Plan for a "Generic Record Sync"

Thom Chiovoloni tchiovoloni at mozilla.com
Tue Jan 29 23:52:56 UTC 2019


Okay, I've converted it into a doc:
https://docs.google.com/document/d/1CpfoO7fUkK3U9ED2OPZZ9wyZpo7UXPXcC-30oB5sZNE/edit#

Maintaining syntax highlighting doesn't seem clear how to do, sorry a lot
of it is code, but it was clearer to organize my thoughts that way.

On Tue, Jan 29, 2019 at 3:18 PM Nicholas Alexander <nalexander at mozilla.com>
wrote:

> Thom,
>
> This is great!  I'm really pleased to see you (and really, the larger a-s
> team!) thinking about extensible Sync 1.5.  Sync 1.6, if you will :)
>
> On Tue, Jan 29, 2019 at 2:29 PM Thom Chiovoloni <tchiovoloni at mozilla.com>
> wrote:
>
>> I've been kicking around ideas for this for a while, and finally wrote
>> them down. Essentially, it's a plan for us to get some of the benefits of
>> Mentat without all of the troubles required to implement Mentat. In
>> particular, it gets us ease of implementing new data types, some degree of
>> schema evolution, and inter-record references (with some limitations).
>>
>> My big fear would be that without something like this, teams will turn to
>> off the shelf sync solutions that don't offer the same crypto/privacy/etc
>> protection that they would with Sync (E.g. things like firebase, etc).
>>
>> It also has a more limited scope (no support for history, no support for
>> record types with constraints as complex as bookmarks).
>>
>> I wrote it as a github gist so that I could embed code with comments and
>> have it syntax hightlighted, although I realize in retrospect that a google
>> doc would allow for more flexibility with commenting and etc.
>>
>> https://gist.github.com/thomcc/d45cebb959a005bb915a1b34cd4fe215
>>
> Mmm, this does make it difficult to comment -- and I have many comments
> :)  There are ways to convert gists to Google docs; could you do that so
> that we can manage the conversation?  Or inline the text here and we'll do
> it by email, although I think many folks won't appreciate the verbiage.
>
> I think we should run this by the Kinto folks as well.  I feel that if we
> can't implement "basically Kinto.js" on top of this then we're probably not
> doing the right thing.
>
> Best,
> Nick
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/sync-dev/attachments/20190129/4e2d22e7/attachment.html>


More information about the Sync-dev mailing list