > > > currently work in progress.
>>> > > > > The remainder ("ideas") can have time estimates in months
but
>>> should
>>> > > not
>>> > > > > have finish dates. Similarly to Pig, we should put a disclaimer
on
>>> > top
>>> > > > that
>>> > > > > we do not guarantee for any feature being developed.
>>> > > > >
>>> > > > >
>>> > > > >
>>> > > > >
>>> > > > > On Thu, Jan 8, 2015 at 9:59 AM, Fabian Hueske <fhueske@apache.org
>>> >
>>> > > > wrote:
>>> > > > >
>>> > > > > > Hi,
>>> > > > > >
>>> > > > > > I think the roadmap should show the long-term development
goals
>>> of
>>> > > > Flink,
>>> > > > > > i.e., show that we are going for a ML library, SQL support,
>>> > > > > Batch-Streaming
>>> > > > > > integration, etc.
>>> > > > > >
>>> > > > > > Right now, it is quite detailed and with very optimistic
time
>>> > > > estimates,
>>> > > > > > IMO.
>>> > > > > > If we would do everything in time, we would be done
with the
>>> > roadmap
>>> > > in
>>> > > > > Q3
>>> > > > > > 2015...
>>> > > > > > I would not even put a time on all issues, esp. on things
which
>>> > > depend
>>> > > > on
>>> > > > > > other developments (which might not even have started).
Also I
>>> > would
>>> > > > make
>>> > > > > > the estimates more coarse-grained. For short-term goals
we could
>>> > use
>>> > > > > > quarters, everything does not need an estimate, IMO.
Issues that
>>> > will
>>> > > > be
>>> > > > > > solved in two months don't even need to be listed.
>>> > > > > >
>>> > > > > >
>>> > > > > > 2015-01-08 7:50 GMT+01:00 Aljoscha Krettek <aljoscha@apache.org
>>> >:
>>> > > > > >
>>> > > > > > > I added some text about my work on the Logical
Query feature.
>>> > > > > > >
>>> > > > > > > On Thu, Jan 8, 2015 at 12:42 AM, Stephan Ewen <
>>> sewen@apache.org>
>>> > > > > wrote:
>>> > > > > > >
>>> > > > > > > > Hi everyone!
>>> > > > > > > >
>>> > > > > > > > It is time we bring the Flink roadmap up to
speed with what
>>> has
>>> > > > > > happened
>>> > > > > > > in
>>> > > > > > > > the last months and what further goals features
ideas have
>>> come
>>> > > up.
>>> > > > > > > >
>>> > > > > > > > The link below leads to a Google Doc that
contains an
>>> initial
>>> > set
>>> > > > of
>>> > > > > > > > suggestions that some of the committers have
come up with.
>>> > Please
>>> > > > > share
>>> > > > > > > > your opinion on those suggestions and feel
free to suggest
>>> > > > additional
>>> > > > > > > items
>>> > > > > > > > to put on the roadmap.
>>> > > > > > > >
>>> > > > > > > >
>>> > > > > > > >
>>> > > > > > >
>>> > > > > >
>>> > > > >
>>> > > >
>>> > >
>>> >
>>> https://docs.google.com/document/d/1QZ0NJC03pOBqE6vbK1Ot4bXwoBcszIqzbZ8a6B5vSEo/edit?usp=sharing
>>> > > > > > > >
>>> > > > > > > > This is specifically open to everyone, not
only committers.
>>> The
>>> > > > link
>>> > > > > > > should
>>> > > > > > > > allow everyone to add suggestions and comments
to the doc
>>> (but
>>> > > not
>>> > > > to
>>> > > > > > > edit
>>> > > > > > > > it directly).
>>> > > > > > > >
>>> > > > > > > > For new suggestions, it would help a lot if
you could also
>>> > > mention
>>> > > > > > > whether
>>> > > > > > > > you would be available to help out with that
feature or
>>> idea -
>>> > > that
>>> > > > > > > helps a
>>> > > > > > > > lot with prioritizing and estimate the time
line.
>>> > > > > > > >
>>> > > > > > > > For general suggestions to the "road mapping"
process,
>>> please
>>> > > > respond
>>> > > > > > to
>>> > > > > > > > this mail.
>>> > > > > > > >
>>> > > > > > > > Greetings and happy drafting!
>>> > > > > > > > Stephan
>>> > > > > > > >
>>> > > > > > >
>>> > > > > >
>>> > > > >
>>> > > >
>>> > >
>>> >
>>>
>>
>>

Mime
View raw message