Taking quotes on building LC external for RethinkDB
mark at livecode.com
Sat Aug 5 16:36:44 EDT 2017
Err - given that revDB is an *SQL* database wrapper and MongoDB is not an SQL database you can imagine that creating an abstraction layer to deal with both might be 'quite' hard - if not impossible.
So, given that we've now been open source for over four years and as such that code has been open for others to contribute to for four years please feel free to try and propose an API which works relatively seemlessly for all types of databases that exist today - I'll happily review it and help ensure it is worthwhile.
Less 'rants' and more actual intellectual effort please. This isn't down to static and dynamic libraries, that is just the colour of the bikeshed at this level - and trying to claim that anything related to that is the issue here is just misleading people.
P.S. My point here is simply that whilst their may be an abstraction which unifies all databases it sits so far up in the abstraction hierarchy that trying to make revDB do it would be entirely pointless. The 'revDB' rewrite you speak of has been subsumed by a number of db libraries which solve the issue of low-level access that revDB provides. Whether that be dblib, sqlyoga, or the way to abstract in a domain specific case as typified in many apps (e.g. The photo app in the createit course) have provided. Indeed, let's not pretend that the issue with supporting all kinds of 'database' here is at the C / API level because it really isn't.
P.P.S. Abstractions are what makes things easy - abstracting too far and all you end up doing is proving that 1 = 1 in numerous not particularly interesting ways. What you might call "category theory 101".
Sent from my iPhone
> On 5 Aug 2017, at 19:39, Mark Wieder via use-livecode <use-livecode at lists.runrev.com> wrote:
>> On 08/04/2017 05:38 PM, Alex Tweedly via use-livecode wrote:
>> I have to admit that rethinkdb sounds really interesting - I hadn't heard of it until your posting.
>> Might be worth a crowdfunding / donation request to spread the cost; while I don't have a *need* for it, it might be a worthy target for (a small amount) of my optional spending of my 'pocket money' ;-)
> <rant continued>
> In 2006 all existing database bugs in bugzilla were rolled into one omnibus 'revDB review' bug report, and the individual report statuses were all changed as 'resolved'. This in favor of 'We will shortly be reviewing revDB' for a major rewrite of the database layer.
> Had this actually been done anytime in the intervening eleven years, adding new database types would be much easier. At some point I tried to add mongodb to the engine and eventually hit a brick wall because of an incompatibility with the existing library structure (a clash of static and dynamic libraries, IIRC)
> I realize revamping the database layer is a bigger task than trying to shoehorn more database types into the existing bucket, but I think it's high time to revisit (crowdfund) this. Otherwise we're just digging ourselves deeper into the existing muck.
> Mark Wieder
> ahsoftware at gmail.com
> use-livecode mailing list
> use-livecode at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your subscription preferences:
More information about the Use-livecode