Scope: the service takes an article name, and serves article content in a text-only format (but including wikilinks and external links).
|Resolved||Deskana||T75616 Tracking: API/backend issues blocking Wikipedia app development|
|Resolved||• GWicke||T87824 Create specialised content API for mobile apps|
|Resolved||bearND||T90228 [Spike 1 hr] Investigate building a Node.js service for mobile apps content|
This sounds to me like awkwardly similar to what mediawiki does now... it could just be a new api endpoint, right? After all we already have the code that does just that (takes a page name in the url, and spits out a json version of the contents of the page).
Also, why does the service need to be "in nodejs"?
There was quite a bit of controversy in the thread you mentioned on the scope, architecture, general usefulness and also maintainability of such a service.
This spike is for an experimental service for us to get to grips with how to build services. The goal is not intended to use this in production, it's for us to build a service from start to finish to learn the pain points.
After this, we'll evaluate the strengths and weaknesses of the approach, and decide whether to build something more fit for our production purposes, as noted in the thread I linked.
And yes, there was concern, but also I've had a lot of people approach me with support. The only way we can truly know whether this works is by cutting through the talk and actually trying it. That's what we'll be doing.