On Sun, 09 Dec 2001 01:38:16 +0100, Michael Hartle <[EMAIL PROTECTED]> wrote:
> Ovidiu Predescu wrote: > > >I'll let you know what's the progress on my work. If anyone is > >interested in seeing intermediate results, I can check the code in the > >scratch-pad CVS area. > > > I am definitely interested in seeing your results. I still have problems > thinking the right way for Prolog or Scheme, but this implicit (right ?) > approach might well be a breakthrough that we are all waiting for > regarding this topic. I am only hoping that the approach itself won't be > a bottleneck when it comes to serving content in a production > environment - what do you think ? We'll have to see how fast the SISC engine is. From my past experience, various Scheme implementations can be as fast or even faster than Java, with less memory consumption in the same time. Check out these pages for example: http://www.bagley.org/~doug/shootout/ Also remember that things like Gimp and Sawfish (a popular Linux window manager) use Scheme to drive their computations. Not to mention Emacs ;-), perhaps the best known Lisp application (even though it's not Scheme). I would say the slowness of Lisp is a myth these days. If we use Scheme appropriately, and the engine is carefully profiled for our usage, we can get really good performance. SISC promises to have good performance, we'll verify this. > Another question, is the current abstraction of a sitemap able to allow > different principles of sitemaps (current one, scheme-based one, etc.) > the admin/developer can choose from depending on his needs ? E.g., I'd > use the current sitemap type at the root, mounting a scheme-based > sitemap under /scheme, then in the scheme-based sitemap I'd mount some > other sitemap type under /scheme/other and so on. The scheme-based sitemap will provide the old semantic of the sitemap as well. We can even support the current syntax by adding a translation layer. In the general case, we can provide two servlets, CocoonServlet and the Scheme servlet in the same WAR files, mounted at different URLs as you mention above. This way the two implementations share the same classes, but each does it's own dispatching. Best regards, -- Ovidiu Predescu <[EMAIL PROTECTED]> http://orion.nsr.hp.com/ (inside HP's firewall only) http://sourceforge.net/users/ovidiu/ (my SourceForge page) http://www.geocities.com/SiliconValley/Monitor/7464/ (GNU, Emacs, other stuff) --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
