By WHAT do you (in parallels) structure? (woof-woof!)

Posted by Amontillado
Oct 24, 2022 at 10:51 PM

22111 wrote:

>In a word then: Outlining’s just an ordering instrument, between mesh
>reality / mesh conception, and then serial, linear presentation, but
>don’t be as naive, or dishonest, as to make it your fallacy: neither for
>the audience, nor, and especially so, for your own thinking.
>

Yes, quite so! Sometimes I use tags in Devonthink to capture little constellations of notes that relate in some way even though they are not about the same topic.

Regarding problems with outlining and sqlite, I respectfully disagree. SQL amounts to a system of file storage, indexing techniques, and what amounts to a macro language. That doesn’t imply any particular structure in what the user sees. An SQL database provides a place to put things and ways to find and retrieve them. You could think of the database as a disk drive with enhancements. You hand it data to store, and ask to get it back when you want.

SQL backends support all kinds of software. Circuit board layout, for example, which is all about networks.

It’s not crazy to think of a database as an analog for a fancy disk drive. There are such things as hardware databases.

For instance, IBM’s Netezza.

You can provision a Netezza system, spanning multiple racks, with petabytes of storage. There is an interesting twist, though.

The database disks aren’t read or written by the operating system, at least not directly. Hardware gate arrays act on the data to both store and retrieve database records. It is sort of like each of hundreds of disks has its own little dedicated database engine to do it’s part in larger operations.

The operating system in a Netezza (Linux) has an SQL system based on Postgres. SQL queries act on the hardware gate arrays, and performance might be 100’s of times faster than anything running a conventional SQL system. Queries on a database with, say, 50 billion records might take less than a second. Serious horsepower. Cost is serious, too.

But I digress.

The most perfect network or mesh based knowledge system will store its data in a computer filesystem, NTFS, AFS, FAT, EXT4, UFS, or something. That underlying filesystem usually doesn’t shape the user’s experience.

If I were going to write any sort of utility that worked with a body of data, I’d strongly consider an SQL engine. Why not? You can tap into well-debugged store-and-retrieve systems for free.