I would go the other way: The visualization and the back-end should not be strictly part of the Zeppelin core, while a generic interface to manage diverse pluggable backends though pluggable front-ends with a common core of managing notebooks and data structure would go a long way towards turning a diversity of notebooks into a single notebook with inherent diversity and customizability - getting us much closer to an R/RStudio-like situation.
A REST-API, in particular for the output of a paragraph, would allow much more flexible development of visualizations. As for the input - I doubt that it's necessary, but having a stricter separation of all the different components would move the community closer to an interchangeable standard of front-end <--> notebook <--> backend. This also helps ease a lot fo the current pains of the build and packaging of Zeppelin. I'm very much in favor of this, and would be willing to invest some (as much as I can...) time into it. On Thu, Nov 5, 2015 at 8:59 AM, IT CTO <goi....@gmail.com> wrote: > I am wondering if this does not make zeppelin an advance Spark-submit > framework for multiple languages? > As I see it, the visual representation of the result combined with the > back-end language support is the core of zeppelin and REST api for > submitting a notebook kind of kills it? > How do we imagen people interact with the results in such option? > Eran > > בתאריך יום ה׳, 5 בנוב׳ 2015, 07:36 מאת moon soo Lee <m...@apache.org>: > >> Hi, >> >> Many people want to have REST api to run a notebook. >> I think there're no reason to not having this feature. >> It's also related to https://issues.apache.org/jira/browse/ZEPPELIN-137. >> >> Thanks, >> moon >> >> On Sun, Nov 1, 2015 at 2:07 PM denny wong <tecgi...@yahoo.com.invalid> >> wrote: >> >> > Hi >> > Is there a plan to add a REST api to run a notebook? >> > Thanks >> > -- > Eran | "You don't need eyes to see, you need vision" (Faithless) >