narf

[Ara.T.Howard]

any strong evidence that it is faster than the built-in cgi.rb?

-a
--
===============================================================================
| email :: ara [dot] t [dot] howard [at] noaa [dot] gov
| phone :: 303.497.6469
| Your life dwells amoung the causes of death
| Like a lamp standing in a strong breeze. --Nagarjuna
===============================================================================

 

[Patrick May]

Hello,

any strong evidence that it is faster than the built-in cgi.rb?

That's interesting, I haven't tested it. I've tested narf on
different platforms using the webunit test suite, but I haven't
compared it with the original cgi.rb. I'll take a look at it tonight.

I have seen that the difference btw cgi and fastcgi/mod_ruby is an
almost an order of magnitude. That's the most direct way to improve
your web performance.

Also, narf is being deprecated. I have been taking a look at the
api, breaking what needs breaking, and putting together something new
and yummy, check it out http://ruby-web.org/.

I've been quiet for a while. I made my commitment to set aside a few
broken aspects of the api, and I want to make sure I have strong
installers for ruby-web. The ruby-web beta should be released by the
end of the week.

My plan is to improve the low end of ruby web development. Rails is
doing a great job with the application framework area, I just want to
get the basics in order.

Cheers,

Patrick

ps if you have patches for the old release of narf, let me know.
Though the future is ruby-web, I'm still maintaining the last release.

 

[Patrick May]

Ara,

any strong evidence that it is faster than the built-in cgi.rb?

narf/ruby-web is not faster than cgi.rb -- they both do the same
amount of work. narf/ruby-web runs under mod_ruby and fast cgi,
where it receives the expected speed benefits.

The main difference between ruby-web and cgi.rb is the api, error
handling, testing support, and installers. Speaking of which, I have
just released the first beta of ruby-web:

http://rubyforge.org/frs/?group_id=873

The windows installer will configure IIS, and the Mac OS X installer
will configure default Apache installation.

This is quite the beta release -- the manual is spotty, and the
installers are fresh. Still, this should give a taste of how I'd
like ruby to be used on the web.

Cheers,

Patrick

 

[Patrick May]

From: (e-mail address removed)
Subject: Re: narf
Date: September 24, 2005 3:19:18 AM EDT
To: (e-mail address removed)

Hello,

I notice much of the documentation still points toward narf instead of
ruby-web. I imagine you do plan to fix this, but I thought you'd be
interested in case you don't already know.

Thanks for the heads up, I'll be doubling back on the docs.

So how different is ruby-web from narf?

The main incompatible differences are in the templating system. I
don't know how many folks were using Narflates. It's not to
difficult to port an app, but there is enough work to tip the old
major version number.

Narf supported the use of "Narflates" via a Web:rint_template
( narf_filename, variables_hash ) method.

In ruby-web, the default templating uses erb. Templates are now
loaded via Web::load( filename ), which conceptually (to me) is more
similar to Kernel::load. Web::load looks up the appropriate handler
for an extension (.rb, .rhtml, add your own, etc.), so it can more
easily support multiple templating systems. Also, the ruby-web
interpreter uses Web::load. Thus, ruby-web is used like eruby, as an
interpreter for html/code documents.

(This points to an out of date bit with the docs -- I have a phprb
templating piece that I like, but I since decided to make erb the
standard in ruby-web for simplicities state. The templating section
of the docs still describes the phprb syntax.)

I'd like to ramble about the name. NARF is a negative definition --
Not Another Ruby Framework. NARF is more of an attitude towards
projects. That what we need aren't big ideas, but problem solving
code. For example, the joke of raa-install was that we didn't need
to define a comprehensive RAA.succ package spec, we just needed an
RAA.suck(package) installer.[1]

Ruby-Web is a better description of this particular project.

Cheers,

Patrick

1. Just of the record, most of the labor of raa-install was done by
Tom Clarke and why the lucky stiff. I pretty much just contributed
the bad pun.