erm... my comment had nothing to do with what backend is used, in case that wasn't clear. I was commenting on ease of use of designing database for the app and integrating it, for which there are exactly zero tools in LC, other that what is afforded by each database - meaning the developer has to program every aspect of database interaction with no assistance from the IDE - the opposite of FMP where the database is both managed by the IDE and integrated into controls, scripting etc.FourthWorld wrote: ↑Sun Oct 02, 2022 5:06 pmOne advantage to things made with the LC technology is that LC has no built-in database, instead letting the developer use any of the most popular DB engines in the world, with all of the documentation, support resources, and tooling that have made them the world leaders.
Did this really make you sad Richard? You mean a file-based system should be completely immune to any sort of problem arising from errors generated by interrupting a write-to-file for example? I find it reassuring, not saddening that this is there. I have never lost any data, ever, by misusing an FMP database, but i have done when misusing an SQLite db (and don't compare with server-based databases - 'repair database' is only for a file-based DB and is not used for server DBs)FourthWorld wrote: ↑Sun Oct 02, 2022 5:06 pmLast time I used FMP I was saddened to see it still had a "Repair Database" item IN THE FILE MENU.
Is this a reflection of FMP and LiveCode, or a reflection of the market segment targeted?FourthWorld wrote: ↑Sun Oct 02, 2022 5:06 pmAfter another few calls from prospective clients leaving FMP, I did a little recon to try to assess mindshare of that platform over time. Just from personal contacts with FMP devs and having attended local FMP user group meetings I had at least an anecdotal sense of attrition rates, but I was surprised to see a subsidiary of one of the wealthiest and most powerful corporations on the planet show search interest only slightly greater than interest in humble LiveCode
Regarding attrition: 99% of the reasons for leaving FMP are the cost. Nothing else. Claris is a greedy subsidiary of Apple and actually reminds be of the worst of the pre-OSX times.
But none of the above mean that FMP's design patterns are not worth emulating.
My point, to re-iterate, was a speculation how to improve ease-of-use of database apps. Appli does that well in some circumstances, by anything more complicated requiring complex relational database isn't really well suited. As you say, Appli serves a different need and does this well.
What would be nice is if LC had a 'database studio' where you could assign any DB back end, graphically design relationships and contexts (similar to SQL views or table occurrences in FMP), and have an integrated way of inserting fields into layouts and data grids for example (or Appli layouts!) as well as integrate with scripting (auto-completes, contextual field selection (eg DbTable::field - typing DbTable:: brings up a list of fields in a popupmenu in the script editor and so on). Actually the scripting integration, rather than just control integration is one of the big strengths of FMP.
Yes, yes, yes - before you say it, i KNOW you can program every aspect of it and this would not confer to new functionality (well, excepting things like calculation fields which you'd have to program by hand in the IDE). But it would speed up dev time significantly as well as making it easier for the less experienced to tackle complex database structures and bringing these directly inside the IDE graphically.
Right now if i'm using SQL for a complex relational DB, I would use Valentina Studio - it's a close approximation to the utility of FMP in designing a database for use in my projects, but lacks several features that made my life easier in FMP.
Just saying that bit would be nice to have these features integrated completely with the IDE, without this being a comment or comparison with Appli or FMP...