couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jan Lehnardt <...@apache.org>
Subject Re: Whither CouchApps (Was: views failing due to fabric_worker_timeout and OS process timed out)
Date Mon, 27 Feb 2017 11:23:26 GMT

> On 27 Feb 2017, at 11:23, Martin Broerse <info@martinbroerse.com> wrote:
> 
> Hi Jan,
> 
> By `the hosting feature` I mean the rewrite rules and the vhost. Good to
> hear that is not going away.

I said attachments are not going away. rewrites and vhosts are probably
not going to be re-implemented in a new http layer (unless we find
someone who wants to write the code and maintain it, *hint* *hint*)

Best
Jan
--

> 
> - Martin
> 
> 
> On Mon, Feb 27, 2017 at 11:04 AM, Jan Lehnardt <jan@apache.org> wrote:
> 
>> Not sure what you mean with `the hosing feature`, but attachments are not
>> going away.
>> 
>> FWIW, I’d be in favour of having a modern CouchApp platform hooked into
>> CouchDB, but in the past 5 years we haven’t found anyone who’d be willing
>> to put in the work.
>> 
>> Best
>> Jan
>> --
>> 
>>> On 27 Feb 2017, at 10:58, Martin Broerse <info@martinbroerse.com> wrote:
>>> 
>>> Hi Jan,
>>> 
>>> If we split thinking about CouchApp in the hosting part and the backend
>>> coding part it would not hurt our usage if we lose the coding part. The
>>> coding part we need on the backend like password resetmails and other
>>> scheduled tasks are not there so the coding part needs to be more
>> powerful
>>> before we can use it. We can solve this tasks with OpenWhisk so perhaps
>>> keep the hosting feature and lose the rest?
>>> 
>>> The Ember guys at LinkedIn found it is faster to eval javascript loaded
>> as
>>> strings than loading the javascript from the backend. We have not tested
>>> this yet but if this is true we can perhaps bootstrap javascript apps
>> from
>>> strings hosted in CouchDB but we still need the CouchDB hosting part for
>>> the bootstrap code.
>>> 
>>> So in the future we are for keeping the hosting and lose the rest.
>>> 
>>> - Martin
>>> 
>>> On Mon, Feb 27, 2017 at 9:18 AM, Jan Lehnardt <jan@apache.org> wrote:
>>> 
>>>> Hi Martin,
>>>> 
>>>> thanks for your comment.
>>>> 
>>>>> On 27 Feb 2017, at 07:52, Martin Broerse <info@martinbroerse.com>
>> wrote:
>>>>> 
>>>>> We use the hosting from couchapp for many projects via
>>>>> https://www.npmjs.com/package/ember-cli-deploy-couchdb so keep it in
>>>>> couchdb. To replace excel sheets in businesses it is super you don't
>>>> need a
>>>>> separate hosting stack. An example couchapp hosted only on Cloudant:
>>>>> https://bloggr.exmer.com
>>>> 
>>>> Existing versions of CouchDB that support CouchApps aren’t going away,
>>>> and I’m sure Cloudant will keep things around for a while, too.
>>>> 
>>>> This is about the future of CouchDB and the non-existent developer
>>>> time that is required to maintain these features as CouchDB evolves.
>>>> 
>>>> Best
>>>> Jan
>>>> --
>>>> 
>>>> 
>>>>> 
>>>>> - Martin
>>>>> 
>>>>> On Sun, Feb 26, 2017 at 6:40 PM, Jan Lehnardt <jan@apache.org>
wrote:
>>>>> 
>>>>>> Aurélien,
>>>>>> 
>>>>>> I see that at least at some point you were subscribed and
>> participating
>>>> on
>>>>>> the couchapp@couchdb.apache.org mailing list. From the stated goal
of
>>>> the
>>>>>> list (find a new technical foundation for CouchApp) and the lack
of
>>>>>> significant engagement (users and devs alike) there, it should have
>> been
>>>>>> clear where this is headed.
>>>>>> 
>>>>>> And just to reiterate:
>>>>>> 
>>>>>> 1. CouchApp was an attempt to revolutionise web development as we
know
>>>> it.
>>>>>> — It failed, in like 2011.
>>>>>> 
>>>>>> 2. It was designed in a world before Node.js. Most folks who want
to
>> do
>>>>>> JavaScript and CouchDB have moved on.
>>>>>> 
>>>>>> 3. There are SEVERE technical limitations, most of which aren’t
as bad
>>>> as
