<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> <title></title> </head> <body bgcolor="#ffffff" text="#000000"> You may want to make a binding to QDBM. (qdbm.sourceforge.net). It has nice features and a good licensing scheme. Berkeley DB is getting more commercial now a days --Hans Noel Welsh schreef: <blockquote cite="mid20060530145807.22371.qmail@web53005.mail.yahoo.com" type="cite"> <pre wrap="">Someone suggested the Berkeley DB. gdbm and dbm are cut-down versions of the same. They match your requirements but you don't seem to have looked at them. There have been several projects to bind Berkeley DB to Scheme (I tried once). One is here: <a class="moz-txt-link-freetext" href="http://wiki.eclab.byu.edu/index.cgi?DbXml">http://wiki.eclab.byu.edu/index.cgi?DbXml</a> Searching for 'gdbm mzscheme' will yield something. I don't think there are any bindings that are up-to-date and/or complete but the interface is fairly simple. All it needs is time, and with lib foreign.ss that amount of time is quite small. N. --- David Richards <a class="moz-txt-link-rfc2396E" href="mailto:vottamusic@verizon.net"><vottamusic@verizon.net></a> wrote: </pre> <blockquote type="cite"> <blockquote type="cite"> <pre wrap="">I've got the same requirements, but haven't found a </pre> </blockquote> <pre wrap="">good solution so </pre> <blockquote type="cite"> <pre wrap="">far. Last time I tried the SLIB database in DrScheme it </pre> </blockquote> <pre wrap="">was broken. </pre> <blockquote type="cite"> <pre wrap="">Moreover, as far as I remember the *disk-based* </pre> </blockquote> <pre wrap="">database was an </pre> <blockquote type="cite"> <pre wrap="">extension for SCM only. SLIBs simple a-list db is </pre> </blockquote> <pre wrap="">in-memory. As far as I can understand it SLIB has a disk-based associative array that allows 256 bytes keys and values. I need blob values -- and perhaps even longer keys, depending on the hash function to be used. I'm not about to get into the 'relational' mess in SLIB. It's way too complex for my needs. I'm not even sure it will work. And in response to another reply, yes, essentially the requirements are the mostly same as a file system. However I wish to potentially store tens of millions of variable sized objects. The 'reiser file system' looks interesting, but I'm not sure I want to tinker with the file system at kernel level on my Mac OS X system. I've been taking a look at distributed hash tables, but that doesn't seem right either for my current needs. I'd love to find a simple C program for a disk-based associative array that stores fixed length keys and variable length values. I could port it to scheme fairly quickly I think. </pre> <blockquote type="cite"> <pre wrap="">It would be nice to have an Mzscheme interface to </pre> </blockquote> <pre wrap="">Metakit </pre> <blockquote type="cite"> <pre wrap=""><a class="moz-txt-link-rfc2396E" href="http://www.equi4.com/metakit.html"><http://www.equi4.com/metakit.html></a>. </pre> </blockquote> <pre wrap="">If wishes were code, programmers would sleep. Best regards, Dave </pre> </blockquote> <pre wrap=""> Email: noelwelsh <at> yahoo <dot> com noel <at> untyped <dot> com AIM: noelhwelsh Blogs: <a class="moz-txt-link-freetext" href="http://monospaced.blogspot.com/">http://monospaced.blogspot.com/</a> <a class="moz-txt-link-freetext" href="http://www.untyped.com/untyping/">http://www.untyped.com/untyping/</a> __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around <a class="moz-txt-link-freetext" href="http://mail.yahoo.com">http://mail.yahoo.com</a> _________________________________________________ For list-related administrative tasks: <a class="moz-txt-link-freetext" href="http://list.cs.brown.edu/mailman/listinfo/plt-scheme">http://list.cs.brown.edu/mailman/listinfo/plt-scheme</a> </pre> </blockquote> </body> </html>