[REBOL] Re: Dialecting......
From: gregg::pointillistic::com at: 7-Feb-2009 10:59
Hi Srini,
SI> Given the power of Rebol, I was wondering whether the guys are
SI> thinking of a datastore that would mimic the natural
SI> expressiveness of Rebol and provide a database without the
SI> issues faced by traditional RDBMSes or OODBMSes.
Two other models have been mentioned in the past, though neither is in
place AFAIK. The first is RIF (REBOL Indexed Format), which is a
lightweight model that is still very block oriented, but doesn't force
you to load an block to retrieve a single item. Carl was planning this
as a native element. In the meantime, someone has already done an
experimental "virtual block" scheme that makes persistent blocks
transparent, or very close, in use.
The second model that was talked about, I think a the Paris DevCon,
though I wasn't there, is an Associative Data Model. I haven't heard
anything more on it, and it would be quite ambitious I think.
Ultimately, I think REBOL should be used to implement all kinds of
DBs, languages, and systems. Not necessarily as production tools, but
as teaching tools.
SI> Of course response time, etc would be a factor, but for small to
SI> medium size datastores this should not be a problem...
I'm not one to suggest that we waste resources frivolously, but I'm
all for using and leveraging the power we have now, and designing
for what is actually needed, rather than some theoretical target.
Plain old REBOL can handle very reasonable amounts of data.
-- Gregg