Non-licensed DBM package, anyone?

[Ozan Yigit]

In article <12497 at ut-emx.UUCP> clyde at ut-emx.UUCP (Clyde W. Hoover) writes:
>Are there any DBM workalikes which are NOT entangled in AT&T licenses? 

Yes !!! (details below)

>I would like to have one to redistribute with a program I have written.  

Sure..

>I recall there was someone who wrote one supposedly based from the 
>ground up but I can't remember enough about it.

I do not know about this... Torek's MDBM is not distributable,
because it is based on dbm/ndbm. Something called dbz was posted: it
is only suitable for B-News, and it too borrowed (?) dbm's hash
function.

I am nearly done with a complete re-write of ndbm (called sdbm for
substitute dbm), based on an algorithm circa 1979 (Extendible
Hashing) by Fagin, Nievergelt, et al. I am not sure if dbm/ndbm was
based on the same algorithm or not (who cares), but sdbm is currently
binary-compatible with Berkeley ndbm. (Bell dbm interface can be
written on top of ndbm, but that is later...) At the file level, it
is similar but non-compatible with dbm/ndbm. For one, sdbm does not create
those famous "holes", and it will (either in the first release or a later
release) support varisize data blocks, so your key+data size is not
restricted to some stupid little number.

Sdbm release will (a la oz) be fully public domain, so you can be free
to do whatever you like, hoarde it or sell it, or claim it to be yours...
It will also be readable, so even ordinary mortals can understand it.

Expect a release within few weeks, after I am done polishing and testing.
[a copy will be sent to BSD folks, so they can replace their licensed code
with it.]

oz

-- 
use the source, luke !!     	        Usenet:    oz at nexus.yorku.ca
uh... we forgot to tell you...          ......!uunet!utai!yunexus!oz
it is unintelligible, but hey, you      Bitnet: oz@[yulibra|yuyetti]
got it, for free (!).                   Phonet: +1 416 736-5257x3976