On 11 Sep 2009, at 1:56 am, Nicholas Indy Ray wrote:
> On Thu, Sep 10, 2009 at 7:19 AM, Aaron W. Hsu <[email protected]>
> wrote:
>> My assertion is that mapping a module language to a specific system
>> behavior such as how the modules are laid out in the file system is
>> a bad
>> idea.
>
> I understand that on some systems you can't even assume a file system
> exists.
Aye. Not just bizarre embedded Schemes, but also ones that live in
unusual worlds such as a database, which might have a *better* place
than the available filesystem.
Case in point: I'm writing a distributed database. I'd like to embed a
Scheme runtime into it for "stored procedures". There's a filesystem
on every node the database runs on, sure... but I'd rather put
libraries in the database, so they are uniformly available across the
nodes, so anything can run anywhere :-)
> However I don't think this should prevent an average scheme
> user trying to write portable code from being able to assume a fairly
> standard module layout for systems with a standard file system.
Exactly. I'd even make the table that stores modules look a bit like a
filesystem, in homage to this, for Least Surprise; I'd probably have
records with a primary key that looks like a pathname ("slib/
whatever.scm"), and whose body contains the source code of the module.
And a command-line tool to load arbitrary filesystem trees into such a
table.
ABS
--
Alaric Snell-Pym
Work: http://www.snell-systems.co.uk/
Play: http://www.snell-pym.org.uk/alaric/
Blog: http://www.snell-pym.org.uk/archives/author/alaric/
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
