Path: news.daimi.aau.dk!news.uni-c.dk!newsfeed.sunet.se!news01.sunet.se!sunic!mn6.swip.net!seunet!news2.swip.net!news.nocom.se!newsfeed.tip.net!news.seinf.abb.se!nooft.abb.no!Norway.EU.net!nntp.uio.no!ubmac86.uio.no!user
From: kolbjorn.aambo@ub.uio.no (Kolbjørn Aambø)
Newsgroups: comp.lang.beta
Subject: Re: B*-Tree and other databases
Date: Wed, 12 Jun 1996 12:12:46 +0100
Organization: University of Oslo Library
Lines: 57
Message-ID: <kolbjorn.aambo-1206961212460001@ubmac86.uio.no>
References: <v7gspc2yd8q.fsf@schroeder>
NNTP-Posting-Host: ubmac86.uio.no

Claus-Peter,

I gave up that idea some months ago because there where no way to be sure
that the Garbage Collection did not mess up my block of memory while I was
reading it from disk. The Mjølner people warned me that if I gave an
adress to a Operating system and asked a Operating system function to
write a block starting at that address the GC could by aksidens start
looking for more memory something that may implyng not finding more! They
gave me no existing mekanism for asking if there was a fixed amount of
memory left that could be used without GC starting but saild that such a
function may emerge in the future.

This was one of the courses that made me start programming the same thing
in C++   ;-|. My work is not a clear cut B-tree implementation, but more a
simple relational system. 

I have been working a little with Beta since 1990 and find it much more 
elegant than anything else. The problem is that not much money and 
only a few peoples efford are put into developing an actual system
development environment. I find this much more developed in systems like
Metroworks Codewarrior, a Macintosh based development system. It also have
a Java but it's still not mature enough for my taste. Probably it will fit
in with pascal, c, c++, 68K asm, PowerPC asm and Intel assembly in one
seemless debugger environment in the next version due in the fall. I find
no such things on my Mac that support BETA. BETA without source-level
debugger, Unicode, 64 bit integer and the before mentioned shortcomings in
memory handling  and not less essential without wendor support is not much
to meet the future with. 

The before mentioned functions should not be new to Mjølner. 
Sun and Apple seperately have sponsored the Mjølner Beta effort to some
extent. Still Sun possibly with open eyes left BETA for a much less
elegant language that have a lot in common with BETA predessor SIMULA-67,
only with C syntax.........

By the way a easily configurable B-tree implementation should have been
integrated into any language system years ago, but there we have the
cultural problem Compiler people know of no other database systems than
file-systems...

And database people ar busy making systems in quite arcane and old
fashioned "languages" like ER and SQL that are too closely connected with
the shorcomings of the underlying technology.


In article <v7gspc2yd8q.fsf@schroeder>, Claus-Peter Knaut
<knaut@ls6.informatik.uni-dortmund.de> wrote:

> Hello,
> 
> I have to write a B*-Tree in BETA. Did someone tried this bevor or
> used the Barkley-DB-Tool for this or some other tool ?
> 
> For futher information in secondary storage in BETA i would be
> thankfull
> 
> CP