>>>>>> a view index generator, but VERY bad for anything OLTP (think CGI
from
>>>> 90s).
>>>>>> 
>>>>>> 4. The features are unmaintained at this point, future refactorings
>>>> might
>>>>>> make the unavailable (e.g. in a http layer rewrite). The last
>>>> significant
>>>>>> work on the relevant code is 5-6 years in the past.
>>>>>> 
>>>>>> 5.We invited the CouchApp community to step up and build a
>> future-ready
>>>>>> version of CouchApps, complete with a design direction and own mailing
>>>>>> list.. Nobody stepped up, and at the end of the day, a project goes
>>>> where
>>>>>> developers can spend time.
>>>>>> 
>>>>>> 6. and to be clear, we are talking about: 1. _show & _list 2.
_update
>>>>>> funs, 3. rewrites // for the time being, we’ll keep
>> validate_doc_update
>>>> and
>>>>>> filter functions, but plan to replace them with per-doc access control
>>>> and
>>>>>> Mango schema enforcement. The idea of design docs, or attachments
on
>>>>>> documents are not going away.
>>>>>> 
>>>>>> In terms of ease of building web apps: a Node.js process running
next
>> to
>>>>>> CouchDB is only minimally more setup hassle and gives you:
>>>>>> 
>>>>>> 1. The same baseline features, plus a lot more.
>>>>>> 2. A simple app building model.
>>>>>> 3. A RICH ecosystem of third party libraries.
>>>>>> 4. WAAAAAAAY better performance and scalability.
>>>>>> 5. A future for you to do just the things you are already doing
>> without
>>>>>> moving to another platform.
>>>>>> 
>>>>>> Best
>>>>>> Jan
>>>>>> --
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>>> On 25 Feb 2017, at 18:22, Aurélien Bénel <aurelien.benel@utt.fr>
>>>> wrote:
>>>>>>> 
>>>>>>> Hi Joan,
>>>>>>> 
>>>>>>>> Your email is aggressive, and your apology is not accepted.
>>>>>>> 
>>>>>>> 
>>>>>>> I didn’t want it to be. I beg you for your pardon then.
>>>>>>> My frustration was real, but I can assure you that I am not an
>>>>>> aggressive person.
>>>>>>> There would not have been any ambiguity in my mother language
:
>>>>>>> discussing technologies in a foreign language is one thing,
>> expressing
>>>>>> your feelings is another.
>>>>>>> 
>>>>>>>> This topic has been discussed to death on the mailing lists
and I am
>>>>>> not going to be pulled into a retread of this argument.
>>>>>>>> http://mail-archives.apache.org/mod_mbox/couchdb-dev/
>>>>>> 201702.mbox/%3CB6DB98EC-42B1-4960-9E43-257F040238F1%40apache.org%3E
>>>>>>> 
>>>>>>> I’m just a « user »… a very dedicated and passionated user
(I’m in
>> the
>>>>>> top 10% on StackOverflow about CouchDB and I taught CouchDB to more
>> than
>>>>>> 150 french software engineers), but a user. That’s why I never
>>>> subscribed
>>>>>> to the « dev »  mailing list (or for a very short period of time).
I
>> now
>>>>>> understand that I should have, but it’s too late.
>>>>>>> 
>>>>>>> My frustration is as high as has been my passion for six years
for
>> this
>>>>>> incredibly interesting project.
>>>>>>> I respect the board decisions but now I will have a hard time
finding
>>>>>> money (which is sparse in academic research) to move all of our
>>>> software to
>>>>>> a different technology stack and arguments to explain to all of my
>>>>>> collaborators that I bet on a technology stack that got rapidly
>>>> deprecated.
>>>>>>> 
>>>>>>> Thank you for your understanding.
>>>>>>> 
>>>>>>> 
>>>>>>> Best regards,
>>>>>>> 
>>>>>>> Aurélien
>>>>>> 
>>>>>> --
>>>>>> Professional Support for Apache CouchDB:
>>>>>> https://neighbourhood.ie/couchdb-support/
>>>>>> 
>>>>>> 
>>>> 
>>>> --
>>>> Professional Support for Apache CouchDB:
>>>> https://neighbourhood.ie/couchdb-support/
>>>> 
>>>> 
>> 
>> --
>> Professional Support for Apache CouchDB:
>> https://neighbourhood.ie/couchdb-support/
>> 
>> 

-- 
Professional Support for Apache CouchDB:
https://neighbourhood.ie/couchdb-support/


Mime
View raw message