Documentation - and what are your plans?

[etoulas]

Hi @foofoofoo97 @lokingwei

I stumbled across your Flutter library.

We were looking for something like this to connect to CouchDB. I see you implemented even two different stores including the very promising ObjectBox.

We are not very sure if we should adopt this library. The listing on pub.dev does not look very trustworthy, but I can see that you are both working for a reputable company.

The lack of even the most basic documentation raises doubts if this library should be considered for any serious app.

Can you briefly explain what are your plans?
Is it a proof-of-concept?
Do you use it already in a production app?
Do you plan to write documentation?

[lokingwei]

Hi @etoulas, thanks for your interest in our library. To answer your question:

Can you briefly explain what are your plans?
Our plan is to create a PouchDB liked database warper for dart and flutter. Our product was previously powered by PouchDB + CapacitorJS, and this project is intended to support migration to flutter

Is it a proof of concept?
Do you use it already in a production app?
Not a POC as the library already running in our production application, doing the live-syncing for about 3 months

Do you plan to write documentation?
Yes, we do understand that basic documentation is needed to encourage external contribution, hence we will start providing documentation once we arranged the man-power

[etoulas]

Hi @lokingwei, thanks for your answers!

That sounds really promising.

Considering that you already run this in production, what would you say is missing compared to the PouchDB library?

Which storage are you using in production: ObjectBox or Hive?

And how is the sync performance working out? We would need to sync a few hundred MB of data (product catalog) to the devices.

I'm trying to figure out if a collaboration on this library would be possible on our side.

[lokingwei]

@etoulas, the following feature is not yet ready

1. Attachment (not in our use case, will need PR)
2. Javascript-based Map-Reduce Function, (still researching suitable JS runtime for flutter)
3. Isolate to handle compaction and indexing (currently implementation, compact then save, index then read)

Our production app is using ObjectBox adapter, both performances are almost the same. But we choose object box instead of hive because of large hive database potentially have memory issue.

Our use case is document count intensive, hence I can't justify the performance in terms of document size. But for your reference, it took around 15min to replicate 10,000+ docs into our android POS system (Rockchip) which has much lower spec compared with modern commercial android phones (Samsung). The performance can be 2x to 3x with a decent android phone.

And also the production app is running on android and windows. iOS will release to production by end of this month.

[etoulas]

Thank you very much for the detail answer @lokingwei!

For context: We have up to 80,000 docs with 3 KB each.

Some users have devices from a few years ago, so not the latest and greatest in terms of performance.

Our target is Android and iOS.

[RokeAlvo]

Hi @lokingwei

Can you tell me what the current status of the repo is? And what are your future plans?

[VladimirCores]

Also would like to have at least basic example, so wont spend time on looking in the code.