[elephant-devel] 4th European Lisp Workshop

Ian Eslick eslick at csail.mit.edu
Wed Apr 11 18:05:24 UTC 2007


You should check out Arthur Lemmon's Rucksack, it's architecture is  
much more ZODB like.

A couple of people are also interested in doing an all-lisp data  
store for Elephant that gets rid of any license issues.  I believe  
that it may be possible to make the native solution run faster than  
BDB.  I'm still stuck on how to perform inexpensive locking or atomic  
sections (for lock-free datastructures) without external libraries.

Ian



On Apr 11, 2007, at 1:37 PM, Ben wrote:

> this is an entirely anecdotal remark, not really relevant, but one of
> the original inspirations for elephant was a database system that was
> used in-house at a company i was at with andrew blumberg and other
> lisp hackers.  i was doing java crap and was a complete lisp neophyte,
> and was amazed to see "change-class" in action.  that was a disrupting
> experience.  also, i had been worried about the object-relational
> impedance mismatch for a while.  BDB-backed elephant was a way to get
> persistence without a table structure, and hopefully reap some
> benefits in flexibility.  the original target applications were web
> stuff and a mother-of-all-email-clients which never materialized.
>
> i still think writing a ZODB-like backend would be a good idea,
> especially given sleepycat's licensing issues.
>
> take care, B
>
> On 4/11/07, Pierre THIERRY <nowhere.man at levallois.eu.org> wrote:
>> Scribit Robert L. Read dies 09/04/2007 hora 18:43:
>> > However, if someone else can go, I am also happy to help or to
>> > co-author a paper with them.
>>
>> It occurred to me that the combination of persistence and the MOP  
>> goes
>> far beyond anything else I know in terms of flexibility WRT to the  
>> data
>> model.
>>
>> That is, even a powerful macro system to build code that accesses  
>> a DB
>> sets in stone the data model at compile-time. But a persistence  
>> library
>> based on the MOP can bring to stored data the dynamic nature of CLOS,
>> and you could completely change how data is organized, while the
>> application runs, without having to build a migration tool, even  
>> if this
>> process is automated (which is hard, I suppose), like cl-migrations.
>>
>> Like Erlang, Common Lisp enables a system to remain live while being
>> updated at the code level. An Elephant-like persisence library  
>> enables a
>> system to remain live while being updated at the data model level  
>> also,
>> almost for free.
>>
>> I think this could easily make a short paper, maybe even a long  
>> one. I
>> will try to write a first draft of it, and probably ask for help  
>> at some
>> point.
>>
>> Quickly,
>> Pierre
>> --
>> nowhere.man at levallois.eu.org
>> OpenPGP 0xD9D50D8A
>>
>> -----BEGIN PGP SIGNATURE-----
>> Version: GnuPG v1.4.6 (GNU/Linux)
>>
>> iD8DBQFGHRogxe13INnVDYoRAvePAKDqISbcjuxdCxcG+v5sgvkrK9pdqgCgiOs1
>> MCQO7iuqMvsotPM1iQA8ZoU=
>> =ez7O
>> -----END PGP SIGNATURE-----
>>
>> _______________________________________________
>> elephant-devel site list
>> elephant-devel at common-lisp.net
>> http://common-lisp.net/mailman/listinfo/elephant-devel
>>
> _______________________________________________
> elephant-devel site list
> elephant-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/elephant-devel




More information about the elephant-devel mailing list