couchdb-marketing mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Samuel Newson <rnew...@apache.org>
Subject Re: The future of couchapps
Date Mon, 21 Sep 2015 15:25:07 GMT
Johs,

I think you’ve taken my reply out of context, but I can see that I’ve forced you to read
between the lines because I didn’t address the topic directly.

I also think the discussion of couchapps does not belong on the marketing@ mailing list. I
feel that marketing@ should be exclusively about marketing topics, not development (current
or future).

It is a fact that no development has been done on couchapps in years, it is also a fact that
couchapps are limited in their abilities. I share Jason’s concerns that it’s a bad experience
to achieve something with couchapps as they exist today only to run into a limit, with no
workaround. 

I define the couchapp support in CouchDB as these endpoints (and only these); _rewrite, _update,
_show and _list. These are each user-programmable server-side endpoints, stored within a design
document.

I see no reason to drop those features. All I believe the PMC is saying is that couchapps
will not become more powerful than these endpoints unless some individual or group steps up
to do it; the core development team are explicitly saying that they are not working on it.

All these endpoints will be in CouchDB 2.0.

Now, to Cloudant. Cloudant is typically used in a database-as-a-service mode, where customers
have a front end and an application middle tier; ergo they do not use the four couchapp endpoints
much (_rewrite for its part in virtual hosting notwithstanding) but we have customers, at
various scales, using couchapp functionality and we plan to continue supporting those users.

In the event that _rewrite, _update, _show and _list become an optional add-on, Cloudant will
deploy that add-on to all our clusters.

Regards,
Robert Newson
CouchDB PMC and IBM Cloudant Senior Database Engineer


> On 21 Sep 2015, at 03:30, Johs Ensby <johs@b2w.com> wrote:
> 
> Thanks Robert,
> for what I regard as a good closing statement for this thread.
> 
> In my environment IBMs support has a very good brand rub-off effect and the support for
couchapps with hosting providers is what I think is important to recruit new developers to
Couch.
> I will pick up the dialogue directly with Cloudant as to what we can expect specifically
on couchapp support going forward.
> 
> Johs
> - looking forward to an Erlang programmer eager to pursue Couch app server functionality
respond to the welcome extended to her or him. 
> 
>> On 20. sep. 2015, at 15.28, Robert Newson <rnewson@apache.org> wrote:
>> 
>> On Cloudant versus couchdb the plan is still code convergence. 2.0 is very close
imo and considerable effort has occurred within Cloudant to prepare to switch to it in our
production environment. 
>> 
>> As for what is staying Cloudant only, it's things like geospatial, the multi-tenancy
and the automation, monitoring and alerting that are needed to deliver a service. The database
and its ability to replicate and form fault tolerant clusters is open source and remaining
so. Cloudant will work directly on the asf repositories that now exist for future enhancements.

