Brian. You don't like the idea. I get that.
If I proposed storing financial data with sales in one flat file. Customer numbers in another. Orders in yet another. And an index to all those files in yet another, you would be amongst the first people to advocate moving that data into a proper database.
The same holds true if it was an airline reservation system. Or a stock control application. Or almost any other set of data that had dependancies and relationships.
Why would I store that in something just to pull it out again? What structure is going to make that more simple?
The rest of the distribution looks the same, and the at the user level most people will not notice a difference.
You have read way to much into what I wrote--I made no proposals for what schema should be used, and you are showing no imagination for what could be done.
How about if the compiled bytecode was also stored the first time the module is used and subsequently that was returned.
How about if modules recorded the maximum or average sizes of arrays created by different calling applications and used that information to preextend their internals arrays accordingly on a per application basis.
How about ...
In reply to Re^4: CP6AN should be a database
by BrowserUk
in thread Notion: CP6AN <strike>should</strike> could be a database
by BrowserUk
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |