Alex Burgers PicoLisp DB tutorial describes indirecty the underlying format: http://software-lab.de/doc/tut.html#db
"block size of 256 Bytes", talking of a CAR and a CDR which indicates, that it is kind of "functional database", comparable to ultra expensive DATOMIC. Alex also writes about kind of btree allowing 'soundex' search which indicates, that there must be specialized index files for faster and tolerant search. PicoLisp has transactions built-in, rollback is possible. PIL or Prolog in PicoLisp allows you to build up highly sophisticated (also remote, distributed) queries from these tree primitives, but it can happen, that suddenly you have accidentally created millions of queries ... NVMe SSD's (>100,000 IOPS) highly recommended ... You finally must know, what you are doing. You can't compare PicoLisp's functional, (btree, soundex) indexed database with highly specialized and optimized engines like PostgreSQL or Neo4J, engines using modern GraphBLAS algorithms like Redis. But finally you will be able to build all kinds of (distributed) databases, even (distributed) graph databases, (distributed) fault tolerant databases with PicoLisp just with a few lines of code. Nothing here speaks against building own, specialized indexing daemons written in either PicoLisp or in C as module, which i have done (Fusion Tree). That's mighty, mighty stuff ... finally far superior to commercial solutions (e.g. MongoDB). PicoLisp is very tiny, very well thought through. One person can fully understand and master (even modify) the whole PicoLisp ecosystem within just a few weeks, C/Lisp knowledge required, of course! That makes PicoLisp unique. C K Kashyap <ckkash...@gmail.com> schrieb am Di., 28. Mai 2019, 22:46: > Hi, > Is there documentation about the file format of the database file in > PicoLisp? > I am looking at the possibility of using it for the tripple store > <https://en.wikipedia.org/wiki/Triplestore>. > Regards, > Kashyap >
