Francois-Rene Rideau <firstname.lastname@example.org>
Sun Dec 2 06:08:02 2001
the Glossary ought to reply to most of your questions about orthogonal
persistence. I'll have a (very slightly) updated version soon.
As for what a persistent system can look like, an excellent read is
A Persistent System in Real Use - Experiences of the First 13 Years
Joche Liedtke (1993)
As for how to achieve it, it's usually done based on MMU.
See e.g. the docs for Grasshopper, the Texas persistent store, Mungi,
or the british FIDE, FIDE2, Staple, etc., systems.
Actually, techniques used in file systems apply (see the Linux Tux2 fs
or journalled filesystems), the main difference being that in a persistent
system, your have persistent at the object level rather than the file level.
As for categorizing methods, orthogonal persistence makes it
completely independent from the persistence of data.
You can do it with hierarchies, association tables, lists, pools,
or whatever method suits the problem at hand.
That is, you access structured data, variables, etc., as if they were mmap()ed
in a safe and coherent way to persistent memory, instead of having to
manually write application-specific synchronization code over and over.
There will likely be hierarchical namespaces as a standard data repository,
but these namespaces will map names to objects (i.e. structured data),
not to "files" (i.e. sequences of raw bytes).
For the email list, see on
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[ TUNES project for a Free Reflective Computing System | http://tunes.org ]
Government is the great fiction through which everybody endeavors to live
at the expense of everybody else. -- F. Bastiat, "Government".
On Wed, Nov 28, 2001 at 07:35:49PM -0600, Ted Shoemaker wrote:
> Date: Wed, 28 Nov 2001 19:35:49 -0600
> From: Ted Shoemaker <email@example.com>
> To: firstname.lastname@example.org
> Subject: Please clarify
> Your TUNES website is very interesting. Anyone who claims that a new
> operating system might be a 10000% improvement over Unix
> has my interest.
> After reading much of your site, including the FAQ
> and the glossary, I still have some questions.
> The site says that you hope to make an operating system that doesn't use
> a file system. I understand the general notion of "orthogonal
> as discussed on your site -- meaning that I understand the objectives.
> I do not understand how you intend to accomplish it. How will objects
> be made orthogonally persistent? What method of categorizing will
> a file system?
> Maybe you are being vague on purpose, and hoping to keep others from
> your ideas. But if not, I would appreciate a more specific explanation.
> I have spoken to some other people who share this opinion.
> I wish you the best. I am excited about your project. If you have an
> email list or forum, please subscribe me to it.
> Thank you very much!