One additional data point here. I mention "serious customers" as narrowly defined in this email as the thousands-to-millions of $ customers. CouchApps probably have a place in the lower end of the market, i.e. shared instance users who have lightweight needs for their applications and are customers of IrisCouch or SmileApps. I wasn't trying to say there isn't a market for this :) The business case to be made for them is very different, i.e. razor thin margins across thousands to millions of people. Such an approach wasn't logical for Cloudant - the shared instances don't drive the company like the dedicated instances do. Because of this data I think CouchApps as a primary user story is very hard road to walk for us.
----- Original Message ----- > From: "Joan Touzet" <woh...@apache.org> > To: marketing@couchdb.apache.org > Cc: "Mike Broberg" <mbrob...@us.ibm.com>, smet...@uk.ibm.com > Sent: Saturday, May 9, 2015 2:26:26 PM > Subject: Re: the future of couchapp > > Hi Miles, > > DISCLAIMER: I am not speaking as an official representative of IBM or > Cloudant. I have cc'ed Mike Broberg, who can speak for them if > necessary. (I also want him to be aware of what I am saying here). > > *** TL;DR: the people who are willing to spend anywhere from > thousands to millions of dollars on a CouchDB-based solution aren't > interested in CouchApps. I think the discussion to date is missing > this, and as such, is entirely unrepresentative of the current > market for Apache CouchDB. > > The answer is that there are practically no customers of Cloudant/IBM > who are banking on CouchApps for any serious need. Every client that > I can think of - meaning they have a dedicated cluster, and aren't > using the shared cluster service - are using either a traditional > three-tier app server structure (Node.JS, Python, PHP, Ruby, Java, > .NET, etc.) or are doing client-side development on mobile platforms > (iOS + TouchDB, Android + PouchDB) where they are replicating back to > the Cloudant clusters for data exchange. In all of these scenarios, > replication is the "killer feature" for CouchDB, with the REST > interface a close second, and the ease of unstructured JSON data as > a third. > > Cloudant built out a document-level (and field-level!) security > solution for one customer, about two years ago now. While there was > initial interest, performance considerations lead to the solution > being backburnered for further consideration. Even in that situation, > CouchApps weren't the primary concern -- database-level enforcement > of security rules *was*. > > Within Cloudant, perhaps Simon Metson was the primary proponent of > using CouchApps for serious purposes. He used them in the "For > Developers" section of the website to help demonstrate various key > features of the platform, including the new MongoDB-inspired Mango > feature that's now a part of CouchDB 2.0. Diana Thayer (@garbados) > picked up on this and built a documentation framework on top of > CouchApps. This, to me, is perhaps the ideal use of CouchApps: > unsecured content, read-only, displayed in different formats based > upon what the end user needs, and self-hosted by CouchDB (so you > can view the product's documentation using the product itself). > More information on this use is at: > > https://mail-archives.apache.org/mod_mbox/couchdb-dev/201410.mbox/%3C28603443.66.1414446738764.JavaMail.joant@Joans-MacBook-Pro.local%3E > > > > > > ----- Original Message ----- > > From: "Miles Fidelman" <mfidel...@meetinghouse.net> > > To: marketing@couchdb.apache.org > > Sent: Friday, May 8, 2015 11:21:28 AM > > Subject: Re: the future of couchapp > > > > Let's be clear. > > (Good) marketing isn't about selling a solution to folks who don't > > have > > a problem in the first place, it's about it's identifying problems > > for > > which we offer a solution. > > > > And.. it occurs to me that Cloudant has been doing market research > > and > > "real" marketing - perhaps some folks from Cloudant might share > > some > > findings related to CouchDB (as opposed to those that might relate > > to > > their commercial extensions and services)? > > > > Miles Fidelman > > > > > > > > Giovanni Lenzi wrote: > > >> translates user@ decisions in "how to drive them to the public"? > > > or maybe better how to drive dev@ implemented features to the > > > public ? > > > > > > 2015-05-08 16:57 GMT+02:00 Giovanni Lenzi > > > <g.le...@smileupps.com>: > > > > > >> Got it, Joan. Thanks for the useful reminder, considered I am a > > >> total > > >> newbie here, I definitely don't know how decision-making process > > >> is driven. > > >> > > >> We will cut the "features" part from this discussion then and > > >> take > > >> it to > > >> the devs@ list > > >> > > >> Here we should then focus on @jan's request about the story for > > >> couchapps.. given that until 2 days ago that was somehow > > >> uncertain > > >> > > >> But I think too this is more a user@ topic... isn't maybe > > >> marketing more > > >> appropriate to translates user@ decisions in "how to drive them > > >> to > > >> the > > >> public"? If you all agree with that, you can move this > > >> discussion > > >> to user@ > > >> or dev@, don't know what is preferable. > > >> > > >> > > >> 2015-05-08 15:56 GMT+02:00 Joan Touzet <woh...@apache.org>: > > >> > > >>> Hi all, > > >>> > > >>> PMC hat on... > > >>> > > >>> Reminding you *again* that we should not be using the MARKETING > > >>> list to > > >>> discuss new FEATURES and functionality for Apache CouchDB. We > > >>> are > > >>> not > > >>> like a company where marketing makes up what they want to do, > > >>> and > > >>> development is forced to implement it. While it's a good idea > > >>> to > > >>> have a > > >>> feedback loop between marketing and development, I am > > >>> especially > > >>> keen to > > >>> not see Apache CouchDB turn into a marketing-driven development > > >>> effort. > > >>> > > >>> If you are proposing new CouchDB features, please make those > > >>> proposals > > >>> on the dev@ mailing list. And if you are willing to *develop* > > >>> and > > >>> *support* those functions - even better. Current CouchDB > > >>> development > > >>> bandwidth is extremely limited, and would best be served by > > >>> helping you > > >>> to understand the current design's constraints, and the > > >>> difficulties > > >>> that may be inherent in what you ask for. > > >>> > > >>> Best regards, > > >>> Joan > > >>> > > >>> ----- Original Message ----- > > >>>> From: "Giovanni Lenzi" <g.le...@smileupps.com> > > >>>> To: marketing@couchdb.apache.org > > >>>> Sent: Friday, May 8, 2015 4:05:12 AM > > >>>> Subject: Re: the future of couchapp > > >>>> > > >>>>> A service-trigger feature could be one of the new features of > > >>>>> Couch > > >>>>> apps. > > >>>> if possible, would be awesome :) > > >>>> > > >>>>> some clear design goals and a very limited set of features to > > >>>>> add > > >>>>> to > > >>>> CouchDB ddocs and focus on an in-browser tool (add features to > > >>>> Fauxton) > > >>>> that removes the need for new developers to learn git and > > >>>> build > > >>>> tools > > >> > > >> > > >> -- > > >> Giovanni Lenzi > > >> www.smileupps.com > > >> Smileupps Cloud App Store > > >> > > > > > > > > > > > > -- > > In theory, there is no difference between theory and practice. > > In practice, there is. .... Yogi Berra > > > > >
