[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: A Conclusion Re: SchemeQL



Noel Welsh <noelwelsh@yahoo.com> writes:

> What are you implementing?  I've had some success using the PLT Web
> Server in my Wiki development.  Its nice and simple to use and quite
> fast.

Wiki development?  I like wikis, so you now have my full attention.

> ODBC is traditionally dog slow.  You may be better off with a native
> driver.  There isn't yet any I'm aware of for Postgres but it would
> make a great contribution!

I'm no C programmer, so if anyone feels like being really good to
me... ;-)

> I doubt MzScheme is causing speed issues, if TCL (dog slow, again) +
> Postgres is fast enough for the OpenACS folk.

I doubt it's MzScheme but this route as a whole isn't good enough, so
I'm looking at alternatives.  The PLT web server is one.  The other is
maybe using something simpler like Berkeley DB, but I've not looked to
see if that's supported yet.  Doing something more understandable to
the people writing the pages, like more SSI stuff, is also an option.

I'm actually not sure whether we are doing anything clever enough to
require a RDBMS and our (volunteer) writers are quite demanding.  They
want it all to be simple for them.  These are the people who keep FTP
in existence and moan constantly because I've disabled it on our
servers.

Interesting that you mention OpenACS, as that's what we're moving
*from* because it's fugly code with no separation of concerns and its
performance is gradually deteriorating as we try to do more and more
with it.  I'm hoping scheme will give us a more comprehensible and
maintainable solution, while allowing us to get back some features we
used to have in the Perl version of the site (NNTP!  I've got the
back-end code written already, although the output methods need some
attention) as well as keeping the new features from openACS that we
actually use, such as the polls and competitions.