In the standard triplet, [key,link,pointer], the bson overloads get a byte of link which now looks like:
[two free bytes,bson codes,graph codes], out of a total of four bytes.
Bson gets us all the basic types and (we hope) all the basic opcodes to support graph microsequences. I am thinking that three bits of graph can point to standard schema, eight of them, with fast code per each. Thinking, but not coding. I could aim this machine straight at the sweet spot in the market, hmm....
How do I market the thing?
A fully BSON, JSON graph traversal search engine built atop sqlite3. Full support for automated web bots, relative freedom to use name:value pairs through, with fast scan thrus. Yet, still, direct execution of natural square sql tables. Loose table schema, immediate (dynamic) schema, inherently multi-processor, and all search directed by ontology graph. Imagisoft would own the market. The thing would self organize, customers just adding Linux boxes as needed. It would be, my Imaginary staff vs MangoDB, a cute sounding fruit.
No comments:
Post a Comment