>> 
>> B. 
>> 
>>> On 20 Sep 2015, at 09:49, Johs Ensby <johs@b2w.com> wrote:
>>> 
>>> Thanks Joan,
>>> for wrapping this up in a 100% PMC support for Jan's "watch me" on this.
>>> It is good to know, so we who could contribute, but not as erlang programmers,
hold our horses.
>>> I understand that we can be a big nuisance as long as there is no one to do the
development work on the team. 
>>> 
>>> Just to sum my take-away:
>>> The PMC will wholeheartedly welcome the continuation of work for what we without
a good breakdown of what it means call "couchapps" provided there is a "hero" to do the heavy
lifting in terms of erlang programming
>>> We should not count on even the present functionality for "Couchapps" being present
in future releases
>>> 
>>> To #1 I attribute these statements:
>>>> Jan Lehnardt 18. sep. 2015
>>>> In my last email I specifically suggested for such a person to step up and
I’d welcome them wholeheartedly.
>>> 
>>>>> Alexander Shorin  September 19, 2015
>>> 
>>>>> this place is vacant and we are welcome everyone who would
>>>>> like to take this duty on his/her shoulders.
>>> 
>>> 
>>> and to #2 I look at:
>>> Jan's "watch me" response to Giovanni's 
>>>>> - you can't remove something that ALREADY exists which A LOT of
>>>>> people like, use and have businesses on top, just because you and a few
>>>>> people
>>> 
>>> and
>>> Alex'
>>>>> Otherwise CouchApps will continue to be feature of secondary type, sort
of accidental side
>>>>> effect from technical decisions of good past days. No need to say
>>>>> what such state means. They may simply not survive next major release.
>>> 
>>> 
>>> Alex' more consequential prediction i presume is the meaning behind Jan's "watch
me", since this still is an open source community, and I am perfectly OK with there being
some emotions in this conversation. It is not about the nuances anymore, it is about  100%
of the PMC backing this.
>>> 
>>> For me it means leaving the community for now.
>>> It doesn't make sense to me to spend time here until an Erlang programmer has
chosen to respond to the wholehearted welcome that has been extended.
>>> I see CouchDB as part of a bigger ecosystem and will try to contribute at some
other node.
>>> 
>>> My best wishes for success and progress with hopes for extended openness and
compatibility with the rest of the Couch ecosystem
>>> 
>>> Johs
>>> 
>>> 
>>> 
>>> 
>>> 
>>>> On 20. sep. 2015, at 00.12, Joan Touzet <wohali@apache.org> wrote:
>>>> 
>>>> I'll concur with Alex here. My survey of the PMC members so far has
>>>> resulted in 100% of us in agreement with Jan's comments.
>>>> 
>>>> This is not a power play, it is a logical, reasoned decision with many
>>>> factors that play a part -- not the least of which is that there are so
>>>> many great web frameworks out there that do such a better job than we do
>>>> with CouchApps. We'd need a massive effort to catch up with them, and a
>>>> significant number of developers just to keep pace. Right now, there's
>>>> no one working on CouchApps period, as has been repeatedly stated on the
>>>> list.
>>>> 
>>>> If you are willing to put in the time to develop the Erlang, JavaScript
>>>> and possibly C to build the new functionality, and are able to sustain
>>>> development on that over the course of the years it would take to match
>>>> e.g. Python or Node.JS's community level for web apps...well, let's just
>>>> say that I look forward to your GitHub pull requests. :)
>>>> 
>>>> Best regards,
>>>> Joan
>>>> 
>>>> 
>>>> 
>>>> ----- Original Message -----
>>>>> From: "Alexander Shorin" <kxepal@gmail.com>
>>>>> To: marketing@couchdb.apache.org
>>>>> Sent: Saturday, September 19, 2015 1:45:52 PM
>>>>> Subject: Re: The future of couchapps
>>>>> 
>>>>> Hi John,
>>>>> 
>>>>> The future of CouchApps is trivial and was already discussed on this
>>>>> ML[1]: this feature stalled over 4 years already, it has quite enough
>>>>> issues and limitations, it lacks maintainer. If something of these
>>>>> won't change in nearest future, CouchApps are doomed. And all
>>>>> solutions lies though the hero(es) who will maintain and develop this
>>>>> feature.
>>>>> 
>>>>> The question here is not only about to keep existed code works. It's
>>>>> also about vision of what CouchApps are, what they have to be, how to
>>>>> evolve them and make really great and useful feature for CouchDB
>>>>> users
>>>>> and not only.
>>>>> 
>>>>> Currently, there is no active leader of CouchApp feature in CouchDB
>>>>> team, so this place is vacant and we are welcome everyone who would
>>>>> like to take this duty on his/her shoulders.  Otherwise CouchApps
>>>>> will
>>>>> continue to be feature of secondary type, sort of accidental side
>>>>> effect from technical decisions of good past days. No need to say
>>>>> what
>>>>> such state means. They may simply not survive next major release.
>>>>> 
>>>>> [1]: Follow this thread:
>>>>> http://mail-archives.apache.org/mod_mbox/couchdb-marketing/201505.mbox/%3CCAPMhwa5R6CdeQP-m_FHfqYsJ4NDwLYiePuwrsi2bRfzifncRpw%40mail.gmail.com%3E
>>>>> - it's long, but has all the same answers.
>>>>> 
>>>>> --
>>>>> ,,,^..^,,,
>>>>> 
>>>>> 
>>>>>> On Sat, Sep 19, 2015 at 8:16 PM, Johs Ensby <johs@b2w.com>
wrote:
>>>>>> Hi all -- both couchapp lovers and skeptics,
>>>>>> 
>>>>>> I want to start a new thread based on the below conversation
>>>>>> between Jan and Giovanni at the end of the thread [VOTE]
>>>>>> WHY/HOW/WHAT.
>>>>>> 
>>>>>> I am in the group of stakeholders that have a business that rely
on
>>>>>> CouchDB. And I use couchApps hosted on Cloudant. I would like to
>>>>>> see the community grow, not split again and there seems to be a
>>>>>> tone developing that I am sorry if I have been a contributor to.
>>>>>> Rather than having wide-reaching discussion about dream features
>>>>>> of future couchapps, I will try to stay on the subject of a future
>>>>>> for couchapps or not.
>>>>>> 
>>>>>> The "watch me" from Jan below is very disturbing to me.
>>>>>> I would appreciate seeing some comments from PMC members on this.
>>>>>> 
>>>>>> Johs
>>>>>> 
>>>>>> 
>>>>>>>> On 18. sep. 2015, at 13.52, Jan Lehnardt <jan@apache.org>
wrote:
>>>>>>>> 
>>>>>>>> - you don't know if tomorrow some erlang skilled couchapp
lover
>>>>>>>> won't get
>>>>>>>> in and start working on the app part (to say the truth some
of
>>>>>>>> them were
>>>>>>>> already active in the past, but they chose to left just because
>>>>>>>> of
>>>>>>>> this/yours direction)
>>>>>>> 
>>>>>>> In my last email I specifically suggested for such a person to
>>>>>>> step up and I’d welcome them wholeheartedly.
>>>>>>> 
>>>>>>> 
>>>>>>>> - you can't remove something that ALREADY exists which A
LOT of
>>>>>>>> people
>>>>>>>> like, use and have businesses on top, just because you and
a few
>>>>>>>> people
>>>>>>>> don't like it
>>>>>>> 
>>>>>>> Watch me.
>>> 
> 


Mime
View raw message